Pedestrian-following service robot applications using chance-constrained target tracking

This paper presents a mobile robot system where the robot tracks a moving target. The system minimizes the probability of losing the target. If the next position of a moving target has the Gaussian distribution, the proposed system guarantees the tracking success probability. In addition, we minimize the moving distance of the mobile robot based on the chosen bound on the tracking success probability. We built a well-designed system on Robot Operating System for applications such as a smart shopping cart, selfie robot, and transporter. The performance of the proposed system is extensively evaluated in field experiments.