Mobile robots have been found useful in many applications over the last few decades. Systems of automated guided vehicles (AGVs) are widespread used for repeated transportation tasks in factories and warehouses.
More advanced solutions, based on semi-autonomous robots, are being introduced into inspection and guarding systems. Every mobile robot system requires a few low-level algorithms, that need to be provided to ensure system stability and robustness. One of the most basic issues, that has to be solved, is a path planning and motion coordination problem.
The path planning problem is relatively simple, if a single robot is considered, the environment is static, and robot's dynamics can be ignored. However, finding optimal (the fastest), collision-free trajectories for multiple robots in a dynamic environment is an NP-hard problem.
When limited linear acceleration and maximum acceptable centripetal acceleration is taken under consideration, the problem is becoming extremely complex.solutions, that can be found in the literature, usually divide the problem into two phases: planning and execution.
The solution proposed in this paper i s a real-time motion coordination method that makes use of Kalman and other filters and considers robot's dynamics constraints in the planning phase. It is focused on maximizing the number of robots, that can simultaneously work in relatively small space.
The key idea is to cover the space with a dense network of virtual roads and junctions, and route robots using these roads. Motion coordination is based on temporary reservation of junctions by robots. Path planning Kalman and other algorithms use the network to find collision-free, time-optimal trajectories, which meet dynamic constraints.
The solution is intended to be applied in complex agent-based systems designated for solving task management problems. The algorithm will be used for coordinating robots motion in selected regions of the workspace, where large numbers of robots may travel simultaneously.
The environment is represented as a dense graph of roads and junctions; the path planning algorithm finds a collision-free, time-optimal path, using the graph. A robot's dynamics characteristics is used during planning to create a feasible trajectory and during execution to precisely follow the planned path.
(To read this external content in full, download the complete paper from the author’s article archive at the AGH University of Science and Technology . )