Machine learning-aided LiDAR range estimation

Abstract

Automotive light detection and ranging (LiDAR) requires accurate and computationally efficient range estimation methods. At present, such efficiency is achieved at the cost of curtailing the dynamic range of a LiDAR receiver. In this Letter, we propose using decision tree ensemble machine learning models to overcome such a trade-off. Simple and yet powerful models are developed and proven capable of performing accurate measurements across a 45-dB dynamic range.