Aiming at the problem of low efficiency and poor real-time performance of traditional iterative closest point（ ICP） algorithm in radar data matching

a fast positioning method based on linearized expression of indoor lidar point cloud data is proposed. The point cloud data are segmented and linearly fitted

and the length

orientation angle and centroid coordinates of the line segments are extracted as the linear expression information of the environment. The main orientation of indoor environment is extracted from linear information to achieve rough rotation matching. On this basis

line segment matching is carried out

and the rotation angle is matched accurately. Compared with the traditional ICP positioning method

this algorithm needs less data to be processed

the positioning speed is obviously improved

the average time used is 35 % of the traditional algorithm and the matching precision is close to the traditional ICP algorithm. The rotation angle deviation of the two algorithms is less than 1°

and the translation deviation is less than 1 cm. That is to say

the algorithm can improve the operation efficiency under the condition that the precision is unchanged.