Workshops

IV2017The28thIEEE

June11-14,RedondoBeach,CA,USA

W1-Workshop1CognitivelyInspiredIntelligentVehicles(9:00–18:00)Room: Redondo 1 Organizer(s): Serge Thill, Interaction Lab, School of Informatics, University of Skövde, Sweden Abstract: The initial focus of a research field interested in the development of autonomous systems is often on algorithms designed from a pure engineering perspective. Research on intelligent vehicles, for example, is currently in this stage. At the same time, there are other fields interested in autonomous systems that have been in existence for longer. One clear example is research on (cognitive) robotics and artificial cognitive systems. A lesson that can be learned from these fields is that the initial engineering-centric methods are eventually supplemented with inspiration from the cognitive sciences. These supplements bring new ways to increase the autonomy of a system, to ensure its ability to deal with events that are not foreseeable at design time, and sometimes even to ensure behavior that is intuitively understandable by humans that interact with these systems. Conversely, an autonomous system (whether a robot or a vehicle) will necessarily have to interact with humans (in the case of intelligent vehicles, these include the passengers inside the vehicle, pedestrians, other vulnerable road users, and drivers of vehicles that are not automated). The vehicle must therefore also be able to understand and predict the actions of others. Overall, such cognitively inspired approaches are now well established in robotics, both at the control level, and where interaction with human users is concerned, but are only beginning to emerge in intelligent vehicle development. The purpose of the present workshop is therefore to give a forum to researchers who either apply cognitive approaches to intelligent vehicles, or have made major contributions to robotics in this manner. The content of this workshop is thus relevant to anyone interested in one of the major under researched areas in the field of intelligent vehicles, now ripe for exploitation. Program 09:00 – 09:10 Organizer’s introduction 09:10 – 10:00 Keynote: Cognitively Inspired Robotics 10:00 – 10:30 Coffee break 10:30 – 12:30 Invited papers and discussion 12:30 – 13:30 Lunch break 13:30 – 14:30 Keynote: Cognitively Inspired User Modelling 14:30 – 15:00 Submitted papers 15:00 – 15:30 Coffee break 15:30 – 16:30 Submitted papers 16:30 – 18:00 Roundtable discussion and conclusions

W2-Workshop 2: Automotive Cybersecurity 2017 (13:30 – 17:30) Room: Marina Organizer(s): Madhusudan Singh and Shiho Kim, YICT-Yonsei University, Seoul, South Korea Abstract: Nowadays intelligent vehicle is an emerging research topic going on around us. The researcher is working on to make intelligent and autonomous vehicles, and transportation system. The more our vehicles become intelligent, the more we need to work on safety and security for vehicles technology. The workshop aims to foster discussion on automotive cybersecurity related problem and solutions and security standardization for not only intelligent vehicles but also discuss security for Infrastructure of Intelligent transportation system. We need to know what type of security requirements needed during design and development of intelligent vehicles. Therefore, it is aimed to investigate security, standardization, security method and process for vehicle communication. The primary objective of this proposed workshop will also discuss vehicle-communication security for the vehicle-to-vehicle (V2V) communication and Vehicle to Infrastructure (V2X) to work with intelligent vehicles and intelligent transportation environment and concentrate on research work for, Cryptography, Embedded security, hardware security for intelligent vehicles. The topics of interest of this workshop cover, but is not limited to, the following scope: Automotive Security threats to cyber-physical systems; Security architecture, design, implementation, and management of intelligent vehicles; Security Techniques and protocols for cooperative vehicles; Data communication security in networked embedded systems; Automotive collision prediction and avoidance in cyber world; Security mechanism for automotive motion planning in dynamic environments; Practical security experiences and testbeds related with intelligent vehicles; Automotive industrial experiences and best practices relevant to safety and security of IV. Program

13:30 – 13:45 An Overview of Automotive Cyber-Security Workshop (AutoCys), Shiho Kim (Yonsei University, South Korea)

13:45 – 14:30 Automotive Cyber Security: Perspective, Challenges and Future Trend, Madhusudan Singh (Yonsei University, South Korea)

14:30 – 15:15 Automotive Technology for Vehicle Safety and Security, Dhananjay Singh (Hankuk University of Foreign Studies, South Korea)

15:15 – 15:30 Coffee break

15:30 – 16:15 Blockchain Technology for automotive cybersecurity, Madhusudan Singh (Yonsei University, South Korea)

16:15 – 16:45 A Secure Cyber Physical Surveillance System for Internet of Vehicles, Antonio Jara (HES-SO, Switzerland)

16:45 – 17:30 Panel Discussion: Cyber-Security Impact on Intelligent Vehicles, Madhusudan Singh; Shiho Kim (Yonsei University, Seoul, South Korea)

W3-Workshop 3: Workshop on Machine Vision and Interfaces in Data Fusion Platforms for Automated Driving (10:00 – 16:30) Room: Redondo 2 Organizer(s): Michael Schilling, HELLA KGaA Hueck & Co.; Cristobal Curio, Reutlingen University Abstract: While the path to series production of automated driving cars still faces many challenges we will focus in this workshop on two topics. One main challenge in building automated driving cars today is the huge variety of interfaces to the environmental fusion system. The current diversity in technologies (e.g. lidar/radar/camera sensors, V2X, maps) and suppliers, spawn a myriad of interfaces which does not allow a quick or easy integration into any data fusion platform. Furthermore, the deployed interfaces of the environmental fusion systems for a specific driving function (e.g. from intersection assist to valet parking) are also not standardized in any way today. When we look at the next step by implementing distributed data fusion in order to get redundancy and enhanced fusion algorithms, the interface challenges will become even more complex. The second topic our workshop will focus on, are the challenges in using high resolution cameras. Human drivers base most of their decision solely on their visual system. Current infrastructure offers rich visual cues to guarantee safe and flexible driving. It seems that today’s semantic vision in combination with high resolution cameras can interpret the richness of the environment very much like the human visual system does and thus will play a big role in future automated driving vehicles. Yet todays systems still face challenges, ranging from adverse weather conditions over pedestrian movement estimation for enhanced predictions to implementing high data stream interfaces to a central fusion platform. Further, high resolution image data also offers the opportunity to produce highly sophisticated simulations, useful e.g. for function optimization and testing purposes. Program

10:00 – 10:30 Workshop Introduction + Interfaces in an Open Fusion Platform, Michael Schilling (HELLA)

10:30 – 11:00 Requirements for Functional Architectures and Interfaces in Centralized Multi-Sensor Systems for Automated Driving, Thomas Dammeier (HELLA Aglaia)

11:00 – 11:30 Coffee break

11:30 – 12:00 Sensor Fusing and Road User Recognition for the WEpods Driverless Shuttle, Floris Gaisser; J. Domhof; R. Happee; P. Jonker (TU Delft)

12:00 – 12:30 Research & Development ADAS Activities at HRI-Europe, Christian Görick (Honda Research Institute)

12:30 – 14:00 Lunch break

14:00 – 14:30 Distributed Data Fusion - A framework for Information Integration in V2X systems, Daniel Clarke (Cranfield University)

14:30 – 15:00 Improving Vehicle Localization Using Semantic and Pole-Like Landmarks, Mohsen Sefati; M. Daum; B. Sondermann; K.D. Kreisköther; A. Kampker (RWTH Aachen)

15:00 – 15:30 Coffee break 15:30 – 16:00 Rudolf Mester (Linköping University) 16:00 – 16:30 Discussion round with all speakers

W4-Workshop 4: Deep Learning for Vehicle Perception (9:00 – 17:00) Room: Redondo 3 Organizer(s): José M. Alvarez; Uwe Franke; Trevor Darrell, NICTA Abstract: Deep learning, and in particular Convolutional Neural Networks, has become the main component of many intelligent vehicle algorithms. Jointly with the explosive growth in the available amount of driving data these data-driven algorithms have certainly enabled the next generation of platforms for reliable autonomous driving as evidenced by the algorithms used by a large number of companies like MobilEye, AutoX, Zoox, Toyota Research, General Motors, Volkswagen, and Daimler among many others. The goal of this workshop on deep learning for vehicle perception (as a second edition of the Deep-Driving workshop held in conjunction with IV2016, follow {http://iv2016.berkeleyvision.org/} for further information) is to foster discussion and to accelerate the study of deep architectures in autonomous driving problems with a focus on the efficiency of the algorithms. Detailed program to be available. Website of Workshop: www.deep-driving.net

W5-Workshop 5: Multi-Sensor Fusion and Extended Object Tracking (9:00 – 17:00) Room: Bay Organizer(s): Stephan Reuter, University of Ulm, Germany; Karl Granström, Chalmers University, Sweden; Marcus Baum, Georg-August-University Göttingen, Germany Abstract: The aim of this workshop is to provide an overview of recent approaches for environment perception based on occupancy grid maps and multi-object tracking and to foster discussions about multi-sensor fusion approaches as well as the reliability of environment perception. Further, the audience is introduced to extended object tracking, with a focus on both the underlying theory and relevant real world applications. The modelling of object shapes and measurements in extended object tracking algorithms are introduced in detail and the integration in multi-object tracking algorithms is outlined. Additionally, several applications of extended object tracking and the multi-sensor fusion approaches are presented by the organizers and invited presenters. Program

09:00 – 10:00 Bayesian Filtering and Modeling of Extended Objects, Baum, Marcus (Georg-August-University Göttingen, Germany)

10:00 – 11:00 Multi-Object Tracking and Sensor Fusion Using Random Finite Sets, Reuter, Stephan (University of Ulm, Germany)

11:00 – 12:00 Multiple Extended Object Tracking, Granström, Karl (Chalmers University, Sweden)

12:00 – 13:30 Lunch break

13:30 – 14:00

GM-PHD Filter for Multiple Extended Object Tracking Based on the Multiplicative Error Shape Model and Network Flow Labeling, Teich, Florian; Yang, Shishan; Baum, Marcus (Georg-August-University Göttingen, Germany)

14:00 – 14:40 Dynamic Occupancy Grid Maps for Autonomous Vehicles, Nuss, Dominik (Mercedes-Benz Research and Development North America, USA)

14:40 – 15:00 Coffee break

15:00 – 15:40 Sensor Fusion Framework for Extended Object Tracking, Svensson Daniel (Zenuity)

15:40 – 16:20 Tracking Objects with Changing Appearance, Vatavu, Andrei (Mercedes-Benz Research and Development North America, USA)

16:30 – 17:00 Summary and discussion round with all speakers

W6-Workshop 6: The 4th Workshop on Naturalistic Driving Data Analytics (9:30 – 17:00) Room: Catalina Organizer(s): Takashi Bando, Denso International America, Inc.; Chiyomi Miyajima, Nagoya University; Pujitha Gunaratne, Toyota Motor Engineering and Mnf. North America Abstract: This workshop on Naturalistic Driving Data Analytics is proposed for the fourth time now, to be collocated at IEEE Intelligent Vehicles Symposium (IV) in 2017. The purpose is two folds, to foster discussions on challenges related to developing and using the right methods to make meaningful inferences and interpretations from large scale unlabeled naturalistic driving data to and to exchange ideas on how to optimally use naturalistic driving data for different purposes, such as driver behavior analysis, further development of ITS technologies, as well as for automated driving. Topics of Interest are

• Data Collection o Naturalistic driving data collection & sharing o Robust data compression, reduction, and annotation for large-scale data

• Data Analytics o Driver behavior analysis o Image processing methods for naturalistic driving videos o Machine learning methods for naturalistic driving data

• NDD applications o Human Machine Interface / Interaction o Advanced Driver Assistance Systems / Automated Driving o Fuel efficiency o Traffic Management Systems

Program

9:30 – 10:00 Trajectory Data Collection for Scene-Aware Driving Behavior Study: An NDDA Activity in Beijing, Huijing Zhao (Peking University)

10:00 – 10:30 Understanding the Performance and Limitations of Advanced Driver State Monitoring System through Naturalistic Driving Experiments, Renran Tian; Yaobin Chen (Indiana University-Purdue University Indianapolis)

10:30 – 11:00 Coffee break 11:00 – 11:30 Dr. Ravi Satzoda (Nauto)

11:30 – 12:00 Data-Driven Motion Planning for Self-Driving Vehicle, Takashi Bando (DENSO International America, Inc.)

12:00 – 12:30 Luke Fletcher (Toyota Research Institute) 12:30 – 14:00 Lunch break (with demonstration)

14:00 – 15:00 Semantic Segmentation of Driving Behavior Data: Double Articulation Analyzer and its Application, Tadahiro Taniguchi (Ritsumeikan University)

15:15 – 16:15 Hands on: How to analyze large-scale driving data

16:30 – 17:00 Panel discussion: How to Effectively Share the Naturalistic Driving Data - from the Results of FOT-NET

W7-Workshop 7: Vision for Intelligent Vehicles and Application-Surround 2017, Workshop & Challenges (13:30 – 17:30) Room: Esplanade Organizer(s): Akshay Rangesh, UC San Diego; Jacob Dueholm, UC San Diego and Aalborg University; Miklas Kristoffersen, UC San Diego and Aalborg University; Rakesh Rajaram, UC San Diego; Eshed Ohn-Bar, UC San Diego; Ravi Satzoda, UC San Diego; Mohan Trivedi, UC San Diego Abstract: The workshop will cover topics relevant to vision-based analysis of surround vehicles from RGB video. Specifically, an extensive U.S. highway dataset collected with 4 perspectives for complete panoramic surround will be presented and made public for the scientific community with complete annotations of vehicle location, ID, state, calibration data, and semantic trajectory labels. We will discuss the following specific challenges, to which motivation, challenges, benchmark, results, and metrics will be presented:

1) Multi-perspective vehicle detection 2) Multi-perspective vehicle tracking (image plane) 3) Multi-perspective vehicle tracking (3D) 4) Multi-perspective trajectory behavior classification and prediction.

Program

13:30 – 14:00 Visual Trajectories for Highly Automated Driving, Mohan M. Trivedi (University of California, San Diego, USA)

14:00 – 14:30 How to Go from Images to Trajectories?, Akshay Rangesh (University of California, San Diego, USA)

14:30 – 15:00 Detection and Tracking from Multiple Perspectives, Kevan Yuen (University of California, San Diego, USA)

15:00 – 15:30 Coffee break and poster session 15:30 – 16:30 Invited talk

16:30 – 17:30 VIVA-Trajectories Dataset, Ravi Kumar Satzoda (University of California, San Diego / Nauto)

W8-Workshop 8: CPSS-based Parallel Driving (9:00 – 12:30) Room: Marina Organizer(s): Fei-Yue Wang, Institute of Automation, Chinese Academy of Sciences, China; Nanning Zheng, Xi’an Jiaotong University, China; Li Li, Tsinghua University, China; Lingxi Li, Indiana University-Purdue University Indianapolis, USA; Dongpu Cao, Cranfield University, U.K. Abstract: Along with ACP-based parallel management and control and its wide real-world applications in the past decade, cyber-physical-social systems (CPSS)-based parallel driving has been steadily developed. This is also greatly correlated with the emerging development in connected and automated vehicles. This workshop aims to compile the latest research and development advances in parallel driving, and present and highlight the emerging new technologies in Connected and Automated Vehicles in China. The topics of this Workshop include, but are not limited to:

• Theories of ACP-based parallel management and control for parallel driving • Development of artificial transportation systems in parallel driving • V2V, V2I and V2S (vehicle

to service) communications and control for parallel driving • Parallel transportation systems for mixed-level automated vehicles • Driver-vehicle interaction and

collaboration between physical and artificial driving systems • Testing, demonstration and assessment of parallel vehicles • Parallel testing approaches for automated vehicles • All the latest advances and emerging new techniques in Connected and Automated Vehicles in

China. Program

9:00 – 9:30 Parallel Vehicles Based on the ACP Theory: Safe Trips via Self-Driving, ShuangShuang Han (Chinese Academy of Sciences, China)

9:30 – 10:00 How Safe is Safe Enough? - Evaluation and Testing of Intelligent Vehicles, Ding Zhao (University of Michigan, USA)

10:00 – 10:30 Driver-Automation Indirect Shared Control of Highly Automated Vehicles with Intention-Aware Authority Transition, Renjie Li, (Tsinghua University, China)

10:30 – 11:00 Jun Ni (University of California, Berkeley, USA) 11:00 – 11:30 Wenshuo Wang (University of California, Berkeley, USA)

11:30 – 12:00 Transforming Driver Assistance Systems to Autonomous Vehicles: Recent Technical Progress in Tsinghua University, Shengbo Li (Tsinghua University, China)

12:00 – 12:30

Parallel Driving: A Novel Paradigm of Cloud-based Unmanned Driving Technologies, Lingxi Li (Indiana University-Purdue University Indianapolis, USA; Qingdao Academy of Intelligent Industries, China; University of Chinese Academy of Sciences, China)

T-Tutorial: Dedicated Short Range Vehicular Communications: Overview, Technical Challenges, and Applications (9:00 – 12:00) Room: Esplanade Organizer(s): Gaurav Bansal and Dr. John Kenney, Toyota InfoTechnology Center, USA Abstract: In this tutorial we cover the most important aspects of Dedicated Short Range Communications (DSRC), also known as Cooperative ITS. This technology is in the early stages of deployment in North America, Europe, and other regions. The US DOT plans to require DSRC in new vehicles in the coming years. DSRC is used to communicate vehicle-to-vehicle (V2V) and vehicle-to/from-infrastructure (V2I), enabling a set of compelling safety, mobility, automated driving, and environmental applications. This tutorial focuses on the safety and automated driving use cases. We explain the DSRC protocol stack, collision avoidance applications, and technical challenges for deployment. We discuss large-scale field tests and early deployment projects in the US, Europe, and Japan, e.g. the US Safety Pilot and the Rotterdam-Vienna Corridor Project. After presenting DSRC basics, we focus on a specific research problem that is currently of great interest: DSRC Channel Congestion. We discuss the merits of various approaches to address congestion, including avoidance and active control, as well as control modalities (message rate, transmit power, etc.). As a case study we present our specific research on adaptive message rate control, which is under consideration for standardization in the US and Europe. We end the tutorial with a discussion of the role DSRC can play in support of automated vehicles, including a framework for communicating dynamic road conditions to nearby vehicles. The primary goal of the tutorial is to empower the attendee to participate in this important emerging technology, whether as a researcher, a developer, or a planner. The objectives of the tutorial are:

a. Master the fundamentals of a critical emerging VTC technology, DSRC b. Design collision avoidance applications based on V2V communication c. Evaluate the impact of strategic investment on DSRC deployment d. Join the DSRC research community, contributing to solutions for congestion control and other

technical challenges e. Incorporate DSRC in automated vehicle development

The syllabus of the tutorial: a. DSRC Technology (15 minutes) b. Vehicular Safety Communications (20 minutes) c. Overview of DSRC Protocols (20 minutes) d. Technical and Policy challenges for deployment (20 minutes) e. Field tests and early deployments (20 minutes) f. DSRC channel congestion control (40 minutes)