Vehicle Technologies Activities of WorkPackage 2

1. Vehicle Technologies Activities of WorkPackage 2 Giancarlo Alessandretti - Centro Ricerche Fiat WP leader WP Partners: CRF, Yamaha-Motor Frog, Robosoft, RUF, Inria, Univ. Bristol, ISR

2. General project objectives Vehicle technologies, together with infrastructure technologies and general methodologies, are a major aspect for the objective of CyberCars project: to accelerate the development of novel urban transportation systems, based on automated vehicles, offering a cleaner and safer transportation mode available to everyone. Vehicle technologies Vehicle technologies Infrastructure technologies Methodologies

3. Contents • Background • Objectives for vehicle technologies development • Approach in CyberCars project • Partners and roles • Main developments and results: • Vehicle controls • Navigation • Obstacle detection • Platooning • Outlook and conclusions

4. Background

5. Full driving automation on : Intercity tracks . City tracks Calm zones Pedestrian zones Local tracks Private vehicles Private grounds Public vehicles Cybercar Road Map

6. Intercity tracks . City tracks Calm zones Pedestrian zones Local tracks Private grounds Industrial automation Base technologies have been developed in the past for robotics; however, applications on road vehicles for passengers require new approaches, with focus on better performance and reduced cost.

7. Intercity tracks . City tracks Calm zones Pedestrian zones Local tracks Private grounds Local tracks • Some preliminary experiences have been done on local dedicated tracks; Cybercars is focused on improvements especially in term of : • Following a complex path with high accuracy • Using modular and standard architectures • Assuring very high safety with no control from a driver • New functions like platooning

8. Intercity tracks . City tracks Calm zones Pedestrian zones Local tracks Private grounds Proper consideration of the urban environment is fundamental for technology developments City zones Current situation's section Future situation's section

9. Longitudinal Control Longitudinal Control Urban Drive Ass . Stop&Go ++ Automatic Driving Stop&Go ACC Rural Drive Ass . + Lateral Control + Lateral Control Avanced Driver Assistance Functions on cars ADASE Road Map

10. Driver Assistance functions on passenger cars LATERAL CONTROL LONGITUDINAL CONTROL

11. Lateral control on a public vehicle

12. Telematics • Exchange of informations between on-board units and a Center for monitoring and control – Fleet management • Infomobility functions for the users (access to services)

13. Objectives for vehicle technologies developments

14. Specific Objectives for vehicle technologies • Improve existing technologies (mostly based on robotics) for automatic navigation, platooning, obstacle detection: • Performance, reliability in various conditions • Safety for users • New concepts less dependent from infrastructure • Reduced costs of equipment • Improve the control architecture for HW and SW: • Modularity and flexibility • Development and validation tools

15. Approach

16. Type of vehicles People Mover (20 persons) 4 seats small vehicle Frog Yamaha Prototype for developments General purpose platforms ISR INRIA and Robosoft

17. Dual mode vehicle (design level) Large collective vehicles (design level) CRF RUF Link with other workpackages Type of vehicles Serpentine

18. Domain of operation • Mostly urban mission for the vehicle (private, public) • Low speed, frequent stops, comfortable driving • Areas requiring ecological solutions • Assistance by infrastructure (different levels investigated) • Interaction with other road users

19. Technological areas Four main tasks: • Vehicle controls • Navigation and guiding • Obstacle detection and collision avoidance • Platooning

20. Partners and roles

21. WP coordinator; study of dual mode vehicle; investigation on automotive radar Regenerative brake; vehicle platform New platform for People Mover; architecture and controls, improved obstacle detection, magnetic rulers New RobuCab platform; wire guidance; camera and laser based techniques; platooning Partners and main roles (1)

22. Guidance for large collective vehicles by magnetic probes SW tool; vision based localization; stereo vision; platooning; study of complex maneuvers General study and simulation for platooning schemes Experimental platform: path tracking; integration of inertial navigation and GPS; development of stereo vision Partners and main roles (2)

23. Main developments and results: • Vehicle controls • Navigation • Obstacle detection • Platooning

24. Vehicle controls Integrated solutions for different functions (traction, braking, steering) under computer control New hardware especially for safety Safe software for distributed processing approach: new specific tools for design and certification

25. Park Shuttle ParkShuttle vehicle platform “Frog box” for vehicle control • Features: • Drive, steer and brake controls integrated • Redundancy for safety critical functions • Different levels for braking (normal, fast, emergency)

26. Vehicle for dual-mode operation Scheme of control system System layout • Small and clean thermal engine plus electric traction (Hybrid) • Manual driving on ordinary roads • Automatic control for collection / redistribution & platooning • Standard automotive components (e.g. Electronic Brake System and Power steering)

27. Cost Uphill Gained downhill Regenerative brake Component installed in the Cybercar Current flow for a representative test uphill Current [A] traject • Implemented on the Floriade 2002: 25 vehicles running 5.000 km each • Range improved, comfort improved

28. Sw tool for distributed architecture CAN Network architecture Implementation in different platforms Data flow graph • The SW allows users to create, implement and certify applications for safe AGV control • Fast prototyping demonstrated • Used on a standard basis by project partners for several vehicle platforms

29. Navigation technologies New flexible and simple techniques based on inertial sensors and magnets for fine-tuning of the path (industrial development) • New techniques requiring no modification of the infrastructure tested at demo level: localization of natural features by laser image processing Advanced techniques for path generation in complex and dynamic environments

30. antenna Magnetic ruler Layout of the vehicle Magnetic field measures and fit • Advanced signal processing for improved accuracy • Large distance of antenna from the road (0,2 m) • High speed (20 km/h @ 0,1 m accuracy)

31. Video based localisation Samples (every 10 images) of the detected characteristics of the road. Experiments in Antibes City Centre and Harbour at 10m/s speed and 25 Hz rate • No specific equipment in the infrastructure • Dual stage algorithm (borders of road; matching left/right images) • Well adapted to capture dynamic parts of the image

32. Obstacle detection Systems based on laser scanners associated with ultrasounds and contact sensing on bumpers Advanced logics to control vehicle motion and negotiate the approach to a potential obstacle or to a curve Research on other collision avoidance techniques using vision or radar (including automotive technologies)

33. Advanced data processing for laser scanner Implementation in the Parkshuttle platform Scanner 2 Scanner 1 • FEATURES: • Measurement at two heights for increased safety • On-path and near-path recognition • New approach for tracking (object memory, probabilistic detection) leading to velocity profiles • Easy configuration: area of interest automatically defined from route plan • Enhanced comfort due to strategies for a gradual stopping

34. Laser based system The laser system has been implemented in the Robucab A protection zone and a warning zone are defined • The validation of a ‘certified’ sensor has been obtained • (with warning range 50m and response time 80 ms)

35. OUT-OF-PATH VS. IN-PATH OBSTACLE RESOLUTION OF TWO OBSTACLES ? ? Lane 3,7m Automotive radar systems Specification phase Obstacle detection test Platooning test Range 37 m Speed 50km/h max • Extensive experimentation in six test scenarios focused on obstacle detection and front car detection. • Suitable performances in the platooning scenario demonstrated • Capability to brake behind a stopping vehicle demonstrated • Field of view to be increased w.r.t current automotive radar.

36. Platooning • First positive results in testing of two techniques: • Scanning laser sensor with reflective beacons • Low cost vision system extracting relevant features Analysis of platooning schemes and manoeuvres by modelling

37. Laser detection technique Scheme of the laser based platoon Implementation • Simple approach demonstrated on RobuCab (catch at 2m, disconnect at 15m) • Provides distance and orientation of preceding vehicle • Same technology can be used for obstacle detection

38. Image processing technique Features of the Cycab used for vehicle recognition • Detects geometric features of preceding Cyber car • Basic building block for a fleet of homogeneous vehicles • Adaptable by simple SW modifications

39. Conclusions

40. Concluding remarks • Technological developments for Cyber Cars are evolving towards: improved integrated control, high safety for all road users, accurate path following. • Work in progress on platooning has shown significant potential of this approach for increased capacity and for the management of empty vehicles. • CyberCars are at present designed for short trips at low speed, in private grounds or in defined urban environments. Significant steps for these tasks have been obtained in the project. • For the long-term, CyberCars would operate on a larger network (possibly in dual mode) in more complex traffic situations. • The developed Cybercars technologies (vehicles + infrastructure) allow demonstrations and evaluations in test areas and also in several controlled city test sites (related work in the CyberMove project).