Co-operative Systems in Support of Networked Automated Driving by 2030
Dr. Angelos Amditis, ICCS , Athens- Greece AutoNet2030 coordinator, firstname.lastname@example.org
Timing and duration of tests
October 2015 – September 2016
Location(s) of tests
Björkvik test track (Sweden), AstaZero test track (Gothenburg Sweden), CRF premises (Trento Italy), INRIA premises (Paris, France)
INRIA testing involves fully automated electric buses. In all other cases heavy-duty trucks (with cooperative ACC capabilities) were/will be tested together with a fully-automated and a manually-driven vehicle.
Level of automation tested
SAE / NHTSA Level 0 - 5
Tested use cases
In the INRIA premises the urban AutoNet2030 use cases will be tested (fail-safe automated driving at urban intersections). In all others the AutoNet2030 highway use cases have partially been/will be tested (formation of trucks’ convoy, merging of manual/automated vehicle into the convoy, leaving the convoy).
Tested transport system
Purposes of testing
- Assessment of impacts on safety (short-term, long-term)
- Interaction with other traffic participants (automated vehicles, conventional vehicles, vulnerable road users)
- Technical assessment, proof of concept (incl. vehicle, background support systems such as communication)
Definition of baseline
No direct one-to-one comparison with some reference system or baseline approach. (What is targeted-for is to reproduce the use cases with the highest possible level of accuracy and realism)
The first step involves testing of individual components performed by responsible partners (lab tests or remote tests over the internet). Then follows the installation of the different components (such as the Manuever Controller, the communications stack, the perception module, HMI etc) over the available vehicle platforms and their operational testing under simple drives (i.e., control of velocity, check of positioning, and communications). The final stage involves the realization of the use cases and the safe maneuvers of the vehicles that will showcace the effective operation of the (integrated) AutoNet2030 system.
Method of testing
Controlled field tests, Simulation, Test track
Test fleet, participants and environment
Number and make of vehicles
2 SCANIA R480 tractors 1 FIAT 500L (manually driven) 1 WV Passat (automated HITACHI vehicle) 2 INRIA Cyberbuses (Yamaha automated electric buses)
Description and number of participants/drivers
5-10 participants representing involved partners in each testing session. Certified drivers from the corresponding OEMs participated in the testing.
Tested environment and facilities
Björkvik test track: 2.5km long track (http://www.bjorkvikring.se/) AstaZero: world’s first full-scale test environment for future road safety (http://www.astazero.com/) Trento: CRF Trento garage and use of 2.5km 2-lanes public road for simple drives (i.e., sensors testing) INRIA: closed testing environment within the INRIA facilities
Legal and ethical aspects
No legal or ethical issues to be reported. (All test sites are certified for testing activities confronting to high security regulations. In Trento the public road was used to test only the vehicles’ positioning system and sensors).
Duration of testing
- Björkvik test track: multiple number of journeys along 2 different testing sessions in October 2015 and April 2016
- AstaZero: multiple number of journeys along 3 different testing sessions to take place in June, September and October 2016
- INRIA (Paris facilities): 3 days testing of communication stack and perception capabilities in November 2016 (without moving the buses).
- Trento: 3 drives for testing only sensory equipment
Input parameters and assumptions of simulation tests
Webots-based (https://www.cyberbotics.com/) simulation to study the scalability (in terms of the number of involved vehicles) of the AutoNet2030 system.
- Communication data (V2V, V2I, V2U, etc.)
- Sensor data
- System internal data
Specifications for the data sources
- Communication data (V2V, V2I, V2U, etc.), specify: exchanged CAM messages
- System internal data, specify: LCM loggings of all internal AutoNet2030 interfaces
- Sensor data, specify: (Camera and Lidar recordings for internal validation purposes)
Key Performance Indicators (KPIs)
No KPIs explicitly defined (see ‘methodology’)
Situational data available
Subjective data collected
Issues that affected the impact assessment
None detected so-far
(Preliminary results suggest that most of the modules are well-integrated and capable of providing the desired output; these results essentially serve as inputs for further testing and validation.) The testing results of a fully-integrated and operational AutoNet2030 system are expected in the upcoming summer 2016 sessions
No publication directly related to testing or demonstration itself; a number of papers presenting different parts of the AutoNet2030 work and the so-far results of the project can be found here http://www.autonet2030.eu/?page_id=76
Other things to report