An Integrated XAI-RL Architecture for Ethical Decision-Making in Autonomous Robots
DOI:
https://doi.org/10.54878/h0csbb74Keywords:
Explainable AI (XAI), Reinforcement Learning, Robot Ethics, Transparent Autonomy, Ethical Decision-Making, Human-Robot Interaction, Value Alignment, Proximal Policy OptimizationAbstract
Can a robot's ethical reasoning be transparent without slowing it down? That is the question we set out to answer. Autonomous robots now make decisions that affect people's lives, and if we cannot understand how they decide, trust becomes a serious problem. We built a unified architecture that couples Explainable AI (XAI) with Reinforcement Learning (RL) through three connected components: a Perception Layer for situational awareness, a Decision Reasoning Layer that breaks agent utility down into measurable ethical sub-rewards (safety, fairness, and well-being), and a contrastive Explainability Engine grounded in a real-time Ethical Ledger. We tested it against Black-Box RL and two post-hoc XRL baselines (LIME and SHAP) across two simulation environments and a physical robot pilot on a TurtleBot3 platform. Our system reached 92±3% ethical alignment in simulation and 89±4% on the physical platform, beating every baseline (p < 0.01 throughout). In a controlled, IRB-approved user study (N=128), participants rated it higher on trust (4.5±0.4), understandability (4.6±0.3), and perceived safety (4.4±0.5) than the alternatives. Explanations were also fast: 4.2±0.6 ms on average, versus 18.3±2.1 ms for SHAP and 12.7±1.8 ms for LIME on the same hardware—well within real-time limits. An ablation study and sensitivity analysis back up these architectural choices.
References
Adadi, A., & Berrada, M. (2024). Explainable AI for trustworthy autonomous systems: A roadmap. Nature Machine Intelligence, 6(1), 12–25.
Alaieri, A., & Vellino, A. (2017). Ethical decision making in robots: A conceptual model and analysis. In Proceedings of We Robot Conference. New Haven, CT, USA.
Cruz, F., Dazeley, R., & Vamplew, P. (2022). Explainable reinforcement learning: A survey and comparative review. ACM Computing Surveys, 55(5), 1–39.
Dazeley, R., Vamplew, P., & Cruz, F. (2021). Explainable AI for robotics: A review. In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA). Xi’an, China.
Dazeley, R., Vamplew, P., & Cruz, F. (2023). Explainable reinforcement learning for complex autonomous systems. In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). Detroit, MI, USA.
Gronauer, P., & Diepold, K. (2021). Human-centered AI for multi-agent deep reinforcement learning: Challenges and opportunities. In Proceedings of the AAAI Conference on Artificial Intelligence. Virtual.
Hellström, M., & Bensch, K. (2020). Explainable robotics: A primer. In Proceedings of the ACM/IEEE International Conference on Human-Robot Interaction (HRI). Cambridge, UK.
Krakovna, T., et al. (2024). Value alignment in human-robot teams: A survey. Autonomous Robots, 48(2), 101–124.
Leurent, E. (2018). Highway-env: An environment for autonomous driving decision-making [Computer software]. GitHub. https://github.com/eleurent/highway-env
Lundberg, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. In Advances in Neural Information Processing Systems (Vol. 30, pp. 4765–4774). Long Beach, CA, USA.
Millar, J. (2016). An ethics point of view: How to design for ethics in robotics. In Proceedings of We Robot Conference. Miami, FL, USA.
Pynadath, D. V., Barnes, M. J., & Wang, N. (2022). Explainable reinforcement learning for human-robot teamwork. Frontiers in Robotics and AI, 9, Article 751592.
Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). “Why should I trust you?”: Explaining the predictions of any classifier. In Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 1135–1144). San Francisco, CA, USA. https://doi.org/10.1145/2939672.2939778
Schulman, J., Wolski, F., Dhariwal, P., Radford, A., & Klimov, O. (2017). Proximal policy optimization algorithms (Version 2). arXiv. https://arxiv.org/abs/1707.06347v2
Theodorou, A., Wortham, R. H., & Bryson, J. J. (2016). Designing and implementing transparency for autonomous systems. In Proceedings of Towards Autonomous Robotic Systems (TAROS) (pp. 243–255). Springer. https://doi.org/10.1007/978-3-319-49274-2_22
Vanderelst, D., & Winfield, A. (2016). The ethical governor: A module for ethical decision-making in robots. In Proceedings of the European Robotics Forum (pp. 1–6). Ljubljana, Slovenia.
Wallach, W., & Allen, C. (2008). Moral machines: Teaching robots right from wrong. Oxford University Press.
Wallach, W., & Allen, C. (2019). On proactive, verifiable, and ethical robotic systems. In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA). Montreal, QC, Canada.
Wang, C., Li, Z., & Liang, B. (2024). Explainable AI in deep reinforcement learning for robotic manipulation. In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA). Yokohama, Japan.
Wang, N., Pynadath, D. V., & Hill, S. G. (2016a). Trust calibration within a human-robot team: Comparing automatically generated explanations. In Proceedings of the ACM/IEEE International Conference on Human-Robot Interaction (HRI) (pp. 109–116). Christchurch, New Zealand. https://doi.org/10.1109/HRI.2016.7451741
Wang, N., Pynadath, D. V., & Hill, S. G. (2016b). Transparency and trust in human-robot interaction. In Proceedings of the IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN). New York, NY, USA.
Yuan, L., Gao, J. H., & Zhang, Z. (2022). A framework for bidirectional human-robot value alignment through explainable AI. IEEE Transactions on Human-Machine Systems, 52(4), 784–796.
