Target Tracking for TarkBot Intelligent Robots by Fusing Lidar and Camera Data

In this article, a target tracking issue is investigated for a TarkBot intelligent robot with a lidar and a camera to identify the state of the target moving around the robot. As the measurement error of the camera increases with distance, a measurement model with multiplicative noise is constructed to describe the distance-related error. Considering the correlation between the disturbance to the target and the measurement noises of sensors caused by complex environment, local estimators for the lidar and the camera are designed in the sense of linear minimum mean square error (LMMSE), respectively. Further, a distributed fusion estimator is constructed and sufficient conditions are provided for ensuring the boundedness of fusion estimation error. Finally, simulation and experiments verify the effectiveness and feasibility of the estimator. Compared with the estimator without considering multiplicative noise, the accuracy of the local estimator of the camera is improved by 61.3%, and the fusion accuracy is improved by 14.7%.