arXiv subjects

Darren Chiu

Publications and source records attributed to Darren Chiu.

6 recordsLinked to original sources

arXiv2026

Steering Multirobot Behavior via Closed-Loop Affine Activation Editing

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.

cs.RO↗

arXiv2025

LEARN: Learning End-to-End Aerial Resource-Constrained Multi-Robot Navigation

Nano-UAV teams offer great agility yet face severe navigation challenges due to constrained onboard sensing, communication, and computation. Existing approaches rely on high-resolution vision or compute-intensive planners, rendering them infeasible for these platforms. We introduce LEARN, a lightweight, two-stage safety-guided reinforcement learning (RL) framework for multi-UAV navigation in cluttered spaces. Our system combines low-resolution Time-of-Flight (ToF) sensors and a simple motion planner with a compact, attention-based RL policy. In simulation, LEARN outperforms two state-of-the-art planners by $10\%$ while using substantially fewer resources. We demonstrate LEARN's viability on six Crazyflie quadrotors, achieving fully onboard flight in diverse indoor and outdoor environments at speeds up to $2.0 m/s$ and traversing $0.2 m$ gaps.

cs.RO↗

arXiv2025

Latent Activation Editing: Inference-Time Refinement of Learned Policies for Safer Multirobot Navigation

Reinforcement learning has enabled significant progress in complex domains such as coordinating and navigating multiple quadrotors. However, even well-trained policies remain vulnerable to collisions in obstacle-rich environments. Addressing these infrequent but critical safety failures through retraining or fine-tuning is costly and risks degrading previously learned skills. Inspired by activation steering in large language models and latent editing in computer vision, we introduce a framework for inference-time Latent Activation Editing (LAE) that refines the behavior of pre-trained policies without modifying their weights or architecture. The framework operates in two stages: (i) an online classifier monitors intermediate activations to detect states associated with undesired behaviors, and (ii) an activation editing module that selectively modifies flagged activations to shift the policy towards safer regimes. In this work, we focus on improving safety in multi-quadrotor navigation. We hypothesize that amplifying a policy's internal perception of risk can induce safer behaviors. We instantiate this idea through a latent collision world model trained to predict future pre-collision activations, thereby prompting earlier and more cautious avoidance responses. Extensive simulations and real-world Crazyflie experiments demonstrate that LAE achieves statistically significant reduction in collisions (nearly 90% fewer cumulative collisions compared to the unedited baseline) and substantially increases the fraction of collision-free trajectories, while preserving task completion. More broadly, our results establish LAE as a lightweight paradigm, feasible on resource-constrained hardware, for post-deployment refinement of learned robot policies.

cs.RO↗

arXiv2024

Collective Bayesian Decision-Making in a Swarm of Miniaturized Robots for Surface Inspection

Robot swarms can effectively serve a variety of sensing and inspection applications. Certain inspection tasks require a binary classification decision. This work presents an experimental setup for a surface inspection task based on vibration sensing and studies a Bayesian two-outcome decision-making algorithm in a swarm of miniaturized wheeled robots. The robots are tasked with individually inspecting and collectively classifying a 1mx1m tiled surface consisting of vibrating and non-vibrating tiles based on the majority type of tiles. The robots sense vibrations using onboard IMUs and perform collision avoidance using a set of IR sensors. We develop a simulation and optimization framework leveraging the Webots robotic simulator and a Particle Swarm Optimization (PSO) method. We consider two existing information sharing strategies and propose a new one that allows the swarm to rapidly reach accurate classification decisions. We first find optimal parameters that allow efficient sampling in simulation and then evaluate our proposed strategy against the two existing ones using 100 randomized simulation and 10 real experiments. We find that our proposed method compels the swarm to make decisions at an accelerated rate, with an improvement of up to 20.52% in mean decision time at only 0.78% loss in accuracy.

cs.RO↗

arXiv2024

SAFE-GIL: SAFEty Guided Imitation Learning for Robotic Systems

Behavior cloning (BC) is a widely-used approach in imitation learning, where a robot learns a control policy by observing an expert supervisor. However, the learned policy can make errors and might lead to safety violations, which limits their utility in safety-critical robotics applications. While prior works have tried improving a BC policy via additional real or synthetic action labels, adversarial training, or runtime filtering, none of them explicitly focus on reducing the BC policy's safety violations during training time. We propose SAFE-GIL, a design-time method to learn safety-aware behavior cloning policies. SAFE-GIL deliberately injects adversarial disturbance in the system during data collection to guide the expert towards safety-critical states. This disturbance injection simulates potential policy errors that the system might encounter during the test time. By ensuring that training more closely replicates expert behavior in safety-critical states, our approach results in safer policies despite policy errors during the test time. We further develop a reachability-based method to compute this adversarial disturbance. We compare SAFE-GIL with various behavior cloning techniques and online safety-filtering methods in three domains: autonomous ground navigation, aircraft taxiing, and aerial navigation on a quadrotor testbed. Our method demonstrates a significant reduction in safety failures, particularly in low data regimes where the likelihood of learning errors, and therefore safety violations, is higher. See our website here: https://y-u-c.github.io/safegil/

cs.RO↗

arXiv2023

Optimization and Evaluation of Multi Robot Surface Inspection Through Particle Swarm Optimization

Robot swarms can be tasked with a variety of automated sensing and inspection applications in aerial, aquatic, and surface environments. In this paper, we study a simplified two-outcome surface inspection task. We task a group of robots to inspect and collectively classify a 2D surface section based on a binary pattern projected on the surface. We use a decentralized Bayesian decision-making algorithm and deploy a swarm of miniature 3-cm sized wheeled robots to inspect randomized black and white tiles of $1m\times 1m$. We first describe the model parameters that characterize our simulated environment, the robot swarm, and the inspection algorithm. We then employ a noise-resistant heuristic optimization scheme based on the Particle Swarm Optimization (PSO) using a fitness evaluation that combines decision accuracy and decision time. We use our fitness measure definition to asses the optimized parameters through 100 randomized simulations that vary surface pattern and initial robot poses. The optimized algorithm parameters show up to a 55% improvement in median of fitness evaluations against an empirically chosen parameter set.

cs.RO↗