Towards Neural-controlled Artificial Legs using High-Performance Embedded Computers

Abstract:

The objective of this study is to develop a high-performance and robust neural-machine interface (NMI) for artificial legs, which can accurately and reliably identify user intent in real-time. Seamlessly integrating human physical neuromuscular control system with cyber systems is challenging: (1) accurate neural decoding using the non-stationary lower limb neuromuscular signals during dynamic movement is difficult; (2) high-robustness must be ensured to deal with environmental disturbance and unexpected sensor failures; and (3) adequate computation power and effective system integration of an embedded system for real-time processing system is required to stream and store multiple sensor data, classify user intent, and process sensor monitoring algorithms at the same time. To design such cyber-physical system (CPS), we have developed novel techniques to accurately and reliably decipher non-stationary neuromuscular control signals and process such signals using a new high-performance embedded system. The designed NMI software consists of a neuromuscular-mechanical fusion algorithm for user intent recognition and a trust management module to ensure the NMI's robustness and trustworthiness. The algorithms have been specifically tailored to high-performance embedded systems (e.g. GPU and FPGA) for real-time implementation. In order to evaluate the system performance, the NMI prototype has been implemented online to control a virtual-reality cyber system. The NMI can accurately identify user intent in real time, potentially useful for neural control of artificial legs. Our breakthrough in design of CPS for lower limb prosthesis control has great potential to significantly improve the quality of lower limb amputees' life. Moreover, the novel concepts and engineering frameworks proposed in this study can benefit the CPS research community to apply high-performance and fast growing computing system in biomedical engineering and health science.