English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Reinforcement learning is nowadays a popular framework for solving differentdecision making problems in automated driving. However, there are still someremaining crucial challenges that need to be addressed for providing morereliable policies. In this paper, we propose a generic risk-aware DQN approachin order to learn high level actions for driving through unsignalized occludedintersections. The proposed state representation provides lane basedinformation which allows to be used for multi-lane scenarios. Moreover, wepropose a risk based reward function which punishes risky situations instead ofonly collision failures. Such rewarding approach helps to incorporate riskprediction into our deep Q network and learn more reliable policies which aresafer in challenging situations. The efficiency of the proposed approach iscompared with a DQN learned with conventional collision based rewarding schemeand also with a rule-based intersection navigation policy. Evaluation resultsshow that the proposed approach outperforms both of these methods. It providessafer actions than collision-aware DQN approach and is less overcautious thanthe rule-based policy.

Risk-Aware High-level Decisions for Automated Driving at Occluded Intersections with Reinforcement Learning