Simon Lacroix

Position estimation in outdoor environments using pixel tracking and stereovision

Presents a method that estimates robot displacements in outdoor unstructured terrain. It computes the displacements on the basis of associations of 3D points sets produced by consecutive stereovision frames, the associations being determined by tracking pixels from one image frame to the other. The paper details the various steps of the algorithms, and first experimental results are presented: they show that the algorithm is able to estimate the 6 parameters of the robot position with a relative error smaller than about 5%, processing several hundreds of images over several tens of meters.

Evaluation of stereo matching algorithms for occupant detection

Vision systems within vehicles offer new opportunities in the automobile industry. The detection and classification of passenger and driver seat occupancy open up new ways to improve the safety and comfort of the passengers. We present results of a stereo system designed for the observation of the cockpit scene in order to provide information about the passenger presence and location within the vehicle to improve the control of the airbag firing. We compare different techniques and examine the effect of random and systematic errors on the performance in precision, robustness and processing speed. These results establish a foundation for on-going work on occupant detection for vehicle safety.

Visual Landmark Constellation matching for spacecraft pinpoint landing

This paper presents a vision-based approach to estimate the absolute position of a planetary lander during the last stages of the descent. The approach relies on the matching of landmarks detected in the descent imagery with landmarks previously detected in orbiter images. The matching process must be robust with respect to scale and radiometry dierences in the image data: it mainly relies on the geometric repartition of the landmarks, rather than on radiometric signatures computed from the image signal. It must also satisfy other requirements like low memory requirement and ecient hardware implementation due to spatial system’s constraints. First results using a simulator are presented and discussed.

Loop closure detection using small-sized signatures from 3D LIDAR data

This article presents a technique for loop closure detection using lidar data for autonomous mobile robots. The technique consists in extracting, indexing and matching a set of small-sized signatures from lidar data. These signatures are based on histograms of local surface normals for 3D point clouds. As the robot moves, some histograms are automatically selected as key-histograms and histograms of newly acquired lidar scans are matched to the key-histograms in order to detect loop-closures. Results with real 3D lidar data validate the proposed technique.

Failure anticipation in pursuit-evasion

This paper presents a new approach for the pursuit of targets by a team of aerial and ground robots under realistic conditions. Mobile target pursuit is often a sub-task of more general scenarios, that call for environment exploration or monitoring activities. Since most of the time a single robot is sufficient to ensure the pursuit of a target, our approach aims at minimizing the team resources devoted to the pursuit: while ensuring the pursuit, a single pursuer evaluates its own potential failures on the basis of the situation defined by the target evolution and the environment structure. It thus assesses its need for team support. When support is necessary to keep the target in view, one or more additional robots are involved, according to a task allocation scheme. We provide mathematical bounds of the complexity of the approach, that ensure that the system has real-time performance. Simulations in a variety of realistic situations illustrate the efficiency of the proposed solution.

Localizing underwater targets using a cooperative AUV architecture

In this paper, we describe a cooperative architecture in which multiple Autonomous Underwater Vehicles (AUVs) cooperate to locate underwater targets. Each AUV is endowed with a sensor that estimates its distance with respect to targets, and cooperate with others to explore an area with the help of an autonomous surface vehicle (ASV) that stores the gathered information map and acts as a communication hub between the AUVs. The communication constraints associated to this context preclude the re-use or adaptation of most of the existing approaches, and require dedicated strategies. In particular, rendezvous are required to exchange information between AUVs and the ASV and to monitor the mission execution. To provide the required autonomy to these vehicles, we build upon an existing system (T-REX), which provides an embedded planning and execution control framework. Simulation results are carried out to evaluate the proposed architecture and adaptive exploration strategy.