A LiDAR Odometry for Outdoor Mobile Robots Using NDT Based Scan Matching in GPS-denied environments

Abstract

The three-dimensional localization problem of outdoor mobile robots is a primary issue for autonomous navigation applications, since the accuracy of positioning will affect the incoming mapping and path planning tasks heavily. In this paper, using INS, encoder and laser LiDAR as the main means for pose perception, a laser LiDAR aided odometry approach based on the normal distributions transform(NDT) scan matching approach is proposed to improve the precision of pose estimates for a mobile robot in GPS denied outdoor environments. The 3D kinematic model of the mobile robot is developed so that the IMU and the encoder data can be fused firstly to obtain the initial localization result with dead reckoning technologies. On that basis, the three-dimensional NDT scan matching method is introduced to compute the pose transformation matrix in real-time using the adjacent two depth frames from LiDAR data, which can get fine pose estimates by calibrate the accumulation errors periodically. Experiments results show the correctness and validity of the presented method.