Highly Automated Learning for Improved Active Safety of Vulnerable Road Users

TL;DR

This paper proposes an autonomous, iterative learning framework for improving traffic participant models in highly automated driving, reducing the need for labeled data through active learning and online adaptation.

Contribution

It introduces a novel three-step autonomous learning process combining detection, novelty detection, and online adaptation for traffic models.

Findings

Effective reduction in labeled data requirements

Improved model accuracy through iterative refinement

Demonstrated online adaptation capability

Abstract

Highly automated driving requires precise models of traffic participants. Many state of the art models are currently based on machine learning techniques. Among others, the required amount of labeled data is one major challenge. An autonomous learning process addressing this problem is proposed. The initial models are iteratively refined in three steps: (1) detection and context identification, (2) novelty detection and active learning and (3) online model adaption.