An improved method to robot´s environment map-building based on the normal distribution transform

According to the problem that the traditional map-building method has high complexity, a map-matching method based on the combination of normal distribution transform (NDT) and environment feature extraction is proposed. The real-time environment information is obtained by two dimension laser sensors, from which the geometric feature is extracted and it will be transformed into normal distribution using NDT method, then the map is matched the by Newton´ s algorithm. The method can reduce the complexity of the traditional NDT process, improving the efficiency of the map-matching. The experiment results demonstrate the effectiveness and superiority of the method.