Model predictive control-based target search and tracking using autonomous mobile robot with limited sensing domain

Chang Liu, J. Karl Hedrick

Abstract: Target search and tracking using autonomous robots is important for both civilian and military applications. In this work, we propose a model predictive control (MPC)-based path planning approach for a ground mobile robot to autonomously search and track a moving target. The robot is equipped with a sensor with limited sensing domain (bounded sensing range and angle of view) for target detection. Both target motion and sensor measurement use linear time-invariant models. Due to the limited sensing domain, we utilize a modified Kalman filter to handle the intermittent measurements. Under the MPC framework, the sensing domain is approximated with a bell-shaped differentiable function and is explicitly considered in the optimization problem. To reduce the computation burden of solving MPC, we propose a two-step procedure: it first considers the limited sensing range and computes a reference trajectory, which is then used for solving the original MPC that considers both limited sensing range and angle. The effectiveness of the proposed method is demonstrated by numerical simulations.