LiDARs detected signal and Target distance estimation: measurement errors from Target reflectance and multiple echos

Abstract

The use of LiDARs in automotive is increasingly widespread. In order to operate in a critical environment such as that of mobility, these systems must offer increasingly high performance. In particular, the ability to estimate the position of objects regardless of their reflectance and presence of diffusing backgrounds is a very sought-after feature by manufacturers. In commercial systems various strategies are used to make the measurement as insensitive as possible to these effects, however, it is not possible to fully compensate for the measurement errors caused by them. In this paper, we propose two simple experimental setups to verify the presence of these measurement errors in two scenarios. Moreover, we report the performance of a commercial LiDAR (MRS 6000 by Sick) using certified reflectance standards (Spectralon® Diffuse Reflectance Standards, by Labsphere). For this LiDAR, the results obtained show that a logarithmic variation of the reflectance of the target of 1.26-log at a target distance 2.4 m provides incompatible measurements. Furthermore, the presence of a background at a distance shorter than 11 cm, 12 cm and 13 cm respectively with 50 %, 75 % and 99 % reflectance also provides incompatible measurements for a target distance of 1.3 m from the LiDAR.

A Method for Estimating Object Detection Probability, Lateral Resolution, and Errors in 3-D LiDARs

“…APPENDIX FUNCTIONS RECORDED AT VARIOUS DISTANCES Figs. [13][14][15][16][17][18][19] show the functions obtained from θ n−1 , θ n , and θ n+1 rays as the knife distance r and the angular distance from the optical axis of the ray vary. As expected from Fig.…”

“…Tests were performed in a controlled environment where both temperature and lighting were monitored. Moreover, to avoid the background could affect the results [19], measurements were performed setting d bg > 100 m (see Fig. 5).…”