The development, selection, and adaptation of assistive technologies such as exoskeletons to assist people with disabilities is associated with a complex decision-making process due to the uncertainty of evaluation criteria. Traditional decision-making methods in this area often fail to address these complex challenges, leading to inefficiencies in the preparation and implementation of the exoskeleton production process and, consequently, reduced product quality. To overcome these challenges, this article proposes an artificial intelligence (AI)-based decision support approach for the development of upper limb exoskeletons. This approach reduces costs, improves production quality, and accelerates exoskeleton design and production, with accuracy reaching 100 per cent using a multilayer perceptron, enabling more accurate and realistic results. The article presents new models supporting the classification of hand dysfunctions, exoskeleton design (indicating the position of exoskeleton actuators, the number of actuators required, and the maximum grip force of the exoskeleton), and estimating manufacturing costs. This shows how to optimize AI-assisted technologies in the form of exoskeletons in support systems for people with disabilities.
Izabela Rojek (Thu,) studied this question.