New control strategies for neuroprosthetic systems

Area of Science:

• Biomedical Engineering
• Neuroscience
• Rehabilitation Technology

Background:

• Artificial muscle excitation techniques offer significant potential for restoring limb movement in individuals with paralysis.
• Neuroprostheses are crucial for stimulating muscles and controlling artificial movements, necessitating sophisticated control strategies.

Purpose of the Study:

• To explore various control methods for neuroprosthetic limb movement, including feedforward, feedback, and adaptive control.
• To evaluate the efficacy of different control strategies in achieving desired movement trajectories and compensating for performance degradation.

Main Methods:

• Development of an artificial motor program for upper extremity neuroprosthetic control.
• Implementation of adaptive feedforward control (cycle-to-cycle controller) for performance compensation.
• Utilizing a neural network controller to customize stimulation parameters and manage muscle fatigue.

Main Results:

• Adaptive feedforward control effectively compensated for performance decreases seen in open-loop systems.
• Neural network controllers demonstrated the ability to personalize stimulation for desired trajectories and maintain performance despite muscle fatigue.
• Lower extremity control achieved optimal results by focusing on perceived gait objectives rather than precise joint trajectories.

Conclusions:

• Practical functional neuromuscular stimulation (FNS) control systems should incorporate features observed in neurophysiological systems.
• Adaptive and neural network control strategies show promise for enhancing the functionality and robustness of neuroprosthetics.
• Tailoring control to individual biomechanics and fatigue is essential for effective limb movement restoration.