View a PDF of the paper titled Security of Deep Reinforcement Learning for Autonomous Driving: A Survey, by Ambra Demontis and 8 other authors
Abstract:Reinforcement learning (RL) enables agents to learn optimal behaviors through interaction with their environment and has been increasingly deployed in safety-critical applications, including autonomous driving. Despite its promise, RL is susceptible to attacks designed either to compromise policy learning or to induce erroneous decisions by trained agents. Although the literature on RL security has grown rapidly and several surveys exist, existing categorizations often fall short in guiding the selection of appropriate defenses for specific systems. In this work, we present a comprehensive survey of 86 recent studies on RL security, addressing these limitations by systematically categorizing attacks and defenses according to defined threat models and single- versus multi-agent settings. Furthermore, we examine the relevance and applicability of state-of-the-art attacks and defense mechanisms within the context of autonomous driving, providing insights to inform the design of robust RL systems.
Submission history
From: Srishti Gupta [view email]
[v1]
Mon, 12 Dec 2022 18:50:49 UTC (967 KB)
[v2]
Mon, 15 Sep 2025 17:46:22 UTC (722 KB)
[v3]
Thu, 25 Sep 2025 11:51:20 UTC (10,394 KB)
[v4]
Fri, 23 Oct 2026 21:56:25 UTC (10,153 KB)
[v5]
Thu, 19 Nov 2026 11:32:46 UTC (10,153 KB)
![[2212.06123] Security of Deep Reinforcement Learning for Autonomous Driving: A Survey [2212.06123] Security of Deep Reinforcement Learning for Autonomous Driving: A Survey](https://techwavedispatch.com/wp-content/uploads/2026/11/221206123-Security-of-Deep-Reinforcement-Learning-for-Autonomous-Driving-A-1024x597.png)