Research On Cooperative Multi-robot Map-building Under Unknown Environment

In this paper, firstly, multi-robot map building problem in a complex and unknown environment is investigated, presents a map building approach based on particle swarm optimization algorithm for global optimization, as well as Hilbert curve on the target region detection of multi-robot collaborative. Experiment of comparing with S shape random exploring algorithm shows that this method will enable the robot to find the approximate optimal target area, reduce the probability of duplicate detection, and improve the efficiency of detection. Second, for dynamic target of trajectory prediction, this paper describes two kinds of trajectory prediction algorithm, a trajectory prediction algorithm based on Kalman filter. One is the Monte Carlo algorithm based on trajectory prediction, using this method to be Gaussian distribution to approximate location of the posterior distribution of state variables. And discusses the algorithm for maneuvering target turning movement in the non-linear trajectory prediction problem, under the flicker noise, a comparison of the two experimental results show that the method does not need to repeat sampling, in filtering accuracy and computing time, the is also superior to other algorithms.