aiqbench analysis
This research introduces DiffCVaR, a reinforcement learning framework that integrates differentiable Conditional Value-at-Risk barrier functions to enable robots to navigate crowded environments with adaptive risk management under uncertainty. By jointly optimizing nominal control inputs and safety margins, the method achieves superior efficiency and generalization compared to traditional optimization-based or purely RL approaches while maintaining strict probabilistic safety constraints. The work advances the intersection of machine learning and robotics by providing a robust solution for stochastic interactions in dynamic settings.
DiffCVaR: Reinforcement Learning for Risk Adaptation via Differentiable CVaR Barrier Functions
arXiv:2605.21257v2 Announce Type: replace Abstract: Planning through crowded environments under uncertain obstacle motions remains difficult, as stochastic interactions often induce overly conservative behavior or reduced efficiency. To address this challenge, we propose an end-to-end risk adaptation framework for crowd navigation under obstacle-motion uncertainty modeled by a Gaussian mixture model. The framework combines reinforcement learning~(RL) with a differentiable quadratic-program safety layer based on Conditional Value-at-Risk~(CVaR) barrier functions, jointly learning nominal control input, risk level, and safety margin and enforcing explicit probabilistic safety constraints. This design enables context-aware adaptation, promoting efficient behavior while invoking caution only when necessary. We conduct extensive evaluations in dynamic, uncertain, and crowded environments across varying obstacle densities and robot models, and further assess generalization under three out-of-distribution cases. Comparisons across optimization-based, RL-based, and integrated RL and optimization methods are provided, and the proposed method is shown to deliver the strongest overall performance in safety, efficiency, and generalization under uncertainty.
Abstract:Planning through crowded environments under uncertain obstacle motions remains difficult, as stochastic interactions often induce overly conservative behavior or reduced efficiency. To address this challenge, we propose an end-to-end risk adaptation framework for crowd navigation under obstacle-motion uncertainty modeled by a Gaussian mixture model. The framework combines reinforcement learning~(RL) with a differentiable quadratic-program safety layer based on Conditional Value-at-Risk~(CVaR) barrier functions, jointly learning nominal control input, risk level, and safety margin and enforcing explicit probabilistic safety constraints. This design enables context-aware adaptation, promoting efficient behavior while invoking caution only when necessary. We conduct extensive evaluations in dynamic, uncertain, and crowded environments across varying obstacle densities and robot models, and further assess generalization under three out-of-distribution cases. Comparisons across optimization-based, RL-based, and integrated RL and optimization methods are provided, and the proposed method is shown to deliver the strongest overall performance in safety, efficiency, and generalization under uncertainty.
Robotics (cs.RO)
arXiv:2605.21257 [cs.RO]
(or arXiv:2605.21257v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2605.21257 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 238 words
Related coverage
News
Quantum Computing Report · Jul 25, 2026
OQC And Trust Base Study Helps Map The Path to Quantum Advantage in Finance
The Quantum Insider · Sep 17, 2026
[AINews] Reality Checks on AI News (Yegge shuts down Gas Town, Databricks’ +60% Astra cost)
Latent Space · Sep 17, 2026
Safety-Contract Graph Multi-Agent Reinforcement Learning for Autonomous Network Security Response
arXiv · Jun 15, 2026