mlsanityData to 7 Oct 2026

Securing Autonomous Vehicle Systems via Twin-Aware Federated Reinforcement Learning

arXiv 2607.08137 ↗First posted 9 Jul 2026cs.CR cs.DC cs.LGcited by 0

On the map: End-to-end autonomous driving → Adversarial scenarios and safety testing

Abstract (arXiv)

Federated reinforcement learning (FRL) is crucial for enabling collaborative learning across multiple agents without sharing raw data, thereby enhancing privacy and scalability in the decision-making process within dynamic vehicular environments. However, poisoning attacks pose a significant threat to the security and reliability of FRL-based systems, particularly in safety-critical autonomous driving, where this vulnerability remains largely unexplored. These attacks can compromise the global control model by subtly injecting malicious system parameters, leading to potential hazards. To counter these challenges, we present alg (underlineSecure underlineAggregation with underlinepoisoning-underlineprevention and historical reinforcement) as a defensive framework aimed at enhancing the robustness of FRL systems designed for safety-critical driving scenarios. alg strategically integrates digital twins for rehearsal-based learning and leverages historical aggregated model parameters along with a selected central gradient to ensure that only benign data is aggregated, effectively mitigating the influence of malicious agents. Theoretical guarantees are provided for the convergence performance of alg in the presence of poisoning attacks. We also validate the effectiveness of alg using developed digital twins that model realistic highway environments to evaluate the control of autonomous vehicles under adversarial conditions.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Adversarial scenarios and safety testing, one of 9 directions in End-to-end autonomous driving. Builds on 35 of its 69 references that are in the map. First posted on arXiv 9 Jul 2026.

Builds on

35 of 69 references.

References · 35

  1. 2023GPT-4 Technical Reportcited by 18,132
  2. 2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,592
  3. 2020Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problemscited by 1,873
  4. 2018How To Backdoor Federated Learningcited by 745
  5. 2018Byzantine-Robust Distributed Learning: Towards Optimal Statistical Ratescited by 698
  6. 2020Deep Reinforcement Learning for Autonomous Driving: A Surveycited by 698
  7. 2019Local Model Poisoning Attacks to Byzantine-Robust Federated Learningcited by 479
  8. 2020FLTrust: Byzantine-robust Federated Learning via Trust Bootstrappingcited by 330
  9. 2019A Little Is Enough: Circumventing Defenses For Distributed Learningcited by 285
  10. 2018Learning to Drive in a Daycited by 266
  11. 2018Mitigating Sybils in Federated Learning Poisoningcited by 158
  12. 2016Stochastic Variance Reduction for Nonconvex Optimizationcited by 143
23 more
  1. 2022Imitation Is Not Enough: Robustifying Imitation with Reinforcement Learning for Challenging Driving Scenarioscited by 124
  2. 2021Back to the Drawing Board: A Critical Evaluation of Poisoning Attacks on Production Federated Learningcited by 116
  3. 2023Learning Agile Soccer Skills for a Bipedal Robot with Deep Reinforcement Learningcited by 111
  4. 2018Stochastic Variance-Reduced Policy Gradientcited by 103
  5. 2018Byzantine Stochastic Gradient Descentcited by 81
  6. 2021Federated Reinforcement Learning: Techniques, Applications, and Open Challengescited by 79
  7. 2022DeepSight: Mitigating Backdoor Attacks in Federated Learning Through Deep Model Inspectioncited by 78
  8. 2016Variance Reduction for Faster Non-Convex Optimizationcited by 72
  9. 2021BACKDOORL: Backdoor Attack against Competitive Reinforcement Learningcited by 66
  10. 2022MPAF: Model Poisoning Attacks to Federated Learning based on Fake Clientscited by 65
  11. 2017Non-convex Finite-Sum Optimization Via SCSG Methodscited by 61
  12. 2022Federated Reinforcement Learning with Environment Heterogeneitycited by 61
  13. 2019An Improved Convergence Analysis of Stochastic Variance-Reduced Policy Gradientcited by 58
  14. 2021Fault-Tolerant Federated Reinforcement Learning with Theoretical Guaranteecited by 52
  15. 2020Resource Management in Wireless Networks via Multi-Agent Deep Reinforcement Learningcited by 41
  16. 2019Federated Deep Reinforcement Learningcited by 34
  17. 2023BadRL: Sparse Targeted Backdoor Attack Against Reinforcement Learningcited by 29
  18. 2024BackdoorIndicator: Leveraging OOD Data for Proactive Backdoor Detection in Federated Learningcited by 17
  19. 2024Poisoning Attacks on Federated Learning-based Wireless Traffic Predictioncited by 10
  20. 2024ParallelSFL: A Novel Split Federated Learning Framework Tackling Heterogeneity Issuescited by 10
  21. 2024Digital Twin-Assisted Data-Driven Optimization for Reliable Edge Caching in Wireless Networkscited by 8
  22. 2024Digital Network Twins for Next-generation Wireless: Creation, Optimization, and Challengescited by 8
  23. 2024Deep Reinforcement Learning for Advanced Longitudinal Control and Collision Avoidance in High-Risk Driving Scenarioscited by 4

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2022Security of Deep Reinforcement Learning for Autonomous Driving: A Survey
  2. 2025FuncPoison: Poisoning Function Library to Hijack Multi-agent Autonomous Driving Systems
  3. 2025AED: Automatic Discovery of Effective and Diverse Vulnerabilities for Autonomous Driving Policy with Large Language Models
  4. 2025Robust Driving Control for Autonomous Vehicles: An Intelligent General-sum Constrained Adversarial Reinforcement Learning Approach
  5. 2025MARL-OT: Multi-Agent Reinforcement Learning Guided Online Fuzzing to Detect Safety Violation in Autonomous Driving Systems
  6. 2026Sparse Threats, Focused Defense: Criticality-Aware Robust Reinforcement Learning for Safe Autonomous Driving
  7. 2023Safety-aware Causal Representation for Trustworthy Offline Reinforcement Learning in Autonomous Driving
  8. 2026ADV-0: Closed-Loop Min-Max Adversarial Training for Long-Tail Robustness in Autonomous Driving