An arm musculoskeletal control scheme incorporating cerebellar and emotional learning models

The cerebrum and cerebellum play a crucial role in motion control and are crucial to perform a variety of fast, precise movements for humans and animals. Emotions are generated in the cerebral cortex, and activate the amygdala, which promotes the storage of information in various regions of the cerebrum. In this paper, cerebellar learning model, emotional learning model, and spinal cord calculation module are incorporated to complete the control of an arm musculoskeletal system, and the redundancy problem of the musculoskeletal system control is solved through the optimized calculation in the spinal cord module. The arm musculoskeletal system can thus complete the end trajectory execution task successfully. It is shown that compared with the cerebellar motion control scheme, the proposed scheme has the advantages of fast learning convergence, simplified synaptic adaptation of cerebellum and strong anti-disturbance ability. It is also verified that the proposed control scheme exhibits good robustness to random noise. The proposed arm musculoskeletal control scheme operates effectively and provides a theoretical reference for the application of biomimetic musculoskeletal system.