MEMS-Scanner Testbench for High Field of View LiDAR Applications

Valentin Baier,Martin Jakobi,Maximilian Fink,Alexander W Koch,Michael Schardt

doi:10.3390/s22010039

Abstract

LiDAR sensors are a key technology for enabling safe autonomous cars. For highway applications, such systems must have a long range, and the covered field of view (FoV) of >45° must be scanned with resolutions higher than 0.1°. These specifications can be met by modern MEMS scanners, which are chosen for their robustness and scalability. For the automotive market, these sensors, and especially the scanners within, must be tested to the highest standards. We propose a novel measurement setup for characterizing and validating these kinds of scanners based on a position-sensitive detector (PSD) by imaging a deflected laser beam from a diffuser screen onto the PSD. A so-called ray trace shifting technique (RTST) was used to minimize manual calibration effort, to reduce external mounting errors, and to enable dynamical one-shot measurements of the scanner’s steering angle over large FoVs. This paper describes the overall setup and the calibration method according to a standard camera calibration. We further show the setup’s capabilities by validating it with a statically set rotating stage and a dynamically oscillating MEMS scanner. The setup was found to be capable of measuring LiDAR MEMS scanners with a maximum FoV of 47° dynamically, with an uncertainty of less than 1%.

Highlights

To achieve the derived merits, we propose a novel measurement technique of characterizing achieve the derived merits, we propose a novel measurement technique of characterizing large-field of view (FoV) mechanical systems (MEMS) scanners by using a projection screen and a so-called ray trace shifting large-FoV MEMS scanners by using a projection screen and a so-called ray trace shifting technique of a position-sensitive detector (PSD) camera setup
This paper introduces a novel approach in dynamically measuring and characterizing large-FoV MEMS scanners, especially for LiDAR sensors
Camera, the steering angle of the scanner can be calculated by measuring the movement of a projected laser spot on an optical diffuser screen

Summary

Introduction

Autonomous vehicles are the step in transportation. To achieve this goal, cars must be equipped with sensors that perceive the environment in detail and accurately. Modern setups ensure safe driving by combining perception sensors such as cameras, radars, and especially LiDARs. LiDAR sensors in particular are one of the key enablers of automotive driving since their output provides direct distance information on every object in the scanned field of view (FoV). LiDAR (short for light detection and ranging) sensors are most commonly used as time-of-flight sensors. They can determine the distance to any object they face by measuring the time that an emitted laser pulse takes to arrive at an object, be scattered by it, and return to the sensor and be detected.

Methods

Results

Discussion

Conclusion