Hi, My team has an opening for a PhD student to work on multi-robot exploration and mapping. We target search and rescue operations with UGVs controlled using PhaROS (Pharo + our client for the Robot Operating System). You'll find the full description below. More info on our team can be found at: http://car.mines-douai.fr <http://car.mines-douai.fr/> Best, Prof. Noury Bouraqadi Ecole des Mines de Douai Lille area France ------------ CONTACT Prof. Noury Bouraqadi, noury.bouraqadi AT mines-douai.fr <http://mines-douai.fr/> Dr. Luc Fabresse, luc.fabresse AT mines-douai.fr <http://mines-douai.fr/> Dr. Guillaume Lozengez, guillaume.lozengez AT mines-douai.fr <http://mines-douai.fr/> DESCRIPTION Robotic exploration and mapping is basic to many robotic applications and more specifically to search and rescue operations. It relies on robots to build a map of the environment as a first step to further mission specific tasks (eg. extracting trapped victims). In scenarios where time is critical such in the aftermath of an earthquake, the use of multiple robots allows to speed up the exploration and to take advantage of distributed sensor to increase the mapping process. Each robot explores a different but connected area. The global map is built by merging local maps built by individual robots. To build a map of an unknown environment, a robot has to perform Simultaneous Localization and Mapping (SLAM). Based on different perceived data (typically odometery and laser scans), metric SLAM algorithms localize the robot while building an occupancy grid. The occupancy grid models the environment as a matrix where each cell denotes the probability of occupancy of a small area. Best metric SLAM algorithms produce precise maps that allow localization with typically centimetric precision. However, an occupancy grid map is materialized as a large matrix. Beside the memory consumption, such a map leads to high CPU consumption for path planning. CPU consumption is also a concern in multi-robot exploration. Indeed, merging metric maps produced by different robots is computationally intensive. The above drawbacks of metrics maps can be mitigated by the use of topological maps. However, topological maps do not allow precise obstacle delimitations needed for autonomous robots exploration. The goal of this PhD is to contribute to our ongoing effort, and propose a solution that gathers the best of both worlds. The expected outcome is a SLAM algorithm that builds a topological map that embeds metric information. The target algorithm should be implemented on top of the ROS (www.ros.org <http://www.ros.org/>) middleware and validated both in simulation and on actual robots. It should also be supported by a set of automated tests that can safely run on an actual robot. BIBLIOGRAPHY - Robotic Mapping and Exploration. C. Stachniss. Springer 2009. - An evaluation of 2D SLAM techniques available in Robot Operating System. J. M. Santos, D. Portugal, and R. Rocha. SSRR 2013. - A Methodology for Testing Mobile Autonomous Robots. Jannik Laval, Luc Fabresse and Noury Bouraqadi. IROS 2013. - Towards Test-Driven Development for Mobile Robots. Luc Fabresse, Jannik Laval and Noury Bouraqadi. ICRA 2013 eigth workshop on Software Development and Integration in Robotics (SDIR VIII). - Team Size Optimization for Multi-robot Exploration, Zhi Yan, Luc Fabresse, Jannik Laval, and Noury Bouraqadi, SIMPAR 2014. - Metrics for Performance Benchmarking of Multi-robot Exploration. Zhi Yan, Luc Fabresse, Jannik Laval, and Noury Bouraqadi. IROS 2015. - Punctual versus continuous auction coordination for multi-robot and multi-task topological navigation, Guillaume Lozenguez, Lounis Adouane, Aurélie Beynier, Abdel-Illah Mouaddib, Philippe Martinet, Autonomous Robot 2015.