ROMJIST Volume 28, No. 4, 2025, pp. 377-384, DOI: 10.59277/ROMJIST.2025.4.06
Sorin LAZARESCU, Florin ISTUDOR, Octavian IONESCU, David DRAGOMIR, Marian ION, Ladislau SZEKELY, Szilard OLASZ, Istvan TUSEZ, Gabriela IONESCU, Stefanos P. ZAOUTSOS, Catalin NICULAE, Bogdan SINATOMA Personalized Movement Algorithms for Neural Forearm Prostheses Using Convolutional Neural Networks
ABSTRACT: The article presents personalized movement algorithms for neural forearm prostheses equipped with AI module. The implementation of personalized motion algorithms is done by using a glove with finger flexion sensors mounted on the patient’s healthy hand, a neural interface with plug electrodes implanted in the motor fascicles of the median and ulnar nerves from the patient’s amputation stump and an AI module based on a convolutional neural network (CNN). For each prosthesis movement selected by the patient, the glove will detect the movement parameters of the fingers from the healthy hand, the neural interface will detect the motor neural signals from the median and ulnar nerves, and the CNN will identify the specific pattern of motor neural signals. The AI module will thus learn to recognize the patient’s commands for the prosthesis and will command its natural movements.KEYWORDS: Convolutional neural networks; implantable neural interface; movement algorithms; neural prosthesesRead full text (pdf)