Contact Force

Contact force research centers on accurately measuring, predicting, and controlling forces during physical interactions between robots and their environments. Current efforts focus on developing robust methods for estimating contact forces using diverse sensor modalities (vision, tactile, audio), often employing machine learning techniques like neural networks (including transformers and recurrent networks) and optimization algorithms to process sensor data and model complex interactions. These advancements are crucial for improving robot safety, dexterity, and performance in tasks ranging from delicate manipulation to human-robot collaboration and legged locomotion in unstructured terrains.