SimuGAN: Unsupervised forward modeling and optimal design of a LIDAR Camera

TL;DR

This paper introduces SimuGAN, an unsupervised generative model that simulates a LIDAR camera's behavior, enabling optimization of its parameters for improved depth accuracy and stability despite noisy signals.

Contribution

The work presents a novel unsupervised forward model of a LIDAR camera using GANs, allowing for differentiable simulation and parameter optimization.

Findings

Successfully models LIDAR behavior on synthetic and real data.

Improves depth estimation accuracy and stability.

Introduces a custom loss function for back-scattered pulse distribution.

Abstract

Energy-saving LIDAR camera for short distances estimates an object's distance using temporally intensity-coded laser light pulses and calculates the maximum correlation with the back-scattered pulse. Though on low power, the backs-scattered pulse is noisy and unstable, which leads to inaccurate and unreliable depth estimation. To address this problem, we use GANs (Generative Adversarial Networks), which are two neural networks that can learn complicated class distributions through an adversarial process. We learn the LIDAR camera's hidden properties and behavior, creating a novel, fully unsupervised forward model that simulates the camera. Then, we use the model's differentiability to explore the camera parameter space and optimize those parameters in terms of depth, accuracy, and stability. To achieve this goal, we also propose a new custom loss function designated to the back-scattered code distribution's weaknesses and its circular behavior. The results are demonstrated on both synthetic and real data.