A model-based reinforcement learning algorithm faced with the laser frequency sweep linearization for FMCW LiDAR

Abstract

Artificial intelligence (AI) has been widely used in various fields of physics and engineering. In this work, we introduce the model-based reinforcement learning (MBRL), which is an important branch of machine learning, to the laser frequency sweep control for frequency modualted continous wave (FMCW) light detection and range (LiDAR). With the well designed neural network structure, the experimental results with control represent a significant improvement in the linearity of laser frequency sweep indicating that our proposed MBRL control method has the potential for complicated optical systems control.