VEHITS 2019 Abstracts

Area 1 - Connected Vehicles
Full Papers Short Papers
Area 2 - Intelligent Transport Systems and Infrastructure
Full Papers Short Papers
Area 3 - Intelligent Vehicle Technologies
Full Papers Short Papers
Area 4 - Data Analytics
Full Papers Short Papers
Area 5 - Sustainable Transport
Full Papers

Area 1 - Connected Vehicles

Full Papers
Paper Nr: 12
Title:

CPD: Crowd-based Pothole Detection
Authors:

Florian Wirthmueller, Jochen Hipp, Kai-Uwe Sattler and Manfred Reichert
Abstract: Potholes and other damages of the road surface constitute a problem being as old as roads are. Still, potholes are widespread and affect the driving comfort of passengers as well as road safety. If one knew about the exact locations of potholes, it would be possible to repair them selectively or at least to warn drivers about them up to their repair. However, both scenarios require their detection and localization. For this purpose, we propose a crowd-based approach that enables as many of the vehicles already driving on our roads as possible to detect potholes and report them to a centralized back-end application. Whereas each single vehicle provides only limited and imprecise information, it is possible to determine these information more precisely when collecting them at a large scale. These more exact information may, for example, be used to warn following vehicles about potholes lying ahead to increase overall safety and comfort. In this work, this idea is examined and an offline executable version of the desired system is implemented. Additionally, the approach is evaluated with a large database of real-world sensor readings from a testing fleet and therefore its feasibility is proved. Our investigation shows that the suggested CPD approach is promising to bring customers a benefit by an improved driving comfort and higher road safety.
Download
Paper Nr: 20
Title:

On the Road with Third-party Apps: Security Analysis of an In-vehicle App Platform
Authors:

Benjamin Eriksson, Jonas Groth and Andrei Sabelfeld
Abstract: Digitalization has revolutionized the automotive industry. Modern cars are equipped with powerful Internet-connected infotainment systems, comparable to tablets and smartphones. Recently, several car manufacturers have announced the upcoming possibility to install third-party apps onto these infotainment systems. The prospect of running third-party code on a device that is integrated into a safety critical in-vehicle system raises serious concerns for safety, security, and user privacy. This paper investigates these concerns of in-vehicle apps. We focus on apps for the Android Automotive operating system which several car manufacturers have opted to use. While the architecture inherits much from regular Android, we scrutinize the adequateness of its security mechanisms with respect to the in-vehicle setting, particularly affecting road safety and user privacy. We investigate the attack surface and vulnerabilities for third-party in-vehicle apps. We analyze and suggest enhancements to such traditional Android mechanisms as app permissions and API control. Further, we investigate operating system support and how static and dynamic analysis can aid automatic vetting of in-vehicle apps. We develop AutoTame, a tool for vehicle-specific code analysis. We report on a case study of the countermeasures with a Spotify app using emulators and physical test beds from Volvo Cars.
Download
Paper Nr: 26
Title:

Cooperative Driving in Mixed Traffic with Heterogeneous Communications and Cloud Infrastructure
Authors:

Rico Auerswald, Roman Busse, Markus Dod, Richard Fritzsche, Alexander Jungmann, Michael Klöppel-Gersdorf, Josef F. Krems, Sven Lorenz, Franziska Schmalfuß, Sabine Springer and Severin Strobl
Abstract: In this paper we introduce an Intelligent Transport System (ITS), designed for enabling cooperative driving manoeuvres in mixed traffic scenarios considering heterogeneous communications and cloud infrastructure systems. We present an architecture that enables connected vehicles to access ITS services independent of their underlying communication technology. This is achieved by introducing a large scale communication system including the road-side infrastructure as well as a heterogeneous cloud. We present insights from the Automated Connected Vehicle (ACV) concept and examine human factors elaborating on the experience of two aspects: driving in an ACV as well as driving in a Non-Automated Connected Vehicle (NACV), interacting with an ACV. Furthermore, we present insights of initial demonstrations, emphasizing that the system works well in real traffic scenarios.
Download
Paper Nr: 28
Title:

Cross-provider Platoons for Same-day Delivery
Authors:

Sînziana-Maria Sebe, Philipp Kraus, Jörg P. Müller and Stephan Westphal
Abstract: Platooning – vehicles travelling close together behaving as a unit – aims to improve network throughput both on highways and in urban traffic. We study the problem of platoon formation in an urban environment using the scenario of logistic service providers equipped with fleets of autonomously driving pods to carry out same-day delivery tasks by creating cross-provider platoons. The novelty of our work is that we investigate the problem of cross-provider platoons, i.e., platoons with members from different self-interested logistic service providers. Our aim is to study platoon formation mechanisms and possible benefits of cross-provider platooning using simulation. We formulate optimal platoon formation as an integer linear optimisation problem (ILP), aiming to find the longest sub-routes to be shared between vehicles by platooning. The proposed method was implemented and tested on a mesoscopic model to simulate platoon formation and operation, on real network data with realistic background traffic models. Comparing our method to a simpler route matching algorithm reveals comparable system level performance; however, our method performs better with respect to local participant utility, i.e.appears more suited to take vehicle/provider preferences into account.
Download
Paper Nr: 44
Title:

Red-Zone: Towards an Intrusion Response Framework for Intra-vehicle System
Authors:

Mohammad Hamad, Marinos Tsantekidis and Vassilis Prevelakis
Abstract: Modern vehicles are increasingly equipped with highly automated control systems both for driving and for passenger comfort. An integral part of these systems are the communication channels that allow the on-board systems to interact with passenger devices (e.g. tablets), ITS systems (e.g. road-side units), and other vehicles. These advances have significantly enlarged the attack surface and we already have numerous instances of successful penetration of vehicular networks both from inside the vehicle and from the outside. Traditional mechanisms for detecting and responding to such attacks are ill-suited to the vehicular domain mainly due to the fact that the entire process of dealing with an attack must be handled automatically and in a way that does not affect safety or severely impacts the continued availability of the vehicle or its key systems. Once a security breach is suspected, the system must evaluate the circumstances in order to determine whether the threat is real (and not a false positive) and select the optimal response through the use of an Intrusion Response System (IRS). Although IRSs have been adopted in other domains, there is a lack of such systems in the vehicular field. In this paper, we investigate the challenges and requirements for integrating such a mechanism inside a vehicle. In addition, we present an Intrusion Response System based on the Red-Zone principle which meets the identified requirements. Finally, we discuss the integration of IRS through the vehicle system development and the different aspects which support such a process.
Download
Short Papers
Paper Nr: 40
Title:

Safety-relevant V2X Beaconing in Realistic and Scalable Heterogeneous Radio Propagation Fading Channels
Authors:

Daniel Bischoff, Harald Berninger, Steffen Knapp, Tobias Meuser, Björn Richerzhagen, Lars Häring and Andreas Czylwik
Abstract: Performance evaluations for heterogeneous communication technologies in the area of V2X safety applications for either improvement, comparison or combination purposes are in general focusing on the realistic representation of the upper communication stack layers, but therefore - often for the sake of simplicity - reducing the radio propagation channel to a maximum range model. The impact and hence the importance to model the environment dependent propagation effects in a representative manner has already been stressed in the literature several times - but separately for ad-hoc or cellular systems and not under the consideration of V2X safety-beaconing applications. By combining a realistic heterogeneous radio propagation channel model with a state-of-the-art V2X communication stack, a representative performance comparison of safety-relevant beaconing applications for 802.11p single-hop broadcast (SHB) and LTE Geocast can be conducted. Our simulation results show that the effects caused by the radio propagation channel cannot be neglected as they significantly impact key communication performance metrics such as channel gain, packet error ratio (PER) and channel load, where we primarily focus on the latter one to give further research directions for an efficient dissemination of safety-relevant V2X beacons.
Download
Paper Nr: 41
Title:

Simulation Platform for Connected Heterogeneous Vehicles
Authors:

Tobias Meuser, Daniel Bischoff, Ralf Steinmetz and Björn Richerzhagen
Abstract: The increasing number of connectivity features in current vehicles poses additional challenges for large-scale vehicular communication systems. Already deployed systems rely on the cellular network infrastructure, while the Wifi-based 802.11p standard will likely be implemented on a large scale in the next years. As real-world tests are costly, simulations are used to develop mechanisms for efficient short-range communication via 802.11p. However, efficient long-range communication between vehicles is pivotal for non-safety related information sharing. Current simulators often focus on short-range communication exchange, while approaches for efficient long-range communication are barely considered in automotive scenarios. To enable rapid development of new approaches, we propose a scalable simulation environment for automotive applications. Our contributions are (i) the realistic modeling of heterogeneous vehicles including sensors and network interfaces, (ii) the automated generation of road properties like accidents and jams, and (iii) a configurable back-end infrastructure distributing events to the vehicles. All of the above contributions enable rapid prototyping and evaluation of automotive applications in various environments. We showcase two exemplary use cases to demonstrate the versatility of our simulation framework: an efficient road-based dissemination approach for long-range information exchange and a distributed information validation approach.
Download
Paper Nr: 86
Title:

A Heuristic Decision Maker Algorithm for Opportunistic Networking in C-ITS
Authors:

Rodrigo Silva, Christophe Couturier, Jean-Marie Bonnin and Thierry Ernst
Abstract: The number of connected devices is growing worldwide and connected and cooperative vehicles should be a major element of such ecosystem. However, for ubiquitous connectivity it is necessary to use various wireless technologies, such as vehicular WiFi (ITSG5, and DSRC), urban WiFi (e.g., 802.11 ac,g,n), 802.15.4, cellular (3G, 4G, and 5G under preparation). In such an heterogeneous access network environment, it is necessary to provide applications with transparent decision making mechanisms to manage the assignment of data flows over available networks. In this paper, we propose the Ant-based Decision Maker for Opportunistic Networking (AD4ON), a Decision Maker (DM) algorithm capable to manage multiple access networks simultaneously, attempting to choose the best access network for each data flow. Moreover, the AD4ON is capable to increase decision’s stability, to reduce the ping-pong effect and to manage decisions flow by flow while maximizing flow’s satisfaction.
Download
Paper Nr: 89
Title:

Collaborative and Distributed QoS Monitoring and Prediction: A Heterogeneous Link Layer Concept towards always Resilient V2X Communication
Authors:

Daniel Hincapie, Ahmad Saad and Josef Jiru
Abstract: Vehicle-To-Everything (V2X) communication is a fundamental pillar of autonomous driving. It enables the exchange of safety-critical data between vehicles, infrastructure and pedestrians to enhance the awareness of the surrounding environment and coordinate the execution of collective functionalities vital to achieve full automation. Due to the safety-critical nature of the interchanged information, V2X communication must be resilient, so that it provides reliable connectivity despite of the very dynamic characteristics of both its environment and network topology. In this position paper, we propose a novel concept that aims at achieving resilient V2X communication. We introduce the Quality of Service Manager (QoSM), a collaborative and distributed implementation concept for the Heterogeneous Link Layer (HLL) that operates on the top of the Medium Access Control (MAC). The QoSM first monitors and predicts QoS indicators of Radio Access Technologies (RAT) in Heterogeneous Vehicular Networks (HetVNET). Then, it determines, under the principle of Always Resiliently Connected (ARC), and sets timely the configuration settings of RAT that meet performance and reliability requirements of autonomous driving applications. Should it not be possible to fulfill applications demands, the QoSM can instruct applications in advance to lower the requirements or run in a safe mode. Like in many autonomous driving applications, the concept of our proposed QoSM is distributed and collaborative to enhance accuracy, self-awareness and safety. QoSMs shall be grouped hierarchically according to correlation of applications demands, conditions of communication links and mobility information. Group’s members share monitored and predicted indicators, as well as configuration settings. This information is used to determine collectively the configuration of the HetVNET. On the one hand, sharing information among QoSMs increases the amount of correlated data used by prediction algorithms, which improves prediction accuracy. On the other hand, hierarchical groups allow to extend the proposed methodology to other hierarchical elements of the access and core network. With this position paper, we intend to open the discussion on the importance of implementing protocols for sharing parameters that allow distributed and collaborative QoS management for resilient V2X communication.
Download
Paper Nr: 1
Title:

Preliminary Study on Road Slipperiness Detection using Real-Time Digital Tachograph Data
Authors:

Jinhwan Jang
Abstract: Faced with the high rate of commercial vehicle-related traffic accidents, digital tachographs (DTGs) are mandatorily installed in commercial vehicles in Korea. However, the current DTGs do not seem to be effective for reducing accidents. One reason for this can be attributed to the absence of useful information for drivers under dangerous road conditions such as black ice. In this study, an innovative technique to identify slippery spots on the road using DTG data is proposed. The DTG can collect two types of vehicle speed: one is wheel rotational speed and the other is vehicle transitional speed. The difference between the two speeds is referred to as wheel slip, which can be exploited as a surrogate measure for detecting road slipperiness. A confidence interval of wheel slip was established using data collected in dry road conditions; if any data point that exceeds the predefined confidence interval is observed, a slippery road spot can be identified. The proposed method was preliminarily tested in four types of winter road conditions and showed satisfactory results.
Download
Paper Nr: 45
Title:

Intelligent Content Management System for Tourism Smart Mobility: Approach and Cloud-based Android Application
Authors:

Alexander Smirnov, Alexey Kashevnik, Nikolay Shilov, Sergei Mikhailov, Oleg Gusikhin and Harry Martinez
Abstract: Intelligent content management systems have become more popular over the last few years in the tourism industry due to the significantly increasing impact on revenue. Such systems are the part of the smart mobility concept. Smart mobility allows tourists to become more comfortable with transportation in an unknown city by providing interesting information about places seen during their trip. Traditional taxies provide quick transportation from the point A to point B but people sometimes are interested in seeing attractions during their trips and are willing to spend more time and money to do so. Integration of the smartphone application with the vehicle infotainment system provides opportunities of new smart services construction that is based on information from vehicle sensors and Internet connection as well as utilization of in-cabin infrastructure and communication with the driver (suggesting the route preferred by the tourist, speed while going around attractions, adjust the temperature, music, and etc.). The paper presents an approach to smart mobility system development and its evaluation by showing the tourist video information about attractions around.
Download
Paper Nr: 77
Title:

Practical Security and Privacy Threat Analysis in the Automotive Domain: Long Term Support Scenario for Over-the-Air Updates
Authors:

Alexandr Vasenev, Florian Stahl, Hayk Hamazaryan, Zhendong Ma, Lijun Shan, Joerg Kemmerich and Claire Loiseaux
Abstract: Keeping a vehicle secure implies provide of a long-term support, where over-the-air updates (OTA) play an essential role. Clear understanding of OTA threats is essential to counter them efficiently. Existing research on OTA threats often exclude human actors, such as drivers and maintenance personnel, as well as leave aside privacy threats. This paper addresses the gap by investigates security and privacy OTA threats relevant for vehicle manufacturers for the whole product lifecycle. We report on a practical scenario “long term support”, its data flow elements, and outcomes of threat analyses. We apply state of the art approaches, such as STRIDE (extended with an automotive template) and LINDDUN, to an automotive case and consider an automotive-specific UNECE OTA threat catalogue. Outcomes indicate complementarity of these methods and provide inputs to studies how well they address practical automotive cases.
Download

Area 2 - Intelligent Transport Systems and Infrastructure

Full Papers
Paper Nr: 10
Title:

Securing Cargo during Transport on Roads of Different Quality
Authors:

Martin Vlkovský and Hana Vlachová
Abstract: The article compares the magnitude of shocks generated by the Tatra T-810 on two types of roads (high-quality – highway and lower-quality – roads paved with granite blocks). As the primary data, sets of acceleration coefficients in the three axes (x, y and z) were used as part of a transport experiment using a three-axis accelerometer with a datalogger and a calibration certificate – OM-CP-ULTRASHOCK-5-CERT. Data analysis is performed using descriptive statistics. The mean values and variations of measured acceleration coefficients on the roads we examined are compared. The graphical comparison of the roads studied is covered in a separate section. The results of the transport experiment show that the magnitude of generated shocks is even higher at a lower average transport speeds on a low-quality roads. The distribution of acceleration coefficient values also differs for both roads.
Download
Paper Nr: 37
Title:

On-demand Ride-sharing Services with Meeting Points
Authors:

Sevket Gökay, Andreas Heuvels and Karl-Heinz Krempels
Abstract: On-demand ride-sharing services propose an alternative transportation mode to public and private transportation. They have similarities with private transportation, since the customers have the convenience of travelling from and to any desired location while defining the departure (or arrival) time. They resemble public transportation in multiple customers sharing a vehicle with similar journeys. This work proposes an approach to improve the throughput of on-demand ride-sharing services by introducing meeting points. The idea bases on combining a vehicle’s nearby location visits (whether for pick-up or drop-off) into one, if temporal and spatial constraints are held, in order to reduce the vehicle detour costs. It, by design, diminishes customer convenience, since walking legs are introduced and departure/arrival times might deviate from what is desired. The trade-offs are evaluated by running two simulations, one without and one with meeting points. The results indicate that even a small customer inconvenience can yield significant increase in the number of satisfied trip requests without increasing vehicle costs.
Download
Paper Nr: 43
Title:

Analyzing Traffic Signal Performance Measures to Automatically Classify Signalized Intersections
Authors:

Dhruv Mahajan, Tania Banerjee, Anand Rangarajan, Nithin Agarwal, Jeremy Dilmore, Emmanuel Posadas and Sanjay Ranka
Abstract: Traffic signals are installed at road intersections to control the flow of traffic. An optimally operating traffic signal improves the efficiency of traffic flow while maintaining safety. The effectiveness of traffic signals has a significant impact on travel time for vehicular traffic. There are several measures of effectiveness (MOE) for traffic signals. In this paper, we develop a work-flow to automatically score and rank the intersections in a region based on their performance, and group the intersections that show similar behavior, thereby highlighting patterns of similarity. In the process, we also detect potential bottlenecks in the region of interest.
Download
Paper Nr: 46
Title:

Global and Local Spatial Autocorrelation of Motorcycle Crashes in Chile
Authors:

Carola Blazquez and María J. Fuentes
Abstract: In Chile, the usage of motorcycles as a mode of transport is growing in unison with the number of crashes that have arisen in recent years. Spatial statistical methods were used in this study to determine whether motorcycle crashes showed spatial clustering over time from a global and local perspective. The global spatial autocorrelation results indicate that high intensity clusters of collisions at intersections with traffic signals and curved road sections resulting in fatalities persisted during the five-year study period. Locally, recurrent high spatial patterns of motorcycle collisions arose along straight road sections and on sunny days due to the loss of control of the vehicle, or the imprudence of the driver or pedestrian. Communes located in the centre zone of Chile, particularly in the city of Santiago and the surrounding areas, presented a large number of highly clustered crash attributes. The findings of this study may help authorities to target efforts towards policy measures to improve motorcycle safety in Chile.
Download
Paper Nr: 62
Title:

Expert Competitive Traffic Light Optimization with Evolutionary Algorithms
Authors:

Yann Semet, Benoit Berthelot, Thierry Glais, Christian Isbérie and Aurélien Varest
Abstract: We present a complete system to optimize traffic lights green phases and temporal offsets based on a combination of microscopic simulation and black box, evolutionary algorithms. We also report the outcome of an AI versus experts comparison workshop conducted with our algorithm and seasoned experts from a specialized traffic engineering office. Experimental results indicate that the proposed algorithmic scheme significantly outperforms expert efforts. Our system entails a memetic (genetic+gradient) calibration module to adapt the Origin/Destination (O/D) matrix to current traffic conditions, an inoculation procedure to incorporate existing traffic light programs, genetic multi-objective optimization capabilities and sound metrics. Experiments are conducted over several real world datasets of operational sizes from the Paris outskirts and various other French urban areas. Our experimental outcome is threefold. First, we report the success of the memetic calibration module in adjusting the simulator’s O/D matrix to a point with variation levels corresponding to recorded sensor data. Second, we confirm the ability of the system to obtain significant gains on that sound basis: gains ranging from 15% to 35% are consistently reached on both traffic jams reduction and pollutant emissions. Most importantly, we report the outcome of the comparison workshop: a formalized methodology followed by experts to manually optimize traffic lights, iterative experimental logs tracing the application of that methodology to two real world cases and comparable results obtained by the algorithm on the same cases. Results indicate that the AI module performs significantly better than experts in both speed and final solution quality.
Download
Paper Nr: 64
Title:

Smart Parking Zones using Dual Mode Routed Bluetooth Fogged Meshes
Authors:

Paul Seymer, Duminda Wijesekera and Cing-Dao Kan
Abstract: Contemporary parking solutions are often limited by the need for complex sensor infrastructures to perform space occupancy detection, and costly to maintain ingress and egress parking systems. For outdoor lots, network infrastructure and computational requirements often limit the availability of innovative technology. We propose the use of Bluetooth Low Energy (BLE) beacon technology and low power sensor nodes, coupled with sensible placement of computational support and data storage near to the sensor network (a Fog computing paradigm) to provide a seamless parking solution capable of providing parking maintainers with accurate determinations of where vehicles are parked within the lot. Our solution is easy to install, easy to maintain, and does require significant alterations to the existing parking structures.
Download
Short Papers
Paper Nr: 2
Title:

Lane Effect of Traffic Flow Analysis in India
Authors:

Tsutomu Tsuboi
Abstract: The aim of this research is to develop quantitative analysis method for emerging countries, especially focusing on major city in India where they are facing negative impact by transportation such as CO2 emission growth, many traffic fatalities, large fuel consumption, and air pollution as a result. In this research, we installed more than ten traffic monitoring traffic counter cameras in Ahmedabad city of Gujarat state of India. The monitoring cameras detect traffic vehicle and capture several traffic data such as vehicle numbers, vehicle speed, traffic occupancy, vehicle density, gap length between vehicle to vehicle and so on. And each data is collected by every minutes per roads. Therefore the collected data becomes more than 432,000 points per months. In order to analyse the traffic data, author recognize special features of the collected traffic data and the Envelope Observation (EO) for traffic flow characteristics by measurement data is useful for obtaining traffic flow equation. The unique feature of emerging counties traffic data is widely spread plots at the traffic basic characteristics such as traffic density to speed curve and traffic density to traffic volume curve. But there is clearly boundary line in those curves. Author uses EO analysis to fit traffic flow parameters to these boundary lines. By defining traffic flow parameters, it is able to obtain the traffic flow value such as free speed traffic flow, critical traffic volume, and critical traffic density. After obtaining traffic parameters, it is able to create traffic flow equation for each measured road and even each lane of its road. The uniqueness of this research is extension of analysis for the road lane effect for the traffic congestion by correlation ratio analysis between driving lane and passing lane of each road. As the result of this analysis, it becomes clear that congestion condition of roads makes the different traffic flow characteristics by driving and passing lane. This is the first time to explain the lane effect for traffic congestion on the basis of the EO method.
Download
Paper Nr: 8
Title:

Traffic Monitoring using an Object Detection Framework with Limited Dataset
Authors:

Vitalijs Komasilovs, Aleksejs Zacepins, Armands Kviesis and Claudio Estevez
Abstract: Vehicle detection and tracking is one of the key components of the smart traffic concept. Modern city planning and development is not achievable without proper knowledge of existing traffic flows within the city. Surveillance video is an undervalued source of traffic information, which can be discovered by variety of information technology tools and solutions, including machine learning techniques. A solution for real-time vehicle traffic monitoring, tracking and counting is proposed in Jelgava city, Latvia. It uses object detection model for locating vehicles on the image from outdoor surveillance camera. Detected vehicles are passed to tracking module, which is responsible for building vehicle trajectory and its counting. This research compares two different model training approaches (uniform and diverse data sets) used for vehicle detection in variety of weather and day-time conditions. The system demonstrates good accuracy of given test cases (about 92% accuracy in average). In addition, results are compared to non-machine learning vehicle tracking approach, where notable vehicle detection accuracy increase is demonstrated on congested traffic. This research is fulfilled within the RETRACT (Enabling resilient urban transportation systems in smart cities) project.
Download
Paper Nr: 27
Title:

Maneuver-based Adaptive Safety Zone for Infrastructure-Supported Automated Valet Parking
Authors:

Valerij Schönemann, Mara Duschek and Hermann Winner
Abstract: One of the major challenges for the release of fully automated driving is the design of safe vehicle automation systems. This work presents a structure to determine a maneuver-specific and adaptive safety zone for collision avoidance. For this, the overall automated driving system is split into functional scenarios that occur during the driving task in the operational design domain. Maneuvers are derived from the given scenarios and car park layouts. Minimum safety distances are determined by injecting worst-case parameters into derived maneuvers. The superposition of these safety distances leads to a new term: the safety zone. The safety zone adapts its size according to the performed maneuver as well as the dynamic driving parameters of the engaged traffic participants such as velocities, timing constraints and deceleration capabilities. The methodology is applied on the example of cooperative automated valet parking (AVP).
Download
Paper Nr: 39
Title:

A Traffic Signal Controller for an Isolated Intersection using Fuzzy Logic Model
Authors:

Nada B. AlNaser and Yaser E. Hawas
Abstract: With the revolution of the new technologies and intelligent transportation systems (ITS) as one category of the artificial intelligent (AI) models, fuzzy logic models (FLMs) were considered as one of the promising methods applied in signalized intersections. In general, results show significant improvements on the efficiency of the traffic networks and intersections. This paper presents a new method of developing an optimal real-time traffic signal controller using the fuzzy logic technique/method (FLM), taking into consideration all various incoming traffic flows. The developed FLM was designed for an isolated intersection with four legs, split phasing, and three different movements (through, right, and left). This research aims at developing an FLM that replicate the control settings of optimized methods. Calibration and validation tests were conducted to ensure accuracy and efficiency of the developed model. Results show that the developed FLM outputs are close to those obtained from optimum methods for traffic signal control systems.
Download
Paper Nr: 61
Title:

Pattern Model Stratification to Decision-making Support of Transport Infrastructure Management
Authors:

Tatyana Mikheeva, Sergey Mikheev and Oleg Golovnin
Abstract: Design and engineering of patterns requires tools to enable these processes by means of universal creation and dynamic modification of objects. The article introduces the basic definitions of pattern design (hereinafter P-models), specifies the types and structures of patterns, determines the organization of object design based on P-models. The introduced axiomatic and systematization are based on the use of original definitions and form the concept of stratified pattern construction of objects and associations. A stratified pattern model is considered as an information, methodological and implementation basis for the design of the decision support system of transport infrastructure management. The synthesis of the decision support system of transport infrastructure management is made on the fundamental methodological basis, determined by the object-oriented paradigm. The synthesis of universal structures based on patterns is considered on the formal declaration of objects invariant to their subject orientation. Along with the synthesis of methodical P-models, the patterns of formalization and systematization of relevant information-logical and functional aspects of transport infrastructure are described. The implementation P-models contains the effectiveness of the pattern design of the decision support system of transport infrastructure management of the processes of modelling, management, experimental studies.
Download
Paper Nr: 67
Title:

Intermodal Containers Transportation: How to Deal with Threats?
Authors:

Sergej Jakovlev, Arunas Andziulis, Audrius Senulis and Miroslav Voznak
Abstract: This paper provides an overview of the port container inspection techniques and procedures (standardized security procedures) relating to the detection of illicit material in containers. These procedures affect the duration of the containers transportation periods in different parts of the transport chain, according to the 2002 Container Security Initiative (CSI) regulations. The main object of this work – to demonstrate the inability of standard systems and associated technologies to deal with current threats and to propose solutions that are in line with the “intelligent containers” worldwide initiative.
Download
Paper Nr: 82
Title:

Bus Regularity Evaluation using the Gini Index and the Lorenz Curve: A Case Study of New Delhi Bus Network
Authors:

Amine Melakhsou and Neila Bhouri
Abstract: The ability of a public transport system to provide regular services is the main attraction for the system users. Assessing the regularity of the provided services from the user’s perspective is thus crucial for stakeholders in order to establish actions for maintaining or improving their system reliability level and therefore increasing the number of the public transport users. The purpose of this paper is to reveal the pertinence of the Gini Index based on the Lorenz curve as headway and travel time regularity indicator and to carry out a case study of the reliability of a bus operator of the city of New Delhi. We began by reconstituting the missed data in the provided automatic vehicle location data using an approximate approach and then, using correlation coefficients, we studied the linear relationships, before and after data reconstruction, between Gini Index and some of the most used regularity measures; headway regularity, headway adherence, standard deviation and travel time variability. Results show that headway adherence and standard deviation are the two indicators that have the higher correlations with the Gini index and that Gini index is less influenced by missing data and errors.
Download
Paper Nr: 87
Title:

A Framework for Personalised HMI Interaction in ADAS Systems
Authors:

Yannis Lilis, Emmanouil Zidianakis, Nikolaos Partarakis, Stavroula Ntoa and Constantine Stephanidis
Abstract: Personalisation features of Advanced Driver Assistant Systems (ADAS) can improve safety and driving experience. However, they are typically developed in an ad-hoc, application-specific and vehicle-specific manner, resulting in tightly coupled implementations that are difficult to extend, while disallowing reuse of personalisation code or even personalisation logic across different setups. In this context, this paper proposes a framework for supporting personalised HMI interaction in ADAS systems, developed in the context of the H2020 ADAS&ME project. The framework is based on a rule engine that uses a customisable and extensible set of personalisation and adaptation rules, provided by automotive domain and HMI experts, and evaluates them according to the driver, vehicle and environment to produce HMI activation and GUI personalisation and adaptation decisions. Personalised HMI modality selection is realised by taking into account all available input and output modalities of the vehicle and maintaining bindings for their activation. At the same time, GUI personalisation is handled automatically through a GUI toolkit of personalisable and adaptable user controls that can be used for developing any GUI application requiring personalisation features. The paper presents the design and development of the framework and validates it by deploying it in two case studies.
Download
Paper Nr: 36
Title:

Time Zone Impact for Traffic Flow Analysis of Ahmedabad City in India
Authors:

Tsutomu Tsuboi
Abstract: This paper describes time zone impact for traffic flow analysis in an one of major city in India based on one month real traffic monitoring big data. The target city is Ahmedabad of Gujarat state where is located in the west part of India. The current population in Ahmedabad is about 7.8 Million and it is one of rapid economic growing city. These days, the traffic congestion in the city become one of major issues. In order to analyse traffic congestion, large amount of the traffic big data is needed and it is collected through the traffic monitoring camera. The measurement of the data is traffic density, traffic occupancy and average of speed of vehicles which is measured at the road by every minute. The traffic data in emerging countries is not well analyzed so far because of difficulty of collecting traffic data. Author has a chance to involve one of traffic project which provides traffic condition to the drivers through traffic information boards and makes suggestions for avoiding traffic congestion. The current judgement of the traffic congestion is based on the occupancy of the road which is one of traffic flow parameters. This occupancy is not so accuracy sometimes because of difficulty of 100 % vehicle sensing. In this paper, it describes the time zone basis traffic flow analysis in the traffic flow characteristics such as traffic density to average vehicle speed curve, traffic density to traffic volume curve, and traffic volume to average vehicle speed. This analysis is able to identify the effect of time zone to traffic flow condition and provide more appropriate occupancy level for traffic congestion.
Download
Paper Nr: 70
Title:

Autonomous Driving of Commercial Vehicles within Cordoned Off Terminals
Authors:

Nathalie Brenner, Andreas Lauber, Carsten Eckert and Eric Sax
Abstract: In recent years, the development of autonomous trucks has progressed rapidly. It can be assumed that such vehicles will be ready within the next decade. In order to make use of the advantages of automated driving along the entire transport chain, it is necessary to use the autonomous vehicles on public roads as well as on the terminal areas. The paper presents the extent to which it is possible to adopt autonomously driving trucks to closed terminal areas. Further it discusses the technical, operational and legal requirements for vehicles, transport service providers and terminals involved. Based on the requirements a concept for autonomous driving of commercial vehicles in cordoned off areas is presented. Afterwards this concept is transformed with the current processes on a fully automated container terminal into a concrete example. This example shows how autonomous commercial vehicles can be integrated in the operational processes of an existing terminal.
Download
Paper Nr: 88
Title:

A High-level Category Survey of Dial-a-Ride Problems
Authors:

Sevket Gökay, Andreas Heuvels and Karl-Heinz Krempels
Abstract: Dial-a-Ride Problem (DARP) is an active research field since 1980. Many on-demand transport concepts like Dial-a-Ride (DAR) services, Demand-Responsive Transport (DRT) and ride-sharing share the common objective of solving DARP. Along with its application areas changing over the years, the problem continues to draw increasing attention with growing diversity of requirements, constraints and features. This paper examines the research on DARP with respect to the feature categories the solutions consider in order to discover DARP variants that most works focus on.
Download

Area 3 - Intelligent Vehicle Technologies

Full Papers
Paper Nr: 15
Title:

Mental Imagery for Intelligent Vehicles
Authors:

Alice Plebe, Riccardo Donà, Gastone Pietro Papini Rosati and Mauro Da Lio
Abstract: The research in the design of self-driving vehicles has been boosted, in the last decades, by the developments in the fields of artificial intelligence. Despite the growing number of industrial and research initiatives aimed at implementing autonomous driving, none of them can claim, yet, to have reached the same driving performance of a human driver. In this paper, we will try to build upon the reasons why the human brain is so effective in learning tasks as complex as the one of driving, borrowing explanations from the most established theories on sensorimotor learning in the field of cognitive neuroscience. The contribution of this work would like to be a new point of view on how the known capabilities of the brain can be taken as an inspiration for the implementation of a more robust artificial driving agent. In this direction, we consider the Convergence-divergence Zones (CDZs) as the most prominent proposal in explaining the simulation process underlying the human sensorimotor learning. We propose to use the CDZs as a “template” for the implementation of neural network models mimicking the phenomenon of mental imagery, which is considered to be at the heart of the human ability to perform sophisticated sensorimotor controls such driving.
Download
Paper Nr: 16
Title:

Dynamic Vehicle Routing under Uncertain Travel Costs and Refueling Opportunities
Authors:

Giorgos Polychronis and Spyros Lalis
Abstract: We study the vehicle routing problem for a system where there is some uncertainty regarding both the cost of travel and the refueling opportunities. Travel cost stands for the energy spent by the vehicle to move between locations. Refueling opportunities are offered at known locations where the vehicle can harvest or re-gain some of the lost energy. The objective is to visit a set of predefined locations without exhausting the energy of the vehicle. We describe the problem in a formal way, and propose a heuristic algorithm for taking routing decisions at runtime. We evaluate the algorithm for a grid topology as a function of the number of locations to be visited and the autonomy degree of the vehicle, showing that the proposed algorithm achieves good results as long as the energy margins are not very tight.
Download
Paper Nr: 22
Title:

Evaluation of Embedded Camera Systems for Autonomous Wheelchairs
Authors:

Cristian Vilar, Benny Thörnberg and Silvia Krug
Abstract: Autonomously driving Power Wheelchairs (PWCs) are valuable tools to enhance the life quality of their users. In order to enable truly autonomous PWCs, camera systems are essential. Image processing enables the development of applications for both autonomous driving and obstacle avoidance. This paper explores the challenges that arise when selecting a suitable embedded camera system for these applications. Our analysis is based on a comparison of two well-known camera principles, Stereo-Cameras (STCs) and Time-of-Flight (ToF) cameras, using the standard deviation of the ground plane at various lighting conditions as a key quality measure. In addition, we also consider other metrics related to both the image processing task and the embedded system constraints. We believe that this assessment is valuable when choosing between using STC or ToF cameras for PWCs.
Download
Paper Nr: 25
Title:

Accelerated RRT* and Its Evaluation on Autonomous Parking
Authors:

Jiri Vlasak, Michal Sojka and Zdeněk Hanzálek
Abstract: Finding a collision-free path for autonomous parking is usually performed by computing geometric equations, but the geometric approach may become unusable under challenging situations where space is highly constrained. We propose an algorithm based on Rapidly-Exploring Random Trees Star (RRT*), which works even in highly constrained environments and improvements to RRT*-based algorithm that accelerate computational time and decrease the final path cost. Our improved RRT* algorithm found a path for parallel parking maneuver in 95 % of cases in less than 0.15 seconds.
Download
Paper Nr: 38
Title:

Lane Accurate Detection of Map Changes based on Low Cost Smartphone Data
Authors:

Florian Jomrich, Daniel Bischoff, Steffen Knapp, Tobias Meuser, Björn Richerzhagen and Ralf Steinmetz
Abstract: Self-driving vehicles rely on High Definition Street Maps (HD Maps) to ensure the safety and comfort of their driving capabilities. However, the road network infrastructure is subject to constant changes (e.g. through constructions works, accidents, ...). Such changes have to be quickly identified to avoid dangerous driving situations, for example through a reduction of driving speed or the safe handover of driving control back to the human. To address this issue we propose a road hazard detection algorithm that identifies and marks the extent of such changes based on crowdsourced GNSS data. To increase the detection speed of our proposed algorithm, we only rely on sensor information in the collection process, that is not only available through vehicles, but as well by cheap and ubiquitous devices carried on by the passengers such as smartphones. To deal with the limited accuracy of the collected data, we enhance existing algorithmic clustering approaches by leveraging additional meta-data such as the quality of the collected GNSS points and the vehicle’s current lane position. Our concept is evaluated with real world measurements in a highway construction site scenario showing improved performance in comparison to the Kernel Density Estimation reference algorithm, used versatile in Related Work.
Download
Paper Nr: 55
Title:

Front-View Vehicle Damage Detection using Roadway Surveillance Camera Images

Authors:

Burak Balci, Yusuf Artan, Bensu Alkan and Alperen Elihos
Abstract: Vehicle body damage detection from still images has received considerable interest in the computer vision community in recent years. Existing methods are typically developed towards the auto insurance industry to minimize the claim leakage problem. Earlier studies utilized images taken from short proximity (< 3 meters) to the vehicle or to the damaged region of vehicle. In this study, we investigate the vehicle frontal body damage detection using roadway surveillance camera images. The proposed method utilizes deep learning based object detection and image classification methods to determine damage status of a vehicle. The proposed method combines the symmetry property of vehicles’ frontal view and transfer learning concept in its inference process. Experimental results show that the proposed method achieves 91 % accuracy on a test dataset.
Download
Paper Nr: 81
Title:

Object Detection Probability for Highly Automated Vehicles: An Analytical Sensor Model
Authors:

Florian A. Schiegg, Ignacio Llatser and Thomas Michalke
Abstract: Modern advanced driver assistance systems (ADAS) increasingly depend on the information gathered by the vehicle’s on-board sensors about its environment. It is thus of great interest to analyse the performance of these sensor systems and its dependence on macroscopic traffic parameters. The work at hand aims at building up an analytical model to estimate the number of objects contained in a vehicle’s environmental model. It further considers the exchange of vehicle dynamics and sensor data by vehicle-to-vehicle (V2X) communication to enhance the environmental awareness of the single vehicles. Finally, the proposed model is used to quantify the improvement in the environmental model when complementing sensor measurements with V2X communication.
Download
Paper Nr: 84
Title:

Variety-aware Routing Encoding for Efficient Design Space Exploration of Automotive Communication Networks
Authors:

Fedor Smirnov, Behnaz Pourmohseni, Michael Glaß and Jürgen Teich
Abstract: The introduction of sophisticated ADAS has given rise to larger and more complex automotive communication networks whose efficient (in effort) and optimal (in quality) design necessarily depends on automated network design techniques. Typically, these techniques either (a) optimize communication routes based on topology-independent constraint systems that encode the inclusion of each network component in the route of a message or (b) depend on a time- and memory-expensive enumeration of all possible transmission routes to identify the optimal route. In this paper, we propose a novel approach which combines the advantages of these two strategies to enable an efficient exploration of the routing search space: First, the given network is preprocessed to identify so-called proxy areas in which each pair of nodes can be connected by exactly one route. Contrary to network areas with a variety of different routing possibilities, proxy areas do not offer any room for optimization. We propose two approaches—both integrable into existing constraint systems—which exploit the knowledge gathered on proxy areas to improve the exploration efficiency during the routing optimization process. Experimental results for two mainstream topologies of automotive networks give evidence that, compared to state-of-the-art routing optimization approaches, the proposed approaches (a) offer an exploration speedup of up to x 185, (b) deliver network designs of equal or higher quality, and (c) enable an automated design of significantly larger automotive systems.
Download
Short Papers
Paper Nr: 7
Title:

Trajectory Planning for Automated Merging on Highways
Authors:

Johannes Potzy, Nadja Goerigk, Thomas Heil, Dennis Fassbender and Karl-Heinz Siedersberger
Abstract: This article introduces a new approach for trajectory planning for merging on highways. The aim of the algorithm, is to find a comfortable driving strategy to merge in a gap on the target lane. Therfore, the proposed algorithm determines a bunch of trajectories to reach surrounding gaps. The trajectory with the lowest costs for each gap is chosen. To obtain the longitudinal component of the trajectory, a five-part section-wise defined polynomial in Frenet space is used to generate comfortable driving behaviour, with as few changes in the acceleration profile as possible. Based on the prediction of surrounding traffic, different variations of deceleration and acceleration are combined. For each longitudinal part, a lateral component to perform a lane change into the target gap is evaluated. The concept allows to evaluate the influence of the longitudinal driving strategy on the dynamics required to change lanes. The algorithm is evaluated in a MATLAB simulation including a runtime estimation.
Download
Paper Nr: 14
Title:

Identification of the Impact of GNSS Positioning on the Evaluation of Informative Speed Adaptation
Authors:

Jamal Raiyn
Abstract: Autonomous vehicles (AVs) are self-driving vehicles that operate and perform tasks under their own power. They may possess features such as the capacity to sense environment, collect information, and manage communications with other vehicles. Many autonomous vehicles in development use a combination of cameras, various kinds of sensors, GPS, GNSS, radar, and LiDAR, with an on-board computer. These technologies work together to map the vehicle’s position and its proximity to everything around it. To estimate AV positioning, GNSS data are used. However, the quality of raw GNSS observables is affected by a number of factors that originate from satellites, signal propagation, and receivers. The prevailing speed limit is generally obtained by a real-time map matching process that requires positioning data based on a GNSS and a digital map with up to date speed limit information. This paper focuses on the identification of the impact of GNSS positioning error data on the evaluation of informative speed adaptation. It introduces a new methodology for increasing the accuracy and reliability of positioning information, which is based on a position error model. Applying the sensitivity analysis method to informative speed adaptation yields interesting results which show that the performance of informative speed adaptation is positively affected by minimizing positioning error.
Download
Paper Nr: 19
Title:

Using the CGAN Model Extend Encounter Targets Image Training Set
Authors:

Ruolan Zhang and Masao Furusho
Abstract: A fully capable unmanned ship navigation requires full autonomous decision-making, large-scale decision model training data to answer for these conditions is essential. However, it is difficult to obtain enough scenes training data in a real sea navigation environment. In response to possible emergency situations even no shore-station support, this paper proposes a method using conditional generative adversarial networks (CGAN) to generate the most executable large-scale target ships image set, which can be used to training various sea conditions autonomous decision-making model. In practice, most of the current research on unmanned ships are based onshore remote control or monitoring. Nonetheless, in some extremely special circumstances, such as communication interruption, or if the ship cannot be guided or remotely controlled in real time on the shore, the unmanned ship must make an appropriate decision and form new plans according to the encounter targets and the whole current situation. The CGAN model is a novel means to generate the target ships to construct the whole encounter sea scenes situation. The generated targets training image set can be used to train decision models, and explore a new way to approach large-scale, fully autonomous navigation decisions.
Download
Paper Nr: 21
Title:

Towards Certification of Autonomous Driving: Systematic Test Case Generation for a Comprehensive but Economically-Feasible Assessment of Lane Keeping Assist Algorithms
Authors:

Thomas Ponn, Dirk Fratzke, Christian Gnandt and Markus Lienkamp
Abstract: Automation of the driving task continues to progress rapidly. In addition to improving the algorithms, proof of their safety is still an unsolved problem. For an automated driving function that does not require permanent monitoring by the driver, a theoretically infinite number of possible traffic situations must be tested. One promising method to overcome this problem is the scenario-based approach. This approach shall enable an economic certification of automated driving functions with sufficient test space coverage. However, even with this approach, the selection of the scenarios to be tested is still problematic. The first step is to consider a driver assistance system in order to reduce complexity. For the Lane Keeping Assist System under consideration, this paper defines a methodology as well as the scenarios for a comprehensive yet economically-feasible certification. Economical-feasibility of the presented methodology is shown in the results by an approximation of the resulting simulation costs for executing the defined test cases.
Download
Paper Nr: 29
Title:

Conditional Privacy: Users’ Perception of Data Privacy in Autonomous Driving
Authors:

Teresa Brell, Hannah Biermann, Ralf Philipsen and Martina Ziefle
Abstract: Connected autonomous driving can be a key for safety enhancement in road traffic and long-term reduction of driver-induced accidents with personal injury. Though, its acceptance is challenging, because of perceived restrictions on data security and privacy concerns. Hence, in this study, the focus was on users’ perception of data privacy in autonomous and connected driving. For this purpose, a two-tiered research approach was conducted, based on semi-structured interviews (N=7) and an online questionnaire (N=100). Special attention was given to data storage and processing, data distribution, as well as personal usage purposes as predictors for the use intention. Results showed that the driver was most likely accepted as data decision maker, whereas the own car was preferred as data receiver. Besides, evaluation profiles indicated user diverse attitudes concerning the willingness to use autonomous driving regularly. These study outcomes contribute to a deeper understanding of user requirements in the context of mobility acceptance.
Download
Paper Nr: 35
Title:

GCCNet: Global Context Constraint Network for Semantic Segmentation
Authors:

Hyunwoo Kim, Huaiyu Li and Seok-Cheol Kee
Abstract: The state-of-the-art semantic segmentation tasks can be achieved by the variants of the fully convolutional neural networks (FCNs), which consist of the feature encoding and the deconvolution. However, they struggle with missing or inconsistent labels. To alleviate these problems, we utilize the image-level multi-class encoding as the global contextual information. By incorporating object classification into the objective function, we can reduce incorrect pixel-level segmentation. Experimental results show that our algorithm can achieve better performance than other methods on the same level training data volume.
Download
Paper Nr: 48
Title:

A Computationally Efficient MPC for Green Light Optimal Speed Advisory of Highly Automated Vehicles
Authors:

Stephan Uebel, Steffen Kutter, Kevin Hipp and Frank Schrödel
Abstract: The current study introduces an approach for energy efficient longitudinal vehicle guidance. The key idea is to utilize a model predictive control (MPC) for the longitudinal vehicle dynamics which explicitly considers the current and the predicted states of multiple traffic lights ahead. Consequently, the vehicle can drive in urban situations much more energy efficient, which can be used to enlarge the range of electric vehicles or save fuel while additionally improving travel time. Modern traffic lights are equipped with transmitters that send information about their actual and upcoming system states. Additionally, traffic lights connected to a traffic control center can broadcast their future signal phases to vehicles many kilometers ahead. This information may be used to adapt the vehicle speed so that engine operation points are optimal and stops can be avoided. These kind of algorithms are referred to as green light optimal speed advisory. This work presents a novel online capable MPC approach that uses a sequential quadratic program to solve the respective optimal control problem. This approach is implemented in a framework introduced as well which allows driving tests in a real vehicle.
Download
Paper Nr: 49
Title:

Where Am I: Localization and 3D Maps for Autonomous Vehicles
Authors:

Farzeen Munir, Shoaib Azam, Ahmad M. Sheri, YeongMin Ko and Moongu Jeon
Abstract: The nuts and bolts of autonomous driving find its root in devising the localization strategy. Lidar as one of the newest technologies developed in the recent years, provides rich information about the environment in the form of point cloud data which can be used for localization. In this paper, we discuss a localization approach which generates a 3D map from Lidar’s point cloud data using Normal Distribution Transform (NDT) mapping. We use our own dataset collected using our self driving car KIA Soul EV equipped with Lidar and cameras. Once the 3D map has been generated, we have used NDT matching for localizing the self driving car.
Download
Paper Nr: 56
Title:

Driver Cell Phone Usage Violation Detection using License Plate Recognition Camera Images
Authors:

Bensu Alkan, Burak Balci, Alperen Elihos and Yusuf Artan
Abstract: The increased use of digital video and image processing technology has paved the way for extending the traffic enforcement applications to a wider range of violations as well as making the enforcement process more efficient. Automated traffic enforcement has mainly been applied towards speed and red light violations detection. In recent years, there has been an extension to other violation detection tasks such as seat-belt usage, tailgating and toll payment violations. In the recent years, automated driver cell phone usage violation detection methods have aroused considerable interest since it results in higher mortality rates than the intoxicated driving. In this study, we propose a novel automated technique towards driver’s phone usage violation detection using deep learning algorithms. Using an existing license plate recognition camera system placed on an overhead gantry, installed on a highway, real world images are captured during day and night time. We performed experiments using more than 5900 real world images and achieved an overall accuracy of 90.8 % in the driver cell phone usage violation detection task.
Download
Paper Nr: 57
Title:

A LSTM Approach to Detection of Autonomous Vehicle Hijacking
Authors:

Naman S. Negi, Ons Jelassi, Stephan Clemencon and Sebastian Fischmeister
Abstract: In the recent decades, automotive research has been focused on creating a driverless future. Autonomous vehicles are expected to take over tasks which are dull, dirty and dangerous for humans (3Ds of robotization). However, augmented autonomy increases reliance on the robustness of the system. Autonomous vehicle systems are heavily focused on data acquisition in order to perceive the driving environment accurately. In the future, a typical autonomous vehicle data ecosystem will include data from internal sensors, infrastructure, communication with nearby vehicles, and other sources. Physical faults, malicious attacks or a misbehaving vehicle can result in the incorrect perception of the environment, which can in turn lead to task failure or accidents. Anomaly detection is hence expected to play a critical role improving the security and efficiency of autonomous and connected vehicles. Anomaly detection can be defined as a way of identifying unusual or unexpected events and/or measurements. In this paper, we focus on the specific case of malicious attack/hijacking of the system which results in unpredictable evolution of the autonomous vehicle. We use a Long Short-Term Memory (LSTM) network for anomaly/fault detection. It is, first, trained on non-abnormal data to understand the system’s baseline performance and behaviour, monitored through three vehicle control parameters namely velocity, acceleration and jerk. Then, the model is used to predict over a number of future time steps and an alarm is raised as soon as the observed behaviour of the autonomous car significantly deviates from the prediction. The relevance of this approach is supported by numerical experiments based on data produced by an autonomous car simulator, capable of generating attacks on the system.
Download
Paper Nr: 69
Title:

Energy-optimal Speed Trajectories between Stops and Their Parameter Dependence
Authors:

Eduardo F. Mello and Peter H. Bauer
Abstract: This paper addresses the problem of energy-optimal vehicle-speed trajectories between stops. The ideal parameter-dependent trajectory is introduced, and it is shown that it reduces transportation energy drastically relative to “typical trajectories” seen in traffic. The resulting trajectories can easily be implemented in self-driving cars and have the potential to significantly reduce transportation energy in networked vehicles and cities. The usage of this energy-optimal speed trajectories between stops can save significant amounts of energy, sometimes in excess of 30% when comparing to conventional traffic flow speed profiles. This paper also addresses the impact that vehicle and segment parameters have on the savings. The role of parameters such as the air drag coefficient, cross-sectional area, vehicle mass, efficiency, segment length, average speed, as well as acceleration capability are investigated. It is shown that optimizing speed trajectories to minimize transportation energy consistently results in energy savings. However, diminishing returns are observed for certain scenarios, such as long, low-speed segments.
Download
Paper Nr: 72
Title:

Quantifying Impacts of Connected and Autonomous Vehicles on Traffic Operation using Micro-simulation in Dubai, UAE
Authors:

Abdul R. Alozi and Khaled Hamad
Abstract: Connected and Autonomous Vehicles (CAVs) will change the transportation system we know with their substantial impacts on the level of safety, traffic operation, fuel consumption, air emissions among other aspects. A large segment of the general public and decision makers are still sceptical of CAVs’ benefits and impacts. This study aims at quantifying the impacts of CAVs on traffic operation using micro-simulation of a 7-kilometer-freeway segment in Dubai, UAE. The simulation was run for different market penetration rates (MPRs) ranging from 0% (no CAVs) up to 100% (all CAVs), in 10% increment. Additionally, multiple scenarios under different traffic volumes were also modelled utilizing PTV VISSIM. To quantify the impacts of CAVs, three performance measures were collected, namely the average delay, average speed, and total travel time. The results showed that the highest impact of CAVs occurs in terms of delay, with a decreased average delay of up to 86%. The other performance measures also show improvement, with 42% speed increase and 25% travel time reduction. Moreover, CAVs show more significant changes at lower traffic volume conditions (off-peak hour).
Download
Paper Nr: 78
Title:

Development of 8x8 All-terrain Vehicle with Individual Wheel Drive
Authors:

Alexander Belyaev, Sergey Manyanin, Anton Tumasov, Vladimir Makarov and Vladimir Belyakov
Abstract: In this article, we consider the problem of developing a rational competitive design of a multifunctional all-terrain vehicle (MATV) with 8х8 axle configuration. Empirical dependencies are proposed to calculate weight-size parameters of these vehicles, such as power and power-to-weight ratio, payload, maximum speed, average ground pressure depending on full vehicle weight. Key dependencies are provided to calculate hydrostatic transmission (HST) parameters used to determine hydraulic unit sizes and connection diagrams. Various HST control algorithms are analyzed in order to increase efficiency and reduce fuel consumption. The results show that the right HST control algorithm can increase efficiency by 10%, and reduce fuel consumption by 18%. General view of the developed MATV is provided.
Download
Paper Nr: 79
Title:

Using an Intelligent Vision System for Obstacle Detection in Winter Condition
Authors:

Marwa Ziadia, Sousso Kelouwani, Ali Amamou, Yves Dubé and Kodjo Agbossou
Abstract: This paper explores the performance of an Advanced Driving Assistance System (ADAS) during navigation in urban traffic and a winter condition. The selected ADAS technology, Mobileye, has been integrated into a hydrogen electric vehicle. A set of three cameras (visible spectrum) has also been installed to give a surrounding view of the test vehicle. The tests were carried out during the dusk as well as in the night in winter condition. Using Matlab, the messages provided by Mobileye system have been analyzed. More than 2800 samples (short sequences of 5s Mobileye messages) have been processed and compared with the corresponding video samples recorded by the three cameras. In average, the selected ADAS device was able to provide 99% of true positive vehicle detection and classification, even in poor ambient lighting condition in winter. However, 72% of samples involving a pedestrian was correctly classified.
Download
Paper Nr: 90
Title:

Optimal Path Planning with Clothoid Curves for Passenger Comfort
Authors:

Edward D. Lambert, Richard Romano and David Watling
Abstract: Highly automated vehicles operating at SAE automation level 4 and 5 will not require the occupants’ attention to be on the road at all. They will be free to amuse themselves as passengers. This will have the side effect of making them more vulnerable to motion sickness. Automated vehicles must plan paths which are feasible for the vehicle and comfortable for its occupants. In railway and highway design, paths with clothoid based transitions provide feasibility and comfort. This paper proposes a method for generating such a path using constrained non-linear optimization and compares it to an existing method based on root finding.
Download
Paper Nr: 4
Title:

Usage of GPS Data for Real-time Public Transport Location Visualisation
Authors:

Aleksejs Zacepins, Egons Kalnins, Armands Kviesis and Vitalijs Komasilovs
Abstract: The concept of the smart city has been fashionable in the political arena in recent years. Cities are trying to be modern and provide various ICT based services for their citizens. An efficient public transportation service is critical for the citizens, but traffic congestions are still a problem in cities and are one of the main reasons for public transport delays. Therefore, it is important for citizens to know where the needed public transport vehicle is located at the moment, to know if the transport has already passed the stop or not. Authors of this research propose a real-time public transport tracking system using a global positioning system (GPS) technology module to receive the location of the vehicle in a real-time. System is based on the Raspberry Pi 3, which is used to transfer positioning data received from GPS module to the remote database. Based on received data, the location of the bus is visualised in the developed Web system.
Download
Paper Nr: 13
Title:

Vision based ADAS for Forward Vehicle Detection using Convolutional Neural Networks and Motion Tracking
Authors:

Chen-Wei Lai, Huei-Yung Lin and Wen-Lung Tai
Abstract: With the rapid development of advanced driving assistance technologies, from the very beginning of parking assistance, lane departure warning, forward collision warning, to active distance control cruise, the active safety protection of vehicles has gained the popularity in recent years. However, there are several important issues in the image based forward collision warning systems. If the characteristics of vehicles are defined manually for detection, we need to consider various conditions to set the threshold to fit a variety of the environment change. Although the state-of-art machine learning methods can provide more accurate results then ever, the required computation cost is far much higher. In order to find a balance between these two approaches, we present a detection-tracking technique for forward collision warning. The motion tracking algorithm is built on top of the convolutional neural networks for vehicle detection. For all processed image frames, the ratio between detection and tracking is well adjusted to achieve a good performance with an accuracy/computation trade-off. Th experiments with real-time results are presented with a GPU computing platform.
Download
Paper Nr: 17
Title:

3D Car Tracking using Fused Data in Traffic Scenes for Autonomous Vehicle
Authors:

Can Chen, Luca Z. Fragonara and Antonios Tsourdos
Abstract: Car tracking in a traffic environment is a crucial task for the autonomous vehicle. Through tracking, a self-driving car is capable of predicting each car’s motion and trajectory in the traffic scene, which is one of the key components for traffic scene understanding. Currently, 2D vision-based object tracking is still the most popular method, however, multiple sensory data (e.g. cameras, Lidar, Radar) can provide more information (geometric and color features) about surroundings and show significant advantages for tracking. We present a 3D car tracking method that combines more data from different sensors (cameras, Lidar, GPS/IMU) to track static and dynamic cars in a 3D bounding box. Fed by the images and 3D point cloud, a 3D car detector and the spatial transform module are firstly applied to estimate current location, dimensions, and orientation of each surrounding car in each frame in the 3D world coordinate system, followed by a 3D Kalman filter to predict the location, dimensions, orientation and velocity for each corresponding car in the next time. The predictions from Kalman filtering are used for re-identifying previously detected cars in the next frame using the Hungarian algorithm. We conduct experiments on the KITTI benchmark to evaluate tracking performance and the effectiveness of our method.
Download
Paper Nr: 18
Title:

Continuously Improving Model of Road User Movement Patterns using Recurrent Neural Networks at Intersections with Connected Sensors
Authors:

Julian Bock, Philipp Nolte and Lutz Eckstein
Abstract: Intersections with connected infrastructure and vehicle sensors allow observing vulnerable road users (VRU) longer and with less occlusion than from a moving vehicle. Furthermore, the connected sensors are providing continuous measurements of VRUs at the intersection. Thus, we propose a data-driven prediction model, which benefits of the continuous, local measurements. While most approaches in literature use the most probable path to predict road users, it does not represent the uncertainty in prediction and multiple maneuver options. We propose the use of Recurrent Neural Networks fed with measured trajectories and a variety of contextual information to output the prediction in a local occupancy grid map in polar coordinates. By using polar coordinates, a reliable movement model is learned as base model being insensitive against blind spots in the data. The model is further improved by considering input features containing information about the static and dynamic environment as well as local movement statistics. The model successfully predicts multiple movement options represented in a polar grid map. Besides, the model can continuously improve the prediction accuracy without re-training by updating local movement statistics. Finally, the trained model is providing reliable predictions if applied on a different intersection without data from this intersection.
Download
Paper Nr: 30
Title:

Differences in Driver Behaviour between Race and Experienced Drivers: A Driving Simulator Study
Authors:

Naman S. Negi, Peter Van Leeuwen and Riender Happee
Abstract: Safety is one of the major areas of concerns today in the field of automotive development. Different safety measures have and are being introduced in order to improve driver/passenger and pedestrian safety. Advanced driver assist systems (ADAS) are therefore becoming increasingly important in their role of reducing driver crash risk. A shortcoming of the ADAS systems is that the variability in drivers based on skill and experience is not taken into account and the system is often designed for average or worst case driver performance thereby compromising on the dynamic behaviour of the vehicle. This study focuses on understanding and quantifying the differences in drivers. This knowledge of driver differences can be helpful in designing an adaptive ADAS by introducing the driver into the control loop. The study investigates differences between race-car drivers and normal (experienced) drivers in a high-speed driving task. The study analyses simulator data for 17 drivers on the Mallory Park test circuit. The driving task required the participants to drive around the circuit to achieve the fastest lap times. Analysis showed that higher steering activity and differences in path strategy were the main reasons for lower lap-times shown by the expert race drivers compared to the non-expert drivers. Steering metrics like average steering rate, steering jerk showed higher values for the expert group and distance traveled around the corner showed a different path strategy adopted by the experts. Both groups showed improvement in performance based on lap-times across the different sessions. Thus the study shows that expert and non-expert drivers have different steering behaviour and path strategy, which can be attributed to differences in driving experience, vehicle dynamics knowledge and vehicle control skills.
Download
Paper Nr: 31
Title:

Differences in Driver Behaviour between Novice and Experienced Drivers: A Driving Simulator Study
Authors:

Naman S. Negi, Peter Van Leeuwen and Riender Happee
Abstract: This study is an extension of a previous work where differences between race-car drivers and normal drivers has been investigated in a high-speed driving task. The study focused on gaining knowledge about driver differences that can be helpful in designing an adaptive ADAS by introducing the driver into the control loop. The present study takes this research forward and is oriented around finding the differences between novice and normal (experienced) drivers while performing a double lane change maneuver and a high-speed cornering task. The study aimed at finding parameters capable of differentiating the two groups with special emphasis on steering behaviour. Part A of the test procedure required the participants to complete a double lane change at various speeds (from 70km/h to 105km/h). Data analysis showed that late initial steering input given by the novices compared to the experienced drivers was the main reason for their poor performance. Steering metrics like timing of steering input, average steering rate and average steering jerk showed statistically significant differences between the two groups. Part B of the experiment required the participants to drive around a flat oval track to achieve the fastest lap times. Analysis showed that higher steering activity and differences in path strategy were the main reasons for lower lap-times shown by the experienced drivers compared to the novice drivers. Steering metrics like average steering rate, steering jerk showed higher values for the experienced group.
Paper Nr: 42
Title:

Automated Bicycle Counting System’s Prototype to Evaluate the Necessity of New Bicycle Lanes in Jelgava City
Authors:

Armands Kviesis, Aleksejs Zacepins, Vitalijs Komasilovs, Normunds Vetra and Nikolajs Bumanis
Abstract: Every year the number of vehicles on the road is increasing. But there are people that start to choose public transport or greener transportation options like bicycles or electric cars over typical fuel cars. Regarding bicycle usage, the problems that arise are related to insufficient bicycle lanes and determination of future lane locations, so that the resources used for bicycle lane construction would be properly invested. To resolve mentioned problems municipality first need to learn where the suitable bicycle lane location for cyclists should be. Such a task can be done by applying a cyclist counting system. This paper describes a portable automatic cyclist counting system’s prototype for bicycle lane location planning and also identifies the limitations for such a system. Proposed prototype is based on rubber tubes and pressure sensors, Wi-Fi module and open source electronic platform Arduino. This study is carried out within the ERANet-LAC project RETRACT (Enabling resilient urban transportation systems in smart cities).
Download
Paper Nr: 53
Title:

Logical Scenario Derivation by Clustering Dynamic-Length-Segments Extracted from Real-World-Driving-Data
Authors:

Jacob Langner, Hannes Grolig, Stefan Otten, Marc Holzäpfel and Eric Sax
Abstract: For the development of Advanced Driver Assistant Systems (ADAS) and Automated Driving Systems (ADS) a change from test case-based testing towards scenario-based testing can be observed. Based on current approaches to define scenarios and their inherent problems, we identify the need to extract scenarios including the static environment from recorded real-world-driving-data. We present an approach, that solves the problem to extract dynamic-length-segments containing a single scenario. These segments are enriched with a feature vector with information relevant for the feature under test. By clustering these scenarios a logical scenario catalog is created, containing all scenarios within the test data. Corner cases are represented as well as common scenarios. An accumulated total length can be calculated for each logical scenario, giving a brief understanding about existing test coverage of the scenario.
Download
Paper Nr: 58
Title:

Parking Occupancy Detection using Thermal Camera
Authors:

Vijay Paidi and Hasan Fleyeh
Abstract: Parking a vehicle is a daunting task during peak hours. The search for a parking space leads to congestion and increased air pollution. Information of a vacant parking space would facilitate to reduce congestion and subsequent air pollution. This paper aims to identify parking occupancy in an open parking lot which consists of free parking spaces using a thermal camera. A thermal camera is capable of detecting vehicles in any weather and light conditions based on emitted heat and it can also be installed in public places with less restrictions. However, a thermal camera is expensive compared to a colour camera. A thermal camera can detect vehicles based on the emitted heat without any illumination. Vehicles appear bright or dark based on heat emitted by the vehicles. In order to identify vehicles, pre-trained vehicle detection algorithms, Histogram of Oriented Gradient detectors, Faster Regional Convolutional Neural Network (FRCNN) and modified Faster RCNN deep learning networks were implemented in this paper. The detection rates of the detectors reduced with diminishing of heat in the vehicles. Modified Faster RCNN deep learning network produced better detection results compared to other detectors. However, the detection rates can further be improved with larger and diverse training dataset.
Download
Paper Nr: 59
Title:

Smart Parking Assistance Services and User Acceptance: A European Model
Authors:

Eleni G. Mantouka, Foteini Orfanou, Martin Margreiter, Eleni I. Vlahogianni and Javier S. Medina
Abstract: Technologies and systems assisting drivers to locate free on street parking space and/or inform on parking availability may significantly reduce the traffic induced from cruising for parking space in cities. This paper attempts to reveal the factors that may affect the acceptability of parking assistance systems in different European cities, based on data collected through a questionnaire survey. The respondents are presented with a real world parking assistance system based on in-vehicle ultrasonic sensors, which detects free parking space in real time, and are, then, asked to respond to a set of questions in relation to their parking choice preferences. The results of the survey are presented and modelled using a genetically optimized Logistic Regression Model. Findings indicate that the proposed system would be useful for people who are not willing to spend too much time in order to find an available parking space as well as to those who are not willing to walk long distances from the parking place to their final destination. Moreover, results revealed that the certainty of the provided recommendation significantly influences the effect of the other parameters on the acceptability of the application. Finally, some further research steps are discussed.
Download
Paper Nr: 74
Title:

Evaluating Mentalization during Driving
Authors:

Giorgio Grasso, Chiara Lucifora, Pietro Perconti and Alessio Plebe
Abstract: The development of artificial intelligence promises important future changes from a social point of view. In particular, the emerging self-driving cars allow today to plan a future where traffic flow will greatly improve, and car accidents will be continuously decreasing. However, we should expect a period when full or partial autonomous vehicles and ordinary cars coexist, during which it would be essential to fully understand the cognitive processes used by ordinary people when driving. We identify as a crucial aspect the shift between quick and automated reactions, and the resort to mentalizing, costly social processes, sometimes necessary to predict intentions of other road users. In our experimental design we investigate the main precursors of mindreading, that is, eye contact and shared attention. We believe that a better understanding of this twofold mecahnisms involved in driving could be used to improve advanced driver assistance systems.
Download
Paper Nr: 75
Title:

Implementation of Autonomous Driving Vehicle at an Intersection with Traffic Light Recognition and Vehicle Controls
Authors:

Changhyeon Park and Seok-Cheol Kee
Abstract: We implemented autonomous driving vehicle system at an intersection equipped with traffic lights. This system was consisted of a traffic light recognition, crossing/stop decision algorithm, vehicle localization, vehicle longitudinal/lateral control, and coordinate map generation. The traffic light recognition was implemented by using camera-based CNN data processing. The crossing/stop decision algorithm decides vehicle longitudinal control whether drive or not depending on recognized traffic light signal. The vehicle localization was implemented by using RTK GNSS and dead reckoning. The longitudinal control was designed by planned path data and the lateral control was designed by processed planned path data and traffic light position/signal recognition results. The overall vehicle control system was implemented based on an embedded control board. Coordinate map was made by saving vehicle’s position data received from RTK GNSS. To evaluate the performance of proposed system, we remodeled a commercial vehicle into autonomous driving vehicle and drove the vehicle on our proving ground. Our own proving ground for the test vehicle driving performance was located in Ochang Campus, Chungbuk National University. As a result, the proposed vehicle successfully drove at intersection equipped with traffic lights with a maximum speed of 40 kph on a straight course and a maximum speed of 10 kph on a 90c̊corner course.
Download

Area 4 - Data Analytics

Full Papers
Paper Nr: 50
Title:

Real-world Test Drive Vehicle Data Management System for Validation of Automated Driving Systems
Authors:

Lars Klitzke, Carsten Koch, Andreas Haja and Frank Köster
Abstract: For the validation of autonomous driving systems, a scenario-based assessment approach seems to be widely accepted. However, to verify the functionality of driving functions using a scenario-based approach, all scenarios that may be relevant for the validation have to be identified. Real-world test drives are mandatory to find relevant and critical scenarios. However, the identification of scenarios and the management of the captured data requires computational assistance to validate driving functions with reasonable effort. Therefore, this work proposes a highly-modularised multi-layer Vehicle Data Management System to manage and support analysing large-scale test campaigns for the scenario-based validation of automated driving functions. The system is capable of aggregating the vehicle sensor data to time-series of scenes by utilising temporal discretisation. Those scenes will be enriched with information from various external sources, providing the foundation for efficient scenario mining. The practical usefulness of the proposed system is demonstrated using a real-world test drive sequence, by finding lane-change scenarios and evaluating an onboard system.
Download
Paper Nr: 54
Title:

Prediction of Bike Mobility in Cascais’s Sharing System
Authors:

Nuno Oliveira, Maricica Nistor and André Dias
Abstract: Bike sharing systems offer a convenient, ecologic, and economic transport mode that has been increasingly adopted. However, the distribution of bikes is often unbalanced, which decreases user satisfaction and potential revenues. Moreover, bike sharing literature is mostly focused on the prediction of demand on large scale systems and uses simulations for the assessment of relocation operations to increase the number of utilizations. We propose prediction models based on machine learning approaches to improve the bike sharing re-balancing in a small city of Portugal. The algorithm aims to improve three metrics, namely (1) increase the number of utilizations, (2) reduce the number of stations without bikes, (3) reduce the time without available bikes in the stations. The relocation operations are validated using real data. Our findings show that (a) the estimated number of utilizations created by this system is substantially higher than the current system by 223%, (b) our model allows the correct identification of more 70%, 165%, 249% empty stations with the same or substantially higher precision than the existing approach, (c) the total time of bike unavailability reduced by the predictive model is 283% higher than the time reduced by current approach (1,394,454 vs 363,971 minutes).
Download
Paper Nr: 83
Title:

On the Evaluation of a Cluster-based Reputation Assessment Mechanism for Carpooling Applications
Authors:

Emmanouil Mastorakis, Athanasios I. Salamanis, Dionysios D. Kehagias and Dimitrios Tzovaras
Abstract: Carpooling is a mobility concept that appears to be the answer when it comes to challenges in urban mobility derived by population growth. In carpooling, the same amount of people move with fewer vehicles leading to reduced traffic congestion and consequently to less CO2 emissions, fuel consumption and drivers frustration. However, there has always been scepticism around carpooling due to the inherent mistrust between drivers and passengers. In recent years, some reputation systems have been proposed to reduce the impact of mistrust on carpooling applications. Among them, the work of Salamanis et al. (Salamanis, 2018), in which a reputation assessment mechanism based on clustering users travel preferences, was introduced. In this paper, we provide an extended version of the previous mechanism and we thoroughly evaluate its robustness in relation with different types of malicious attacks and clustering algorithms. In addition, we compare our mechanism with a benchmarking reputation system that utilizes the simple arithmetic mean to calculate reputation values based on users ratings. The evaluation results indicate that the extended reputation assessment mechanism exhibits more robust behavior compared to the benchmarking system in all types of attacks when using the hierarchical clustering algorithm.
Download
Paper Nr: 91
Title:

Optimal Driving Profiles in Railway Systems based on Data Envelopment Analysis
Authors:

Achilleas Achilleos, Markos Anastasopoulos, Anna Tzanakaki, Marius Iordache, Olivier Langlois, Jean-Francois Pheulpin and Dimitra Simeonidou
Abstract: The present study focuses on the development of a dynamically re-configurable Information Communication Technology (ICT) infrastructure to support the sustainable development of railway network. Once data have been collected, the extracted knowledge is used to develop a set of applications that can improve the energy efficient operation of railway systems. A typical example includes the identification of the optimal driving profiles in terms of energy consumption. In the present study, this is achieved through the adoption of an optimization framework based on Data Envelopment Analysis (DEA). The performance of the proposed scheme is evaluated based on actual data collected at an operation tramway system. Preliminary results illustrate that when the proposed method is applied, a 10% reduction in the overall power consumption can be achieved.
Download
Short Papers
Paper Nr: 47
Title:

Training and Validation Methodology for Range Estimation Algorithms
Authors:

Patrick Petersen, Adam T. Thorgeirsson, Stefan Scheubner, Stefan Otten, Frank Gauterin and Eric Sax
Abstract: Estimating the range of battery electric vehicles is one of the most challenging topics for the current trend in the automotive industry, the electrification of vehicles. Range anxiety still limits the adoption of battery electric vehicles. Since the range estimation is dependent on different influencing factors, complex algorithms to accurately estimate the vehicles consumption are required. To evaluate the accuracy of data-driven machine learning algorithms, an exhaustive training and validation procedure is mandatory. In this paper, we propose a novel methodology for the development and validation of range estimation algorithms based on machine learning validation approaches. The proposed methodology considers the evaluation of driver-specific and driver-unspecific performance. In addition, an error measure is introduced to assess the performance of range estimation algorithms. This approach is demonstrated and evaluated on a set of recorded real-world driving data. It is shown that our approach helps to analyze the performance of the range estimation algorithm and the influences of different parameter sets.
Download

Area 5 - Sustainable Transport

Full Papers
Paper Nr: 6
Title:

A Novel Approach for Development of Neural Network based Electrical Machine Models for HEV System-level Design Optimization
Authors:

Christian Gletter, Andre Mayer, Josef Kallo, Thomas Winsel and Oliver Nelles
Abstract: To find the optimal system-level design of hybrid electric vehicles (HEVs), component models are used in simulations to evaluate a large number of different designs within a high dimensional design space. As the electrical machine (EM) represents a key component of the HEV powertrain in terms of energy consumption, models require scalability and sufficient accuracy with manageable computational effort. This paper presents a novel approach for the development of scalable EM models based on Neural Networks (NN). The models are trained with data derived by a Finite Element Analysis (FEA) based scaling procedure and capable to represent the characteristics of a wide range of EM designs without the incorporation of further details. Once a model is trained, it can be directly used in system-level design optimization. The practicality of the model is proven within an exemplary simulation study and its goodness of fit to the training data is validated by a statistical analysis. This approach can help to reduce the computational effort of EM efficiency maps calculation, since only a small number of time-consuming FEA based scaling simulations must be performed prior to the optimization.
Download