Crash course learning: an automated approach to simulation-driven LiDAR-basedtraining of neural networks for obstacle avoidance in mobile robotics

Authors

STANKO KRUZIC
JOSIP MUSIC
MIRJANA BONKOVIC
FRANTISEK DUCHON

DOI

10.3906/elk-1907-112

Abstract

This paper proposes and implements a self-supervised simulation-driven approach to data collection used for training of perception-based shallow neural networks for mobile robot obstacle avoidance. In the approach, a 2D LiDAR sensor was used as an information source for training neural networks. The paper analyzes neural network performance in terms of numbers of layers and neurons, as well as the amount of data needed for reliable robot operation. Once the best architecture is identified, it is trained using only data obtained in simulation and then implemented and tested on a real robot (Turtlebot 2) in several simulations and real-world scenarios. Based on obtained results it is shown that this fast and simple approach is very powerful with good results in a variety of challenging environments, with both static and dynamic obstacles.

Keywords

Autonomous mobile robots, obstacle avoidance, neural networks, simulation-based learning

First Page

1107

Last Page

1120

Recommended Citation

KRUZIC, STANKO; MUSIC, JOSIP; BONKOVIC, MIRJANA; and DUCHON, FRANTISEK (2020) "Crash course learning: an automated approach to simulation-driven LiDAR-basedtraining of neural networks for obstacle avoidance in mobile robotics," Turkish Journal of Electrical Engineering and Computer Sciences: Vol. 28: No. 2, Article 35. https://doi.org/10.3906/elk-1907-112
Available at: https://journals.tubitak.gov.tr/elektrik/vol28/iss2/35