Intelligent Control Technologies for Smart Prosthetic and Orthotic Devices: A Systematic Review

Smart prosthetic and orthotic systems are evolving from isolated assistive devices into closed-loop human–machine systems that integrate biological sensing, intelligent intent decoding, adaptive actuation, and sensory feedback. This systematic review brings together evidence from 70 sources on recent advances in intelligent prosthetic and orthotic systems. The reviewed approaches include the following: surface and high-density electromyography, peripheral neural interfaces, multimodal sensing, artificial intelligence, adaptive and reinforcement-learning-based control, sensory restoration, digital twins, cybersecurity, and clinical translation. The review was conducted and reported in accordance with the PRISMA 2020 guidelines. Records were retrieved from Scopus, PubMed/MEDLINE, and Web of Science and screened using predefined eligibility and scope criteria. Of the 312 records identified, 48 duplicates were removed, leaving 264 records for screening. Following the screening process, 115 full-text articles were assessed for eligibility, and 70 sources were ultimately included in the qualitative synthesis. Collectively, the findings indicate that, the evidence points to a gradual shift from conventional, fixed myoelectric control toward more adaptive and multimodal control architectures. Approaches such as targeted muscle reinnervation, regenerative peripheral nerve interfaces, implanted electrodes, and bidirectional neural interfaces may improve signal specificity and expand the number of controllable degrees of freedom. However, high decoding accuracy achieved under controlled laboratory conditions does not necessarily translate into meaningful clinical benefits. Important factors, including signal drift, the need for repeated recalibration, response latency, long-term stability, usability, cost, and performance during real-world activities, remain inconsistently evaluated across studies. Based on these findings, the review proposes an end-to-end framework that links biological and environmental sensing with signal conditioning, user-intent interpretation, adaptive actuation, sensory feedback, and progressing from laboratory research and prototype development through validation and clinical testing to regulatory approval, clinical implementation, and ultimately real-world use. The review also the review identifies several key priorities for future research, including standardized benchmarking protocols, longitudinal multicenter validation, interpretable and uncertainty-aware artificial intelligence, durable neural and sensing interfaces, cybersecurity, and patient-centered outcome measures. Therefore, addressing these areas will be essential for moving intelligent prosthetic and orthotic technologies from promising experimental systems toward reliable, clinically deployable solutions.