Learning Actions for Drift-Free Navigation in Highly Dynamic Scenes

TL;DR

This paper introduces a deep reinforcement learning approach to enable autonomous vehicles to navigate in highly dynamic scenes with minimal localization drift, using LIDAR data and simulation-based training.

Contribution

It presents a novel method for learning navigation actions that reduce drift in dynamic environments, leveraging deep reinforcement learning with LIDAR sensors.

Findings

Significant reduction in localization drift compared to baseline methods.

Effective learning of drift-minimizing actions in synthetic scenes.

Demonstrated superiority over non-policy-based approaches.

Abstract

We embark on a hitherto unreported problem of an autonomous robot (self-driving car) navigating in dynamic scenes in a manner that reduces its localization error and eventual cumulative drift or Absolute Trajectory Error, which is pronounced in such dynamic scenes. With the hugely popular Velodyne-16 3D LIDAR as the main sensing modality, and the accurate LIDAR-based Localization and Mapping algorithm, LOAM, as the state estimation framework, we show that in the absence of a navigation policy, drift rapidly accumulates in the presence of moving objects. To overcome this, we learn actions that lead to drift-minimized navigation through a suitable set of reward and penalty functions. We use Proximal Policy Optimization, a class of Deep Reinforcement Learning methods, to learn the actions that result in drift-minimized trajectories. We show by extensive comparisons on a variety of synthetic, yet photo-realistic scenes made available through the CARLA Simulator the superior performance of the proposed framework vis-a-vis methods that do not adopt such policies.