A Model Predictive Control for Dynamic Obstacle Avoidance of a Mobile Robot Based on ROS

Abstract

This paper presents a Model Predictive Control (MPC) for mobile robots to avoid dynamic obstacle in an indoor environment. The model is implemented using the Robot Operating System (ROS) platform and Gazebo simulation environment. Conventionally, Wheeled Mobile Robots (WMRs) in the practical world require flexibility, safety, and potential impact avoidance. MPC was used for local planning combined with A * global planning to generate the trajectory. Dynamic obstacle avoidance MPC based on dynamic obstacle action prediction includes two stages object motion prediction and motion planning. In particular, MPC avoided singularities to increase the robot's reliability. In addition, the A * algorithm is used to find the optimal path from the start point to the destination point with the static map. Simulation results in Gazebo demonstrate the robot's ability to track the trajectory while smoothly avoiding obstacles.