Multi-Robot Motion Planning: A Modified Receding Horizon Approach for Reaching Goal States

Abstract

Abstract. This paper proposes the real-time implementation of an algorithm for collision-free motion planning based on a receding horizon approach, for the navigation of a team of mobile robots in the presence of obstacles of different shapes. The method is simulated with three robots. The impact of the parameters is studied with regard to computation time, obstacle avoidance and travel time.

Real-time distributed receding horizon motion planning and control for mobile multi-robot dynamic systems

“…In the DRHMP in [10], each robot in the team computes its own local trajectory. Analogous to an MPC, a prediction time-horizon T p and an implementation/computation timeslot T c are defined (Fig.…”

“…Differently from [10], [11], we do not consider that all robots involved in a conflict have finished Step 1 and exchanged information when any of them starts executing Step 2. Here, the robot estimates trajectories for the conflictual robots based on the available information at the end of Step 1.…”

“…The work presented here builds directly on the mathematical programming approach presented in [10], called DRHMP. The aim is to extend that approach so it can take into account dynamics constraints.…”

Experimental Validation of a Multirobot Distributed Receding Horizon Motion Planning Approach

“…It takes the desired final configuration as a hard constraint in the optimization problem and uses the time for reaching the goal as the cost function to be minimized. Refer to works [12], [21] for more details.…”

“…Instead of a complete geometric description of obstacles as used in previous works [12], [21], only 2D costmaps are available for the purpose of obstacle avoidance when using ROS navigation stack. These costmaps are based on occupancy grids generated from 3D sensor data and user defined inflation radius that is meant to take the robot's footprint into account.…”