Marc Revilloud

Map-aided localization with lateral perception

Accurate localization of a vehicle is a challenging task as GPS available on the market are not designed for lane-level accuracy application. Although dead reckoning helps, cumulative errors from inertial sensors result in a integration drift. This paper presents a new method of localization based on sensors data fusion. An accurate digital map of the lane marking is used as a powerful additional sensor. Road markings are detected by processing two lateral cameras to estimate their distance to the vehicle. Coupled with the map data in a EKF filter it improves the ego-localization obtained with inertial and GPS measurements. The result is a vehicle localization at an ego-lane level of accuracy, with a lateral error of less than 10 centimeters.

An improved approach for robust road marking detection and tracking applied to multi-lane estimation

In this paper, an original and innovative algorithm for multi-lane detection and estimation is proposed. Based on a three-step process, (1) road primitives extraction, (2) road markings detection and tracking, (3) lanes shape estimation. This algorithm combines several advantages at each processing level and is quite robust to the extraction method and more specifically to the choice of the extraction threshold. The detection step is so efficient, by using robust poly-fitting based on the point intensity of extracted points, that correction step is almost not necessary anymore. This approach has been used in several project in real condition and its performances have been evaluated with the sensor data generated from SiVIC platform. This validation stage has been done with a sequence of 2500 simulated images. Results are very encouraging : more than 95% of marking lines are detected for less than 2% of false alarm, with 3 cm accuracy at a range of 60 m.

Accurate lateral positioning from map data and road marking detection

Abstract We are witnessing the clash of two industries and the remaking of in-car market order, as the world of digital knowledge recently made a significant move toward the automotive industry. Mobile operating system providers are battling between each other to take over the in-vehicle entertainment and information systems, while car makers either line up behind their technology or try to keep control over the in-car experience. What is at stake is the map content and location-based services, two key enabling technologies of self-driving cars and future automotive safety systems. These content-based augmented geographic information systems (GIS) as well as Advanced Driver Assistance Systems (ADAS) require an accurate, robust, and reliable estimation of road scene attributes. Accurate localization of the vehicle is a challenging and critical task that natural GPS or classical filter (EKF) cannot reach. This paper proposes a new approach allowing us to give a first answer to the issue of accurate lateral positioning. The proposed approach is based on the fusion of 4 types of data: a GPS, a set of INS/odometer sensors, a road marking detection, and an accurate road marking map. The lateral road markings detection is done with the processing of two lateral cameras and provides an assessment of the lateral distance between the vehicle and the road borders. These information coupled with an accurate digital map of the road markings provide an efficient and reliable way to dramatically improve the localization obtained from only classical way (GPS/INS/Odometer). Moreover, the use of the road marking detection can be done only when the confidence is sufficiently high (punctual use). In fact, the vision processing and the map data can be used punctually only in order to update the classical localization algorithm. The temporary lack of vision data does not affect the quality of lateral positioning. In order to evaluate and validate this approach, a real test scenario was performed on Satory’s test track with real embedded sensors. It shows that the lateral estimation of the ego-vehicle positioning is performed with a sub-decimeter accuracy, high enough to be used in autonomous lane keeping, and land-based mobile mapping.

A new multi-agent approach for lane detection and tracking

This paper proposes an unconventional approach for multi-lane detection and tracking based on a reactive multi-agent system. Most of the algorithms use camera information with a two-step process to detect road marking (1) extraction of road marking features, (2) lane estimation and tracking, performed by studying the extracted point distribution. However, our proposed method is based on a confidence map instead of lane marking features, and a multi-agent model instead of geometric fitting. This approach takes better account of the specific features of road markings, and more precisely, parts defined by clothoids. The method has been tested on a real-world dataset of images in real condition and evaluated with a sequence of more than 2500 synthetic images provided by the SiVIC platform. First results are very promising, with more than 98% for ego lane detection and 97% for multi-lane detection with 4% of false alarm. Furthermore, this approach gives us new opportunities to improve lane detection which would be difficult to implement in a more conventional approach.

A cooperative fusion architecture for robust localization: Application to autonomous driving

The localization of a vehicle is a central task of autonomous driving. Most of the time, it is solved by considering a single algorithm with a few sensors. In this paper, we propose a cooperative fusion architecture based on two main algorithms: a laser-based Simultaneous Localization And Mapping (SLAM) process and a lane detection and tracking approach using a single camera. Both algorithms are designed individually as cooperative fusion processes where other sensors (GPS and proprioceptive information) and dedicated maps are integrated to strengthen the advantages of each system. The whole architecture is formalized around key components (ego-vehicle, roadway, obstacle and environment). A final decision layer, that takes into account the state of each algorithm, allows the system to choose the most appropriate ego-vehicle localization mean based on the current road situation and the environmental context.

A lane marker estimation method for improving lane detection

In this paper we present a new lane markers detection and estimation algorithm aiming to improve lane detection methods. We first estimate the area of lane marking using the profile of the lane estimation in a confidence map. After that a fitting method is applied to improve the lane marker detection accuracy. To track our lane markers over time and make the association between two iteration, we use transferable belief model. The final lane markers are then used to filter out noise and to improve the lane estimation. The presented method is validated on both real and synthetic datasets. Without the integration of lane markers, we obtain a 98% ego detection rate on synthetic data for the lane detection algorithm presented in [8]. With the use of our lane marker detection algorithm, we are able to improve ours results by 1.5% thus reaching a detection rate of 99.5%. For the real validation, we use a highway sequence of about 4000 pictures and get 95.02% of good detection with 0.45% of false alarm for the first thousand images and 99% for ego-lane detection with no false alarm for all the sequence.

PerSEE: A central sensors fusion electronic control unit for the development of perception-based ADAS

Automated vehicles and Advanced Driver Assistance Systems (ADAS) face a variety of complex situations that are dealt with numerous sensors for the perception of the local driving area. Going forward, we see an increasing use of multiple, different sensors inputs with radar, camera and inertial measurement the most common sensor types. Each system has its own purpose and either displays information or performs an activity without consideration for any other ADAS systems, which does not make the best use of the systems. This paper presents an embedded real-time system to combine the attributes of obstacles, roadway and ego-vehicle features in order to build a collaborative local map. This embedded architecture is called PerSEE: a library of vision-based state-of-the-art algorithms was implemented and distributed in processors of a main fusion electronic board and on smart-cameras board. The embedded hardware architecture of the full PerSEE platform is detailed, with block diagrams to illustrate the partition of the algorithm on the different processors and electronic boards. The communications interfaces as well as the development environment are described.

Ecodriving performances of human drivers in a virtual and realistic world

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.