Vision Based SLAM for Mobile Robot Navigation Using Distributed Filters

摘要

In this chapter, a vision based SLAM and encoder integrated system is presented for mobile robot navigation. By considering the nonlinear measurement model and feature point availability around the trajectory, a distributed particle filter approach is applied. Simulation results demonstrate the performance of the implemented mobile robot. Further results confirm that the estimation performance largely depends on the number of feature points and particles, which will be mutually associated while implementing the embedded navigation computer. It also depends on the attached angle of a vision sensor and the landmark.