A rapid impassable terrain recognition algorithm for mobile robot based on hypotheses testing theory

The recognition of impassable terrain in three dimension environment plays an important role in the motion planning of mobile robot. However, the most common used sensor like range finders, stereo vision sensors, which can provide the clues of the surrounding environment suffer from their inaccurate nature. Thus, this paper proposed a rapid impassable terrain recognition algorithm based on hypotheses testing theory using low-cost range finders with different diffusion angle. By applying a slope recognition algorithm using two different range finders mounted with different height, slope of object is capable to be recognized. By treating the historical recognition result as a sampling set to the real slope, the hypothesis testing theory is then introduced, with which a judge threshold value can be induced to decrease the influence of inaccurate recognition. Then the stable recognition for impassable terrain would be then acquired. The experiments results indicate that this algorithm can provide a reliable recognition for the impassable terrain based on lower cost range finders with different diffusion angle while only little computation is required.