| On 2011-03-14 at 16:00:00 (Brussels Time) |
Abstract
We study a distributed approach to the path planning problem. We focus on holonomic kinematic motion in a plane with static obstacles. The problem consists in planning the path of a rigid object that has to be transported from an initial to a final location through a constrained path. The planner observes the environment from above through a visual system. We consider the case in which the path covers a large area, such that the planner architecture consists of a wireless network of observer nodes which each can see a portion of the area. A centralized solution is neither robust nor scalable. To overcome these difficulties we propose a fully distributed approach: each observer node locally calculates the part of the path relative to the area that it sees, and communicates to its neighbor the information which allows the execution of the planning. Our goal is to calculate effective paths in a way that is scalable, resource efficient, and robust to calibration and alignment errors. As reference models for the planner nodes, we consider the eye-bot robots, small flying robots that can attach to the ceiling, equipped with an omnidirectional video camera, and a system for wireless communications and for measuring the relative bearing and distance between two robots.
Keywords
path planning, distributed vision system, flying robots swarm, cooperative motion planning, cooperative multi-robot systems
