Proprioceptive State

Proprioceptive state estimation focuses on determining a robot's position and orientation using only internal sensors, like IMUs and joint encoders, a crucial capability for robust locomotion and manipulation in environments lacking external sensing. Current research emphasizes improving accuracy and reducing drift using advanced filtering techniques, such as invariant Kalman filters, often combined with neural networks to enhance robustness and handle complex dynamics. These advancements are significantly impacting robotics, enabling more reliable control of legged robots, soft manipulators, and even humanoid robots in challenging terrains and during complex tasks.