Using Neural Networks for the Recognition of Specific Motion Symptoms of the Parkinson’s Disease

It is proposed a wearable sensing system based on Inertial Measurement Unit which uses Artificial Neural Networks (ANNs) for the detection of specific motion disorders typical of the Parkinson’s disease (PD). The system is made of a single inertial sensor positioned on the head by the ear. It recognizes noticeable gait disorders potentially dangerous for PD patients and can give an audio feedback. The algorithm of recognition based on ANNs is extremely versatile and correctly operating for any individual gait feature. It provides robust and reliable detection of the targeted kinetic features and requires fast and light calculations. The final headset system will be extremely energy efficient thanks to its compactness, to the fact that the ANN avoids computational energy wasting and to the fact that the audio feedback to the patient does not require any wired/wireless connection. This improves the system performance in terms of battery life and monitoring time.