Starting date : May 2015 > Apr. 2018
Lifetime: 36 months
Program in support : H2020-MG-3.5a-2014 _2Stages: mobility for growth
Status project : complete
CEA-Leti's contact :
Benoît Denis
Bernard Strée
Project Coordinator: Jacobs University Bremen (DE)
Partners: - DE: German Aerospace Center, Ibeo Automotive Systems, IT21 Robert Bosch, Objective Software, Zigpos
- FR: BeSpoon, CEA-Leti, Eurecom
- LU: FBConsulting
- NL: PaulsConsultancy, Tass International
- SE: Chalmers University of Technology Gothenburg
Target market: n/a
Publications:
“Breaking the Gridlock of Spatial Correlation in GPS-aided IEEE
802.11p-based Cooperative Positioning”, G.-M. Hoang, B. Denis,
J. Härri, D. Slock, IEEE Transactions on Vehicular Technology
(IEEE Trans. VT), Connected Vehicles Series, Vol. 65, Is. 12,
pp. 9554-9569, Aug. 2016.
“Cooperative Localization in VANETs: An Experimental
Proof-of-Concept Combining GPS, IR-UWB Ranging and V2V
Communications”, G.-M. Hoang, B. Denis, J. Härri, D. Slock,
IEEE Workshop on Positioning, Navigation and Communications
2018 (IEEE WPNC’18), Bremen, Oct. 2018.
Wideband Pedestrian to Vehicle Propagation Channel
Characterization and Modeling”, G. Makhoul, R. D’Errico,
C. Oestges, European Conference on Antennas and Propagation
2018 (EuCAP’18), London, April 2018.
“Rethinking Cooperative Awareness for Future V2X Safety-critical
Applications”, I. Khan, G.-M. Hoang, J. Härri, IEEE Vehicular
Networking Conference 2017 (IEEE VNC’17), Torino, Nov. 2017.
“Robust and Low Complexity Bayesian Data Fusion for Hybrid
Cooperative Vehicular Localization,” G.-M. Hoang, B. Denis,
J. Härri, D. Slock, IEEE International Conference on
Communications 2017 (IEEE ICC’17), Paris, May 2017.
“Cooperative Localization in GNSS-aided VANETs with Accurate
IR-UWB Range Measurements”, G.-M. Hoang, B. Denis,
J. Härri, D. Slock, IEEE Workshop on Positioning, Navigation
and Communications 2016 (IEEE WPNC’16), Bremen,
Oct. 2016 > Best Paper Award.
“On Communication Aspects of Particle-Based Cooperative
Localization in GPS-aided VANETs”, G.-M. Hoang, B. Denis,
J. Härri, D. Slock, IEEE Intelligent Vehicles Symposium (IEEE IV’16),
Workshop on Cooperative Communication and Positioning (CCP),
Gothenburg, June 2016.
Investment: € 6.0 m.
EC Contribution: € 6.0 m.
| Stakes
Design and proof-of-concept validation of cooperative algorithms for high-accuracy vehicle positioning (within ~ 0.25 m) relying on standardized technologies (e.g. ITS-G5 V2X connectivity, GNSS absolute positioning, Impulse Radio - Ultra Wideband relative ranging) and with a limited footprint in vehicle-to-vehicle (V2V) communications (in terms of channel congestion, overhead, latency), including: - A flexible hybrid data fusion framework based on advanced Bayesian filtering - Selection mechanisms based on theoretical performance bounds, integrating only the most informative contributions from neighboring vehicles (e.g. based on Bayesian performance bounds) - Optimized V2V broadcast policies (i.e. controlling transmit power, message payload and transmit rate, parametric message approximation) to ensure virtually constant localization accuracy, while complying with standardized decentralized congestion control mechanisms and message size - Mitigation of harmful effects inherent to V2V cooperation at both signal processing and scheduling levels (e.g. space-time correlations of fused observations, particle depletion and filter overconfidence in high-dimensional problems, error propagation among vehicles, poor dilution of precision in the cross-track dimension, etc.) - Vehicle-to-infrastructure extensions of the cooperative fusion framework to mitigate long-term GNSS denials.
Refinement of existing V2V radio channel models (e.g. geometry-based stochastic model implementation) and unprecedented vehicle-to-pedestrian channel sounding/ characterization.
Studies and recommendations regarding the integrity/security of cooperative V2x radio links against jamming and interception attacks.
Contributions to integration and final proof-of-concept demonstration of the V2V-aided cooperative localization system at project consortium level (including 4 equipped vehicles on a highway section).
Cooperative - Intelligent Transport System (C-ITS) applications require precise knowledge of vehicle geographical positions. Unfortunately, conventional satellite-based navigation systems (e.g. GPS and Galileo) cannot provide sufficient accuracy in practical operating environments (e.g. urban canyons, tunnels) or else they can suffer from excessive convergence times (cold start). The HIGHTS project has addressed this problem by combining V2X data communications (e.g. relying on the ITS-G5 standard), GNSS/GPS data, ranging technologies (e.g. based on Ultra Wideband) and, optionally, other on-board sensor measurements (e.g. LIDAR, camera lane detection, inertial navigation units, wheel speed counter, etc. when available). Specific semantic representations have also been proposed to manage and share multi-layer Local Dynamic Maps (LDM) among vehicles and/or infrastructure. LDMs account for the relative positions of passive and/or cooperative entities around each vehicle as well as those of geo-referenced active road side units, etc.…to support new cooperative localization functions and to extend the nominal perception range of each vehicle. Relevant contributions are intended for possible integration into the “facilities” layer of the “C-ITS station” communication architecture (ISO/ETSI), thereby ensuring availability to various applications in challenging use cases such as autonomous driving and platooning. As a contribution to defining a European-Wide Positioning Service Platform (EWPSP), the HIGHTS system architecture has also been advanced to select contextually the best combination of positioning algorithms depending on the context (i.e. environment, available sensor data) and the needs of the applications and services running on top (typically, an a priori accuracy class). Finally, the project context has prompted integration of a scaled proof-of-concept system demonstrator involving up to 4 cooperating vehicles. Experimental validations conducted on a highway section have shown that significant performance gains can be achieved through V2X cooperation (typically, within 0.25m accuracy in steady-state regimes) compared with standard standalone GNSS capabilities (typically, above 2m accuracy or even under GNSS denial).
IMPACT
In general terms, HIGHTS’ concept and platform pave the way for upcoming road safety applications and higher automation levels, thus contributing to reducing fuel consumption and (near-)crashes. HIGHTS has contributed to ETSI standardization (TC ITS) regarding the format of V2X messages for high-accuracy localization or the format and interfaces of Local Dynamic Maps. The developed system architecture (complying with EWPSP and ETSI visions) is open so as to authorize integration of new technologies.
|
|