Reactive Self-Collision Avoidance for Dual-Arm Robots Using a Temporal Feature Modeling and Fusion Network

Dual-arm robots (e.g., humanoid robots) possess substantial potential for executing collaborative tasks in universal scenarios. However, the workspaces of the individual arms frequently overlap, rendering self-collision avoidance crucial for maintaining safe robotic operations. Existing motion planning-based methods demonstrate inadequate real-time efficacy, and learning-based distance proxy methods are subject to elevated false positive rates. To address these challenges, we present a novel minimum distance prediction neural network for reactive collision avoidance of dual-arm robots, which considers the continuous motion of the robotic arms and the interrelation of joint configurations. The temporal joint configurations are encoded and divided into historical and current features. A state-space model is utilized to capture the temporal dependency of historical features. A self-attention mechanism is employed to model the hidden relationships among current features. The integration of historical and current features via a cross-attention mechanism followed by a gated fusion module allows for precise prediction of the minimum distance between the dual arms. Simulations and real-world experiments, including human-robot interaction and autonomous grasping tasks, were conducted using two redundant robotic arms. The proposed method achieves an average error of 1.804 cm in minimum distance prediction. In dual-arm autonomous grasping experiments, an average error of 1.471 cm is attained. Our approach has improved accuracy by 38.08% over the state-of-the-art methods. No collisions occurred throughout all real-world experiments. This method holds promise for extensive applications of dual-arm robots. The code is accessible at https://github.com/XuejinLuo/SelfCollision Note to Practitioners - Dual-arm robots have shown great potentials in collaborative tasks. The collision between the two redundant robotic arms has to be addressed during the motion. A straightforward solution is predicting the minimum distance between the two arms in real time and incorporating the distance information into the framework of optimal control for self-collision avoidance. Existing state-of-the-art methods have accuracy issues in the distance prediction. We propose a novel neural network for real-time minimum distance prediction for reactive collision avoidance of dual-arm robots. we introduce a temporal model and self attention mechanisms into the network, which is supposed that the temporal correlations between the joint configurations of adjacent frames can be effectively modeled and fused, leading to more accurate and smoother prediction of the minimum distance. Our proposed network outperforms the state-of-the-art methods, achieving higher accuracy in minimum distance prediction with a reduced false positive rate. The proposed network is further incorporated into a quadratic optimization framework for realizing the self-collision avoidance during task execution. Comprehensive real-world experiments have demonstrated the effectiveness of our method.