Search papers, labs, and topics across Lattice.
This paper integrates adaptive experience replay and online uncertainty estimation within a safe actor-critic framework, addressing the limitations of treating these components as isolated modules. The proposed architecture enhances the control barrier function by dynamically updating obstacle geometry based on uncertainty estimates, while prioritizing replay based on intervention filtering and estimation residuals. In a challenging robot-navigation task with corrupted obstacle measurements, the integrated approach achieved zero contacts and consistently reached the goal across multiple evaluation seeds, outperforming configurations that relied solely on uncertainty estimation.
An integrated architecture for safe actor-critic control achieved flawless navigation in extreme conditions, demonstrating the power of coupling uncertainty estimation with experience replay.
Safe actor-critic control often treats barrier filtering, uncertainty estimation, and experience replay as separate modules, even though each changes the data used for learning and control. We develop an integrated architecture in which the uncertainty estimate updates the obstacle geometry used by a control barrier function, filter interventions and estimation residuals determine replay priority, and the critic learns from the executed rather than nominal action. We instantiate the architecture on a two-dimensional robot-navigation task with corrupted obstacle measurements and compare six component-matched configurations under common training budgets, random seeds, sensor streams, exploration, and disturbances. Evaluation includes a moderate post-training test, an eleven-level perception-noise sweep, and an exploratory extreme-stress test at multiplier $6.0$. In the extreme test, the integrated configuration recorded no contacts and reached the goal in all five evaluation seeds. Its mean cost was $7.63\pm0.44$ and its obstacle-belief root-mean-square error was $3.52\pm0.55$ cm. The uncertainty-estimation ablation also recorded no contacts but reached the goal in four of five seeds, with mean cost $8.96\pm2.08$ and belief error $11.08\pm1.23$ cm. A finite-training bound clarifies replay exposure, and a robust barrier condition states the required estimation-error and feasibility assumptions. The results support coupling estimation, safety filtering, and replay on this benchmark; broader safety and convergence claims require further study.
