Abnormality detection and robust perception model for laser range finder

Laser range finder has been widely used in map building, locating, and tracking in robotics. Accurate and robust perception model is critical to achieve accurate locating, consistent mapping and effective tracking. In particular, the perception model is a key factor for weighting in particle filters (or sequential Monte Carlo) methods that are widely employed in many tasks for mobile robots. A robust perception model for laser range finder is presented in the paper for these purposes. A dasiareverse ray tracingpsila method is put forward to compute expectation distance of beams quickly. The robustness of the provided model is two folds, (1) error beams of laser range finder are filtered out with segment analysis method, (2) beams occluded by dynamical objects are also picked out. Experimental results show the robustness and accuracy of the presented model.