Target tracking of cloud robot in intelligent space

This paper presents a new method for target tracking based on cloud robot in intelligent space. First, a multi-level information fusion structure using the distributed data fusion tree is designed to realize the real-time target and robot localization. A modified A* weighting algorithm is employed to conduct path planning on dynamic danger degrees map for searching target. Due to uncertain transform delayed in a certain range, time inconsistency from different sensors is obvious. To solve the problem, multi-source heterogeneous data fusion through the key frame collaborative detection and alignment is implemented. Finally, by introducing the cloud coefficient, the intelligent space and cloud robot are combined to reduce the service execution time, to increase the tracking success rate, and to share knowledge between robots. Simulation results show the effectiveness of the proposed method.