Cooperative driving and braking applications, Maurice Kwakkernaat
2 Who is TNO? TNO The Netherlands Organisation for Applied Scientific Research Founded by law in 1932 Statutory, non-profit research organization 600 m turnover 4000+ employees worldwide ~50% with academic degree Headquarters in Delft, The Netherlands World-wide sales offices (TASS) Mission TNO connects people and knowledge to create innovations that boost the sustainable competitiveness of industry and well-being of society
3 Mobility in different expertise at TNO Injury Driver Safe and Clean Monitoring ICT & Traffic Prevention Behaviour Vehicle Systems & Sensors Management
4 Communication supported automated driving
5 Cooperative Driving Influencing individual vehicles through advisory or autonomous actions in order to optimize their collective behaviour Main enablers Wireless vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communications Benefits Advanced driving assistance systems (ADAS) Intelligent, but simple human-machine interfaces (HMI) To improve safety of cars, pedestrians, and cyclists To reduce traffic jams and increase throughputs of the roads Eco-efficient: less number of stop/go pollution less fuel consumption, less
6 Cooperative driving & TNO TNO made its first cooperative application in 1999 Involved in many EU and National projects, e.g. SAFESPOT, COMPASS4D, DRIVEC2X, SPITS, etc.
7 Cooperative driving facilities @ TNO Helmond 15 intersections 7 km highway Inter urban and urban
8 TNO application projects Cooperative Adaptive Cruise control Large scale experiments shockwave damping: driver advice and automated driving Contrast System integration of speed advice function using test site infrastructure, 6 month evaluation by 80 users Spits Live Large scale deployment of cooperative technology Shockwave damping 2015: Start Pilot 1000 vehicles => 100000 before 2020
9 Main cooperative driving applications Warning applications Road works warning, Traffic jam, Emergency vehicle, weather, Efficiency Floating car data In-vehicle signage GLOSA (Green Light Optimised Speed Advice) Shockwave damping Active Safety Collision avoidance technologies (e.g. Cooperative Automated Emergency Braking)
10 Introduction AEB and collision avoidance systems Sensor systems Real-time feedback Small time constrains Limitations in effectiveness Reaction-time (sub-second) Range (operational velocity) Cross-range (crossing scenarios) Line-of-sight (e.g. around a corner) Beyond field-of-view (e.g. blind spot) Vehicle Mercedes Volvo Mitsubishi Fiat Ford System PRE-SAFE Brake City Safety & Collision Warning with Full Auto-Brake Forward Collision Mitigation City Brake Control Active City Stop
11 *ASPECSS D1.1, 2013 Introduction Challenges Vulnerable road users (VRUs) Lateral movement Obstructed views Earlier response Extended range and field-of-view Reliable object detection and prediction Enabler Wireless communication (V2X) Cooperative automated emergency braking (C-AEB)
12 Example scenario 10 m/s FOV = 60 deg. Building TTC ~ 1 s
13 Objective Show that C-AEB can improve sensor only AEB Focus on improved bicycle collision avoidance Improved situational awareness Information fusion with in-vehicle sensors Early object detection Improved lateral tracking Reliable object classification Reduce need for expensive sensors Provide necessary level of redundancy and performance
14 Technology used Toyota Prius Dummy bicycle Wireless communication ITS-G5 (ETSI CAM) GPS and motion sensors Radar and camera sensors Information fusion, object tracking Vehicle and bicycle host tracking AEB control algorithmic m 0.5 m
15 System architecture and design TNO s Intelligent Vehicle Safety Platform (ivsp)* Centralized safety awareness platform Scalable and layer-based Process and exchange information from multiple sources Generates real-time situational awareness Sensor Layer Application Layer Information Layer Communication Layer Interface and authorization layer * Kwakkernaat et al.: Layer-based Multi-Sensor Fusion Architecture for Cooperative and Automated Driving Application Development, AMAA, Berlin, 23 June 2014.
16 Host vehicle Manual brake Vehicle actuation HMI Application layer Bicycle object Host tracking Actuator control Bicycle object data Risk level, TTC C-AEB control Risk estimation Status info User control mode Vehicle motion data Information layer -- Bicycle motion data ITS-G5 communication (CAM) Bicycle motion data Object tracking ITS-G5 communication (CAM) -- Host tracking -- Communication layer Sensor data Radar, camera data Sensor data On-board GPS, motion sensors, odometer On-board radar, and camera On-board GPS, motion sensors, odometer Sensor layer
17 Host and object tracking Host tracking GPS delay correction Sensor fusion Linear Kalman filter Object tracking Logical fusion approach Discrete Kalman filters Sensor noise covariance matrix Sensor delay compensation GPS Motion sensors Host Tracking KF Latitude Longitude Velocity Heading Yaw-rate Acceleration Logical fusion d d d α α Radar 1 1 2 2 3 Camera 2 3-1 1 Communication 3 2 1 3 2 Sensor data Preprocessing Feature filtering Data clustering Object state estimation objects
18 Testing and evaluation (indoor)
19 Testing and evaluation (outdoor)
UTM - Y [m] 20 Test results Vehicle-cyclist crossing scenario Partly obstructed sight 5.706 x 106 UTM plot of test scenario 5.706 Cyclist - OxTS Object tracking output Host Radar Field-Of-View (start) Radar Field-Of-View (end) 5.706 5.706 5.706 5.706 Building 5.706 6.8226 6.8227 6.8227 6.8228 6.8228 6.8228 6.8229 UTM - X [m] x 10 5
21 Conclusions Obstructed vehicle-cyclist crossing scenario Early object detection possible (> 3 sec.) Improved lateral tracking GPS not sufficient, host tracking required Current and future work Infrastructure detection and communication Distributed sensing and communication User-tests (HMI)
22 Demo at imobility challenge 2013
23 Related work for further enhancements Extended host tracking and prediction using vehicle state estimation Collision avoidance steering strategies for adverse conditions
24 Combined CAEB - CACC
25 THANK YOU FOR YOUR ATTENDANCE Contact info: (Sven.jansen@tno.nl, +31 8886 65743) TNO Technical Sciences / Automotive Steenovenweg 1, 5708 HN, Helmond, The Netherlands Reference: Kwakkernaat, Maurice*; Ophelders, Frank; Vissers, John; Willemsen, Dehlia; Sukumar, Premnaath;, COOPERATIVE AUTOMATED EMERGENCY BRAKING FOR IMPROVED SAFETY BEYOND SENSOR LINE-OF-SIGHT AND FIELD-OF-VIEW, FISITA 2014 World Automotive Congress, Maastricht, The Netherlands, June 2014