Assistive technology for neuro-rehabilitation applications using machine learning techniques

Machine learning (ML) has become a part of every human’s day-to-day life. Assistive technology (AT) is a device with programming software, which helps or assists people with disabilities to do work by themselves by giving commands through speech, eye/hand movement, or medical signals. The main focus of the area of AT is developing devices for disabled people who are struggling with communication, cognition, and using physical strength, etc. In this chapter, Assistive Technologies for stroke patients are discussed. Especially accessing the computer by voice and hand gestures. The term “neurological disorders” refers to disorders of the brain and nerves in the body, as well as the spinal cord. There are a variety of symptoms that can result from abnormalities of the brain, spinal cord, or other nerves. A person with this condition may experience loss of memory, sensation, muscle weakness, seizures, pain, paralysis, and altered mental state. The process of assisting and educating/training people having a neurological disorder like Parkinson's disease with proper guidance from doctors and experienced physical therapists is called neuro-rehabilitation. ML helps the devices to learn, understand, and classify the data or command by using training datasets either online or offline without manual interaction. The performance of the ATs depends on how fast the device learns the command; this mainly depends on the performance of learning algorithms. Accuracy and training dataset plays a crucial role in ATs. Different ML techniques, such as support vector machine, naïve Bayes, k-nearest neighbor, discrete tree, Random Forest, and so on, are used for training, learning, and classifying the audio, image, and medical signals. Depending on the algorithm performance the ATs performance was calculated.