Papers by Dominique Gruyer
2021 AEIT International Conference on Electrical and Electronic Technologies for Automotive (AEIT AUTOMOTIVE), 2021
For a decade, a lot of researches have focused on the development and the deployment of automated... more For a decade, a lot of researches have focused on the development and the deployment of automated mobility services and means for road networks (urban, suburb, rural, highway). In the development of autonomous driving embedded systems, several stages are needed. The first one concerns the perception layers. The second one is dedicated to the risk assessment, the decision and strategy layers and the optimal path planning. The last one addresses the control/command parts. Based on the Autonomous Decision-Support Framework (ADSF) of the EU project Trustonomy, this paper proposes a risk assessment and decision-making framework to the second stage and improves an existing virtual co-pilot by combining a new emergency mode and corresponding trajectory planning algorithm. After introducing the project framework for risk management and the general co-pilot concept developed in the University Gustave Eiffel, the Decision-Support framework, implemented in RTMaps platform, is demonstrated within a realistic 3D simulation environment called Pro-SiVIC. Both the previous virtual copilot and the new emergency algorithm are combined and a switching strategy between the different modes is tested in a near-accident situation.
2021 IEEE International Intelligent Transportation Systems Conference (ITSC), 2021
This paper addresses the proposal of a new multidimensional risk estimator using spatial uncertai... more This paper addresses the proposal of a new multidimensional risk estimator using spatial uncertainty modeling for automated vehicles risk assessment. This risk estimator, on which relies AV decision-making, is based on an extended version of the distance of Gruyer (DG). This estimator provides an answer and a solution to the risk assessment needed as a part of a generic and extended architecture dedicated to the building of a generic driving meta-model usable for multimodal driving behavior simulation (personal vehicle, connected vehicle, connected and automated vehicles and autonomous vehicle). The proposed estimators have been tested, evaluated, and analyzed on a set of representative highway scenarios with three key performance indicators. Results show that the proposed risk estimator (RIMUM) is more realistic, extended DG more reversible. Future works will concentrate on improving their prediction capacity which is lower than reference estimators.
Eco-driving is about energy efficient use of vehicles. There are many approaches to improve eco-d... more Eco-driving is about energy efficient use of vehicles. There are many approaches to improve eco-driving. Eco-driving driving style is hard to learn even with the help of driving assistance systems. This paper presents an immersive driving simulation training tool to support eco-driving training. We address the following questions: - (1) How gamification concepts can improve eco-driving evaluation, training and adoption in simulated environments and - (2) How to setup such elements in a 3D immersive driving simulator. We present an implementation of gamification concepts in a driving simulator architecture built upon pro-SiVIC software and a 3D Helmet Mounted Display. The gamification functions is then used to motivate drivers to be eco-friendly. We conclude with future work and open issues.
Modeling and validation of a new generic virtual optical sensor for ADAS prototyping
2012 IEEE Intelligent Vehicles Symposium, 2012
In the early design stages of embedded applications, it becomes necessary to have a very realisti... more In the early design stages of embedded applications, it becomes necessary to have a very realistic simulation environment dedicated to the prototyping and to the evaluation of these Advanced Driving Assistance Systems (ADAS). This Numerical simulation stage is gradually becoming a strong advantage in active safety. The use of realistic numerical models enabling to substitute real data by simulated data is primordial. For such virtual platform it is mandatory to provide physics-driven road environments, virtual embedded sensors, and physics-based vehicle models. In this publication, a generic solution for cameras modelling is presented. The use of this optical sensor simulation can easily and efficiently replace real camera test campaigns. This optical sensor is very important due to the great number of applications and algorithms based on it. The presented model involves a filter mechanism in order to reproduce, in the most realistic way, the behaviour of optical sensors. The main filters used in ADAS developments will be presented. Moreover, an optical analysis of these virtual sensors has been achieved allowing the confrontation between real and simulated results. An optical platform has been developed to characterize and validate any camera, permitting to measure their performances. By comparing real and simulated sensors with this platform, this paper demonstrates this virtual platform (Pro-SiVIC™) accurately reproduces real optical sensors' behaviour.
Positioning, 2016
This paper presents the Optimized Kalman Particle Swarm (OKPS) filter. This filter results from t... more This paper presents the Optimized Kalman Particle Swarm (OKPS) filter. This filter results from two years of research and improves the Swarm Particle Filter (SPF). The OKPS has been designed to be both cooperative and reactive. It combines the advantages of the Particle Filter (PF) and the metaheuristic Particle Swarm Optimization (PSO) for ego-vehicles localization applications. In addition to a simple fusion between the swarm optimization and the particular filtering (which leads to the Swarm Particle Filter), the OKPS uses some attributes of the Extended Kalman filter (EKF). The OKPS filter innovates by fitting its particles with a capacity of self-diagnose by means of the EKF covariance uncertainty matrix. The particles can therefore evolve by exchanging information to assess the optimized position of the ego-vehicle. The OKPS fuses data coming from embedded sensors (low cost INS, GPS and Odometer) to perform a robust ego-vehicle positioning. The OKPS is compared to the EKF filter and to filters using particles (PF and SPF) on real data from our equipped vehicle.
A Cooperative Vehicle Ego-localization Application Using V2V Communications with CBL Clustering
2018 IEEE Intelligent Vehicles Symposium (IV), 2018
The cooperative vehicle paradigm offers new opportunities for enhancing various vehicle functions... more The cooperative vehicle paradigm offers new opportunities for enhancing various vehicle functions through distributed applications. The localization function is one of the most important since its services are needed by numerous applications ranging from driver navigation to autonomous vehicle guiding. Though the GPS (Global Positioning System) provides honorable service for driver navigation, its precision is not accurate for allowing an autonomous vehicle to localize itself correctly on the road lanes. Recently an ego-localization technique where a group of vehicles exchange their position and the related correction through V2V communications has been proposed in order to enhance the precision of the 10- cation of each node in the group. In this paper, we evaluate the performance that this application can expect from V2V communication services supplied by the CBL (Chain-Branch-Leaf) clustering scheme. The simulation results show that CBL achieves delays and packet delivery ratio adapted to various rates of the ego-localization application traffic.
Technological Forecasting and Social Change, 2019
Ridesourcing services play a crucial role in metropolitan transportation systems and aggravate ur... more Ridesourcing services play a crucial role in metropolitan transportation systems and aggravate urban traffic congestion and air pollution. Ridesplitting is one possible way to reduce these adverse effects and improve the transport efficiency, especially during rush hours. This paper aims to explore the potential of ridesplitting during peak hours using empirical ridesourcing data provided by DiDi Chuxing, which contains complete datasets of ridesourcing orders in the city of Chengdu, China. A ridesplitting trip identification algorithm based on a shareability network is developed to quantify the potential of ridesplitting. Then, we evaluate the gap between the potential and actual scales of ridesplitting. The results show that the percentage of potential cost savings can reach 18.47% with an average delay of 4.76 min, whereas the actual percentage is 1.22% with an average delay of 9.86 min. The percentage of shared trips can be increased from 7.85% to 90.69%, and the percentage of time savings can reach 25.75% from 2.38%. This is the first investigation of the gap between the actual scale and the potential of ridesplitting on a city scale. The proposed ridesplitting algorithm can not only bring benefits on a city level but also take passenger delays into consideration. The quantitative benefits could encourage transportation management agencies and transportation network companies to develop sensible policies to improve the existing ridesplitting services.
Evidential model and hierarchical information fusion framework for vehicle safety evaluation
17th International IEEE Conference on Intelligent Transportation Systems (ITSC), 2014
Vehicle safety evaluation is a systematic and comprehensive process involving vehicles, road envi... more Vehicle safety evaluation is a systematic and comprehensive process involving vehicles, road environments, and driver behaviours. In real road conditions, due to great uncertainty, evaluation based on singular information source lacks in sufficient accuracy and stability. In this paper, we proposed a vision-based real-time vehicle safety evaluation system using lane and driver's eye information, which were modelled in the framework of evidence theory. Vehicle safety was assessed via hierarchical fusion of driver drowsiness detection and distracted and impaired driving performance. The system was validated in real world scenarios. Experimental results demonstrate that it is promising to improve the robustness and temporal response of vigilance of vehicle safety.
2015 IEEE Intelligent Vehicles Symposium (IV), 2015
In this study, results of an ecodriving challenge that took place during the Paris Motor Show in ... more In this study, results of an ecodriving challenge that took place during the Paris Motor Show in 2014 are presented. The principle of this challenge was to drive a simulated passenger car as far as possible with a limited quantity of energy (15 cL). The experimental setup, constituted of the SiVIC software, an Oculus Rift Helmet and a fuel consumption model, is also detailed. 1211 trips of visitors were validated during the 17 days of the event. Results showed that high acceleration without kickdown is desirable and that constant speed can lead to significant reduction in energy consumption. Next work will concentrate on improving the simulation and the scenario to increase the immersion realism and the ecodriving behavior sensitivity.
Descripteurs de forme pour l'indexation de maillages 3D
Techniques et sciences informatiques, 2003
Cet article traite de l'indexation d'objets 3D mailles a l'aide de descripteurs de fo... more Cet article traite de l'indexation d'objets 3D mailles a l'aide de descripteurs de forme (DF), sous contraintes d'invariance geometrique et de robustesse topologique. Le spectre de forme 3D (SF3D), propose par les auteurs et retenu comme DF dans MPEG-7, est tout d'abord introduit. Intrinsequement invariant aux transformations geometriques, le SF3D n'est pas robuste vis-a-vis des representations topologiques multiples. C'est pourquoi un nouveau DF (DH3D), intrinsequement stable topologiquement, est propose. Derive de la transformee de Hough 3D, le DH3D n'est pas invariant aux transformees geometriques. Nous montrons mathematiquement comment il peut etre associe de facon optimale en termes de compacite de representation et de complexite de calcul a une procedure d'alignement spatial lui conferant alors un comportement d'invariance geometrique. Cela conduit a definir le DH3D optimal (DH3DO). Apres avoir specifie les mesures de similarite utilisees lors des applications de requete et avoir decrit la base de 1300 modeles utilisee, les deux DF sont evalues et compares objectivement en termes de score Bull-Eye, ce critere etablissant une nette superiorite du DH3DO.
Manoeuvre-based trajectory planning for highly autonomous vehicles on real road with traffic
2009 European Control Conference (ECC), 2009
This paper presents the design and simulation results of a vehicle path planning algorithm that t... more This paper presents the design and simulation results of a vehicle path planning algorithm that takes into account traffic and other obstacles on a highway. The proposed algorithm is designed to require little computation in order to achieve a real time evaluation in an embedded environment such as an ECU. Hence, the trajectory planning is proposed as a two-step algorithm. The first step defines at high level the feasible manoeuvres according to the environment. The output of this first step is a target group of manoeuvres such as accelerating, decelerating and changing lanes. The second step gives a lower level description of various trajectories within these given manoeuvres, computing all possibilities if the computation power is available. The output is a trajectory described in the vehicle frame that represents the recommended vehicle state (position, heading, speed and acceleration) for the next seconds.
2017 IEEE Intelligent Vehicles Symposium (IV), 2017
In Autonomous driving applications, the LIDAR is becoming one of the key sensors for the percepti... more In Autonomous driving applications, the LIDAR is becoming one of the key sensors for the perception of the environment. Indeed its work principle which is based on distance ranging using a laser beam scanning the environment allows highly accurate measurements. Among sensors commonly used in autonomous driving applications, which are cameras, RADARs and LIDARs, the LIDAR is the most suited to estimate the shape of objects. However, for the moment, LIDARs dedicated to pure automotive application have only up to four measurement layers (4 laser beams scanning the environment at different height). Hence objects detection algorithm have to rely on very few layers to detected and classify the type of objects perceived on the road scene, that makes them specific. In this paper we will present an Detection and Tracking of Moving Objects (DATMO) algorithm featuring an objecttype classification based on the belief theory. This algorithm is specific to automotive application therefore, the classification of perceived vehicles is between bike, car and truck. At the end of this paper we will present an application of this algorithm in real-world context.
This paper deals with real-time obstacle detection and tracking using multi-layer LIDAR data. We ... more This paper deals with real-time obstacle detection and tracking using multi-layer LIDAR data. We present two algorithms to cluster raw data coming from LIDAR sensors. The rst algorithm is based on a dynamic clustering approach while the second one relies on the connectivity between the laser impacts. Both algorithms take into account the inaccuracy and the uncertainty of the data sources. We propose a tracking approach based on the belief theory to estimate the dynamic state of the detected objects in order to predict their future maneuvers. The objects are then ltered using an intelligent ROI that depends on a dynamic evolution area computed from proprioceptive information of the ego-vehicle. We evaluate and validate the whole chained process on real data-sets.
Most of the solutions proposed in fault detection focusing on one part of the analyzed system (Ac... more Most of the solutions proposed in fault detection focusing on one part of the analyzed system (Actuators, System, Sensors) supposing other parts in a nominal and well known state. In this paper a new method aiming to complete a sensor fault detection without any assumptions made on the comportment of the rest of the complete system is proposed. For that, a faults identification will be proposed in order to reject defaults due to environmental perturbations or failure in the system or the actuators, as the authors focus in this paper on sensor fault detection.
IMM-based sensor fault detection and localization for a drive-by-wire vehicle
With the development of the embedded application and driving assistance systems, it becomes relev... more With the development of the embedded application and driving assistance systems, it becomes relevant to develop parallel mechanisms in order to check and to diagnose these new systems. This is particularly true in the case of partially and fully autonomous vehicles. In this paper, we focus our researches on a subpart of this issue dedicated to the detection and the identification of sensor faults for a drive-by-wire road vehicle. An Interacting Multiple Model approach has been implemented, based on a non-linear vehicle dynamics observer, including tire-road force and actuator modeling. Experimental validation tests, relying on real vehicle signals, show a fast and robust fault identification.
Applied Sciences, 2021
For a decade, researchers have focused on the development and deployment of road automated mobili... more For a decade, researchers have focused on the development and deployment of road automated mobility. In the development of autonomous driving embedded systems, several stages are required. The first one deals with the perception layers. The second one is dedicated to the risk assessment, the decision and strategy layers and the optimal trajectory planning. The last stage addresses the vehicle control/command. This paper proposes an efficient solution to the second stage and improves a virtual Cooperative Pilot (Co-Pilot) already proposed in 2012. This paper thus introduces a trajectory planning algorithm for automated vehicles (AV), specifically designed for emergency situations and based on the Autonomous Decision-Support Framework (ADSF) of the EU project Trustonomy. This algorithm is an extended version of Elastic Band (EB) with no fixed final position. A set of trajectory nodes is iteratively deduced from obstacles and constraints, thus providing flexibility, fast computation, a...
This paper presents an online trajectory planning strategy with a modified potential field method... more This paper presents an online trajectory planning strategy with a modified potential field method on distributed architectures for autonomous vehicles. The approach overcomes the well-known artificial potential field method (APFM) issue, which is due to local minima that induce the APFM to stick in. Thus, the standard APFM is no longer useful. The advantage of the new proposed method reverse to those that resort to the global optimization methods is the low computing time which borders up the A-Star (A*) method. The strategy consists of looking for a realistic path in the potential field—according to the potential gradient descent algorithm (PGDA)—and affix a repulsive potential, to the current state, in the case of blocking configuration, a local minimum. When the PGDA reaches the global minimum, a new artificial potential field will be constructed with only one minimum which matches to the final destination of the vehicle, the global minimum. Finally, to determine the achievable t...
2017 IEEE International Conference on Systems, Man, and Cybernetics (SMC), 2017
Dominique (2017) A study on AI-based approaches for high-level decision making in highway autonom... more Dominique (2017) A study on AI-based approaches for high-level decision making in highway autonomous driving. In Cheng, I & Pedrycz, W (Eds.
2017 15th International Conference on ITS Telecommunications (ITST), 2017
Vehicular communications can be achieved through the infrastructure (Vehicle-to-infrastructure ne... more Vehicular communications can be achieved through the infrastructure (Vehicle-to-infrastructure network, V2I), as well as directly through vehicle-to-vehicle communication (V2V) via ad hoc networks. In V2V communications, the routing protocols are designed in order to optimize the dissemination of messages. This paper presents an evaluation of routing protocols such as the Optimized Link State Routing (OLSR), Ad hoc On-demand Distance Vector (AODV), Dynamic Source Routing (DSR), and Geographic Routing Protocol (GRP), while considering both vehicular safety application requirements and mobility models based on real-world traces of vehicular traffic. The results show that, though proactive routing protocols perform better in this context, the four routing protocols fail to fulfill the safety application requirements on the delay metric even for a reasonable number of vehicles.
2017 IEEE Intelligent Vehicles Symposium (IV), 2017
This paper presents a modified potential field method for robot navigation. The approach overcome... more This paper presents a modified potential field method for robot navigation. The approach overcomes the wellknown artificial potential field (APF) method issue, which is due to local minima that induce the standard APF method to trap in. Thus, the standard APF method is no longer useful in such case. The advantage of the new proposed method, as opposed to those that resort to the global optimization methods, is the low computing time that lines up with the standard A-Star (A*) method. The strategy consists of looking for a practical path in the potential field-according to the potential gradient descent algorithm (PGDA)-and adding a repulsive potential to the current state, in case of blocking configuration, a local minimum. When the PGDA reaches the global minimum, a new potential field will be constructed with only one minimum that matches the final destination of the robot, the global minimum. Finally, to determine the achievable trajectory, a second iteration is performed by the PGDA.
Uploads
Papers by Dominique Gruyer