CEASE: Collision-Evaluation-based Active Sense System for Collaborative Robotic Arms

TL;DR

The paper tackles collision avoidance for collaborative robotic arms by replacing stationary external sensing with CEASE, an active-sense framework that uses rotatable RGB-D cameras to widen observation of dynamic obstacles. It introduces the State Confidence Envelope (SCE) to quantify obstacle state certainty, the Observation-based Uncertainty Evolution (OUE) law to predict how uncertainty evolves with observations, and the Collision-free Optimal Observation Sequence (COOS) to plan the best sequence of vision actions via an MDP. The approach yields substantial gains in temporal observation coverage of dynamic humanoid obstacles, demonstrated by up to 168% improvements in simulation and robust real-world obstacle avoidance where occlusions hinder fixed-camera setups. Together, CEASE reduces collision risk and enhances safety for human-robot collaboration, with practical deployment benefits due to minimal installation requirements and ROS-based integration.

Abstract

Collision detection via visual fences can significantly enhance the safety of collaborative robotic arms. Existing work typically performs such detection based on pre-deployed stationary cameras outside the robotic arm's workspace. These stationary cameras can only provide a restricted detection range and constrain the mobility of the robotic system. To cope with this issue, we propose an active sense method enabling a wide range of collision risk evaluation in dynamic scenarios. First, an active vision mechanism is implemented by equipping cameras with additional degrees of rotation. Considering the uncertainty in the active sense, we design a state confidence envelope to uniformly characterize both known and potential dynamic obstacles. Subsequently, using the observation-based uncertainty evolution, collision risk is evaluated by the prediction of obstacle envelopes. On this basis, a Markov decision process was employed to search for an optimal observation sequence of the active sense system, which enlarges the field of observation and reduces uncertainties in the state estimation of surrounding obstacles. Simulation and real-world experiments consistently demonstrate a 168% increase in the observation time coverage of typical dynamic humanoid obstacles compared to the method using stationary cameras, which underscores our system's effectiveness in collision risk tracking and enhancing the safety of robotic arms.

Figures (9)

• Figure 1: Overview of our CEASE system.
• Figure 2: Schematic representation delineating the safe region with the inclusion of the two evolutionary approaches, denoted as $\mathbb{T}_{se}$ and $\mathbb{T}_{sc}$.
• Figure 3: Exposition of the COOS Search method: the left figure illustrates the generation of the action set derived from the robotic arm trajectory, considering both known obstacles and safe spaces; the right figure depicts Markov decision process search.
• Figure 4: Count of successful obstacle avoidance experiments for each method across various tests.
• Figure 5: Static obstacle experiments: Figs. \ref{['fig:subfig7']} and \ref{['fig:subfig8']} depict individual methods alongside the original trajectories of the robotic arm in Experiment A and Experiment B. Figs. \ref{['fig:subfig9']} and \ref{['fig:subfig10']} illustrate the schematic representation of the TCP observation method, where the robotic arm collides with a static obstacle.