Sensor technology with gait as a diagnostic tool for assessment of Parkinson's disease: a survey

Parkinson's Disease (PD) is the most precarious chronic disorder of the human brain affecting millions of people globally. In today's world of technical inventions, the diagnosis of PD at its initial stage is a serious issue. The high dependency on the patient's medical history and clinical rating scales have certain limitations as they are very tiring and include recall bias. Therefore, more stable and objective measurements are needed to execute automated and effective detection of PD. The biometric gait has been developed as a reliable diagnostic tool for such a purpose because of its unobtrusiveness. In past years, the applications of sensor devices have been utilized the most to evaluate PD via gait. So, the purpose of this article is to analyze the past and current research toward sensor-based (SB) diagnosis of PD motor symptoms. In this article, we provide a brief description of PD and the related concepts also defining its impact on human gait. This article comprehensively surveys the SB technology and the role of different sensors in PD gait recognition. This study investigates the machine learning paradigms used in PD analysis and their performance evaluation. Several SB PD gait datasets are surveyed and explored considering literature from the last ten years. Also, we exhaustively provide a discussion section in this article to give a clear picture of the results concluded from the analysis of prior sections. At last, this article examines some gaps in the existing studies that need to be addressed and also suggests some measures to tackle such issues using advanced paradigms.