A Neural Network-based Method for Fast Angle Calibration of MEMS LiDAR

MEMS LiDAR has been drawing more and more attention as a representative achievement in low power consumption and miniaturization integration of LiDAR research. To address the problem that traditional angle calibration methods for mechanical LiDAR are not completely applicable in MEMS LiDAR due to different sources of angle errors, a neural network-based method for fast angle calibration of MEMS LiDAR is proposed. The calibration field only consists of a rotating platform and an angular reflector. By modelling the calibration field and analyzing the error sources of MEMS LiDAR, including the mounting error, the nonlinear error, and the systematic error, a neural network model was then developed to correct the angle measurement errors. The experimental results show that the method is reasonable and effective and is able to achieve fast angle calibration of MEMS LiDAR.