# Stability-aware Residual Reinforcement Learning Framework for Robotic Manipulator Disturbance Compensation

arXiv: https://arxiv.org/abs/2609.21307 · HTML: https://mlsanity.com/paper/2609.21307/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in RL for robot locomotion and manipulation, one of 15 directions in Deep reinforcement learning methods.
- Builds on 17 of its 33 references that are in the map.
- First posted on arXiv 18 Sep 2026.

## Placed in

- Deep reinforcement learning methods / [RL for robot locomotion and manipulation](https://mlsanity.com/deeprl/rl-for-robot-locomotion-and-manipulation/)

## Builds on (most cited first, 17 of 17)

- [Unsupervised Learning of Visual Features by Contrasting Cluster Assignments](https://mlsanity.com/paper/2006.09882/) (2020)
- [Sinkhorn Distances: Lightspeed Computation of Optimal Transportation Distances](https://mlsanity.com/paper/1306.0895/) (2013)
- [Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World](https://mlsanity.com/paper/1703.06907/) (2017)
- [Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics](https://mlsanity.com/paper/1705.07115/) (2017)
- [Deep Reinforcement Learning for Robotic Manipulation with Asynchronous Off-Policy Updates](https://mlsanity.com/paper/1610.00633/) (2016)
- [Residual Reinforcement Learning for Robot Control](https://mlsanity.com/paper/1812.03201/) (2018)
- [Residual Policy Learning](https://mlsanity.com/paper/1812.06298/) (2018)
- [A predictive safety filter for learning-based control of constrained nonlinear dynamical systems](https://mlsanity.com/paper/1812.05506/) (2018)
- [IndustReal: Transferring Contact-Rich Assembly Tasks from Simulation to Reality](https://mlsanity.com/paper/2305.17110/) (2023)
- [Hybrid Internal Model: Learning Agile Legged Locomotion with Simulated Robot Response](https://mlsanity.com/paper/2312.11460/) (2023)
- [Towards bridging the gap: Systematic sim-to-real transfer for diverse legged robots](https://mlsanity.com/paper/2509.06342/) (2025)
- [DiAReL: Reinforcement Learning with Disturbance Awareness for Robust Sim2Real Policy Transfer in Robot Control](https://mlsanity.com/paper/2306.09010/) (2023)
- [Residual Policy Learning for Vehicle Control of Autonomous Racing Cars](https://mlsanity.com/paper/2302.07035/) (2023)
- [Disturbance Observer-based Robust Control and Its Applications: 35th Anniversary Overview](https://mlsanity.com/paper/1902.09032/) (2019)
- [RL-augmented Adaptive Model Predictive Control for Bipedal Locomotion over Challenging Terrain](https://mlsanity.com/paper/2509.18466/) (2025)
- [Residual MPC: Blending Reinforcement Learning with GPU-Parallelized Model Predictive Control](https://mlsanity.com/paper/2510.12717/) (2025)
- [DOB-Net: Actively Rejecting Unknown Excessive Time-Varying Disturbances](https://mlsanity.com/paper/1907.04514/) (2019)

## Abstract (arXiv)

Although conventional controllers and disturbance observers (DOBs) are the standard for precision tracking in manipulators, they suffer from parameter uncertainty, nonlinear friction, and compound disturbances. This study proposes a residual reinforcement learning DOB framework that pairs an analytical observer with an RL policy. The deterministic baseline operates within a reliable region, whereas the RL policy explicitly targets the residuals that the model cannot capture. To make this compensation disturbance-aware, an estimator network aligns the observation history with a privileged disturbance context, organizing the latent space by disturbance regime and enabling rapid adaptation across disturbance transitions. To guarantee stability, we derived and enforced a state-dependent action bound on the RL policy from an input-to-state stability (ISS) analysis such that the closed loop provably confines the tracking error to a certified envelope for arbitrary policy outputs. Experiments on a 6-DOF manipulator demonstrated consistent improvements in disturbance estimation and tracking, including a 27.8% tracking-error reduction on real hardware under zero-shot sim-to-real transfer and a 38.0% reduction under a base-vibration disturbance that was not observed during training.