Multi-robot exploration with space-based potential field map
The potential field method is a widely used approach for obstacle avoidance path-planning. It performs path-planning efficiently with multiple obstacles and robots. In this thesis we propose an approach to controlling multiple robots for area exploration missions that focuses on robot dispersion. The proposed method is based on a potential field approach that leverages knowledge of the overall bounds of the area to be explored. This additional information allows a simpler potential field control strategy for all robots but which nonetheless has good dispersion and overlap performance in all the multi-robot scenarios while avoiding potential minima. The research presents a path-planning algorithm, the space-based potential field map (SBPF), that uses a potential field map integrated with information about exploration, and allows simultaneous evaluation of multi-robot path-planning. The map is designed to achieve efficient communication and synchronization while still being able to represent the environment for effective robot exploration. In addition, we also model the sensors and robots as realistically as possible, including sensor span and robot heading. Modelling these variables adds more complexity to our calculation, but allows us to transition the simulation to real-robot experiments easily. Lastly, we implement a flexible ad-hoc peer-to-peer communication structure so that more robots can join the exploration after the mission starts. This communication schema also allows robots reconnect after communication interrupts. A set of performance measures was introduced to evaluate the performance of SBPF and compare with that of a similar approach proposed by Renzaglia. A thorough performance evaluation was carried out as part of this research, including a set of simulations using different maps, as well as experimental trials using the Pioneer 3-AT robots with SICK laser sensor. The proposed control strategy was simpler compared to Renzaglia's approach, while still being able to successfully finish the exploration without getting trapped by obstacles or in local minima. The additional robot details we introduced made the robot model more realistic than Renzaglia's more theoretical results, and yet our results were still on par with Renzaglia's. Finally, our flexible framework allows dynamic robot teams, which have improved exploration performance efficiency compared to Renzaglia's static robot teams.
Computer Engineering|Robotics|Computer science
Liu, TsungMing, "Multi-robot exploration with space-based potential field map" (2014). ETD Collection for Fordham University. AAI1568374.