End to end collision avoidance based on optical flow and neural networks

TL;DR

This paper presents a neural network approach utilizing optical flow for collision avoidance, demonstrating its effectiveness on a specially adapted vehicle, inspired by biological flight mechanisms.

Contribution

It introduces a novel neural network method that leverages optical flow for collision avoidance, specifically tested on a custom-fitted car.

Findings

Effective collision avoidance demonstrated on a refitted vehicle.

Optical flow can be successfully integrated with neural networks for real-time navigation.

Biologically inspired approach shows promise for autonomous vehicle safety.

Abstract

Optical flow is believed to play an important role in the agile flight of birds and insects. Even though it is a very simple concept, it is rarely used in computer vision for collision avoidance. This work implements a neural network based collision avoidance which was deployed and evaluated on a solely for this purpose refitted car.