摘要
Multi-robot map merging without inter-robot observations, common landmarks, and the knowledge of initial relative poses is a challenging problem due to lack of features or local maxima. This Letter proposes a multi-hypothesis map merging algorithm with sinogram-based particle swarm optimisation to solve the problem. Experimental results show that the proposed algorithm can merge individual robot maps much more accurately than the existing algorithms.
- 出版日期2016-7-7