Automatic Grouping of Redundant Sensors and Actuators Using Functional and Spatial Connections: Application to Muscle Grouping for Musculoskeletal Humanoids
Kento Kawaharazuka, Manabu Nishiura, Yuya Koga, Yusuke Omura, Yasunori Toshimitsu, Yuki Asano, Kei Okada, Koji Kawasaki, Masayuki Inaba
TL;DR
The paper tackles reducing computational complexity in controllers for robots with distributed redundant sensors and actuators by automatically grouping functionally related elements. It introduces a relational graph that fuses functional connections learned via AutoEncoders with spatial proximity constraints, and a randomized grouping algorithm that respects group-size constraints. Applied to musculoskeletal humanoids, the method groups muscles into interpretable regions without a geometric model, achieving high consistency with ground-truth groupings when using a combined functional and spatial approach. This approach enables scalable, interpretable, and potentially online learning-based control across complex, highly redundant robot bodies.
Abstract
For a robot with redundant sensors and actuators distributed throughout its body, it is difficult to construct a controller or a neural network using all of them due to computational cost and complexity. Therefore, it is effective to extract functionally related sensors and actuators, group them, and construct a controller or a network for each of these groups. In this study, the functional and spatial connections among sensors and actuators are embedded into a graph structure and a method for automatic grouping is developed. Taking a musculoskeletal humanoid with a large number of redundant muscles as an example, this method automatically divides all the muscles into regions such as the forearm, upper arm, scapula, neck, etc., which has been done by humans based on a geometric model. The functional relationship among the muscles and the spatial relationship of the neural connections are calculated without a geometric model.