Rehabilitation is a central part of healthcare that promotes recovery and improves the quality of life of patients. It is widely recognized that rehabilitation must be highly personalized and tailored to individual needs. A comprehensive assessment of the patient’s condition is essential to identify functional impairments, set functional goals, and develop individualized treatment plans. Regular reassessments allow for monitoring progress and adjusting interventions to optimize rehabilitation outcomes. An important part of the assessment is the evaluation of motor functions, such as walking, sitting, and grasping. Traditionally, this assessment has been done through observations by healthcare professionals, which is a subjective approach that often leads to variability in results and limitations in precision and reliability. There is a great need to develop a new method for assessing motor functions. We propose the use of biomedical radar for motor function assessment in the rehabilitation field. Biomedical radar works by emitting electromagnetic waves towards a person and analyzing the reflected signals. Due to its non-contact nature, biomedical radar enables non-invasive and continuous health monitoring. Over the past decade, biomedical radar has been investigated for a variety of applications, such as sleep monitoring, respiratory and heart rate monitoring, fall detection, and behavioral and activity recognition. In this project, we will use biomedical radar to assess motor functions in Parkinson's disease, with a focus on finger tapping. The finger tapping test is used to evaluate bradykinesia in a Parkinson's patient and is one of the tests included in the standardized clinical assessment tool MDS-UPDRS (Movement Disorder Society-Unified Parkinson's Disease Rating Scale). Within this scale, three aspects of bradykinesia are assessed with finger tapping: frequency, amplitude, and rhythm, resulting in a total score ranging from 0 (normal performance) to 4 (severe). The goal of this project is to develop a machine-learning algorithm to classify the various categories of bradykinesia level using biomedical radar measurement data.