Steering Multirobot Behavior via Closed-Loop Affine Activation Editing

• URL: http://arxiv.org/abs/2606.11489v1
• Date: Tue, 09 Jun 2026 22:20:07 GMT
• Title: Steering Multirobot Behavior via Closed-Loop Affine Activation Editing
• Abstract summary: CLAE is an inference-time framework for steering the behavior of a frozen policy.<n>It edits intermediate activations while keeping the base policy weights and downstream action head untouched.<n>We show that CLAE can steer individual robot behavior by controlling each robot's velocity profile.
• Score: 16.82661659623732
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Real-world robots need to adapt their behavior beyond the envelope of their pre-trained policy. Policy finetuning or retraining are options, but they risk catastrophic forgetting, degrading the pretrained policy's base performance. To combat this, we introduce CLAE: Closed-Loop Affine Activation Editing, an inference-time framework for steering the behavior of a frozen policy by editing intermediate activations while keeping the base policy weights and downstream action head untouched. CLAE approaches behavior steering as a closed-loop problem whose outputs edit policy activations that adapt online to the robot state, environment, target behavior, and multi-robot context. It trains a sparse autoencoder over frozen-policy activations, selects behavior-relevant latent features via post-hoc probing, and learns a lightweight RL-based steering policy that applies state-dependent affine edits to selected latents during inference. We validate CLAE on a frozen multi-quadrotor navigation policy trained to perform a single task: navigating robots to a set of goal locations while avoiding obstacles. Through extensive simulations and physical tests, we show that while navigating to their goal positions, CLAE can 1. steer individual robot behavior by controlling each robot's velocity profile; 2. coordinate multirobot behavior by preserving a desired formation; and 3. produce entirely new behavior wherein robots are required to reduce their exposure to surveillance cameras in the environment.

Related papers

• ORPA: Online Residual Policy Adaptation for Robot Manipulation Control with Human Feedback [1.711436622393758]
  Online Residual Policy Adaptation (ORPA) is a framework that enables feedback-driven correction of robot actions without modifying the underlying policy parameters.<n>ORPA augments a pretrained control policy with a lightweight, feedback-conditioned module that predicts residual adjustments directly in joint space.<n>We evaluate ORPA on a set of precision-sensitive manipulation tasks using the ALOHA platform, demonstrating improvements in success rate and recovery from small perturbations compared to baseline control policies and rule-based inverse kinematics corrections.
  arXiv  Detail & Related papers  (2026-08-18T03:30:12Z)
• AgilePE: Autonomous UAV Pursuit-Evasion via Self-Play Reinforcement Learning [31.828745787124316]
  We present AgilePE, a complete system for autonomous UAV pursuit-evasion via self-play reinforcement learning.<n> AgilePE integrates agile low-level control, competitive policy optimization, and sim-to-real deployment in a unified framework.<n>For training, we use competitive self-play with Prioritized Fictitious Self-Play (PFSP) and a diversified opponent pool, enabling agents to improve against historical policies.<n>Real-world experiments reproduce pursuit-evasion tactics observed in simulation, including rapid dodging and flanking, and demonstrate interactive two-agent zero-shot deployment.
  arXiv  Detail & Related papers  (2026-08-14T09:41:16Z)
• ENPIRE: Agentic Robot Policy Self-Improvement in the Real World [40.75426390954145]
  ENPIRE is a harness framework for coding agents that instantiates a physical feedback routine with four core modules.<n>Powered by ENPIRE, frontier coding agents can autonomously train a policy to achieve a 99% success rate on challenging, dexterous manipulation tasks.
  arXiv  Detail & Related papers  (2026-06-18T09:21:27Z)
• VAIC: Vision-Guided Humanoid Agile Object Interaction Control via Decoupled Commands [56.029505206531155]
  Vision Guided Agile Interaction Control (VAIC) is a unified framework that bridges the gap between current controllers and humanoid robots.<n>First, a privileged teacher policy masters diverse interaction skills using precise object kinematics and exact environmental states.<n>Second, a deployable student policy distills these capabilities by replacing full body tracking with velocity targets across multiple axes.
  arXiv  Detail & Related papers  (2026-06-08T09:52:55Z)
• TAM: Torque Adaptation Module for Robust Motion Transfer in Manipulation [32.68351241308298]
  In contact-rich, dynamic manipulation, even small motion discrepancies can result in failure to track reference motion.<n>We introduce the Torque Adaptation Module (TAM), a learned module that adapts the torque commands sent to the robot to match the behavior of an ideal robot.
  arXiv  Detail & Related papers  (2026-06-04T14:31:54Z)
• Closed-Loop Action Chunks with Dynamic Corrections for Training-Free Diffusion Policy [52.106797722292896]
  We present DCDP, a Dynamic Closed-Loop Diffusion Policy framework that integrates chunk-based action generation with real-time correction.<n>In dynamic PushT simulations, DCDP improves adaptability by 19% without retraining while requiring only 5% additional computation.
  arXiv  Detail & Related papers  (2026-03-02T15:04:18Z)
• Vision-based Goal-Reaching Control for Mobile Robots Using a Hierarchical Learning Framework [4.014524824655106]
  This paper decomposes the whole system into a set of tightly coupled functional modules.<n>The proposed framework guarantees uniform exponential stability of the actuation system and safety of the whole operation.
  arXiv  Detail & Related papers  (2026-01-02T08:41:47Z)
• CHIP: Adaptive Compliance for Humanoid Control through Hindsight Perturbation [70.5382178207975]
  hIsight Perturbation (CHIP) is a plug-and-play module that enables controllable end-effector stiffness.<n>CHIP is easy to implement and requires neither data augmentation nor additional reward tuning.<n>We show that a generalist motion-tracking controller trained with CHIP can perform a diverse set of forceful manipulation tasks.
  arXiv  Detail & Related papers  (2025-12-16T18:56:04Z)
• Steering Robots with Inference-Time Interactions [0.5801621787540268]
  When a pretrained policy makes errors during deployment, there are limited mechanisms for users to correct its behavior.<n>My research proposes an alternative: keeping pretrained policies frozen as a fixed skill repertoire while allowing user interactions to guide behavior generation at inference time.<n>Specifically, I propose (1) inference-time steering, which leverages user interactions to switch between discrete skills, and (2) task and motion imitation, which enables user interactions to edit continuous motions while satisfying task constraints defined by discrete symbolic plans.
  arXiv  Detail & Related papers  (2025-06-17T07:59:07Z)
• Task and Domain Adaptive Reinforcement Learning for Robot Control [0.34137115855910755]
  We present a novel adaptive agent to dynamically adapt policy in response to different tasks and environmental conditions. The agent is trained using a custom, highly parallelized simulator built on IsaacGym. We perform zero-shot transfer to fly the blimp in the real world to solve various tasks.
  arXiv  Detail & Related papers  (2024-04-29T14:02:02Z)
• Learning Vision-based Pursuit-Evasion Robot Policies [54.52536214251999]
  We develop a fully-observable robot policy that generates supervision for a partially-observable one. We deploy our policy on a physical quadruped robot with an RGB-D camera on pursuit-evasion interactions in the wild.
  arXiv  Detail & Related papers  (2023-08-30T17:59:05Z)
• Nonprehensile Planar Manipulation through Reinforcement Learning with Multimodal Categorical Exploration [8.343657309038285]
  Reinforcement Learning is a powerful framework for developing such robot controllers. We propose a multimodal exploration approach through categorical distributions, which enables us to train planar pushing RL policies. We show that the learned policies are robust to external disturbances and observation noise, and scale to tasks with multiple pushers.
  arXiv  Detail & Related papers  (2023-08-04T16:55:00Z)
• SABER: Data-Driven Motion Planner for Autonomously Navigating Heterogeneous Robots [112.2491765424719]
  We present an end-to-end online motion planning framework that uses a data-driven approach to navigate a heterogeneous robot team towards a global goal. We use model predictive control (SMPC) to calculate control inputs that satisfy robot dynamics, and consider uncertainty during obstacle avoidance with chance constraints. recurrent neural networks are used to provide a quick estimate of future state uncertainty considered in the SMPC finite-time horizon solution. A Deep Q-learning agent is employed to serve as a high-level path planner, providing the SMPC with target positions that move the robots towards a desired global goal.
  arXiv  Detail & Related papers  (2021-08-03T02:56:21Z)
• Neural Dynamic Policies for End-to-End Sensorimotor Learning [51.24542903398335]
  The current dominant paradigm in sensorimotor control, whether imitation or reinforcement learning, is to train policies directly in raw action spaces. We propose Neural Dynamic Policies (NDPs) that make predictions in trajectory distribution space. NDPs outperform the prior state-of-the-art in terms of either efficiency or performance across several robotic control tasks.
  arXiv  Detail & Related papers  (2020-12-04T18:59:32Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.