Robotic Tremor Simulation for Experimental Validation of Inertial Sensors in Neurodegenerative Disease Monitoring

Neurodegenerative diseases such as Parkinson’s require precise and repeatable diagnostic tools to enable early and objective symptom assessment. Inertial measurement units (IMUs) have shown strong potential for capturing motor impairments, but clinical validation is often limited by variability in patient behavior and uncontrolled environments. This paper presents a novel method using a programmable robotic arm (KUKA KR 4 R600) to generate synthetic tremor profiles that replicate key features of Parkinsonian motion. The motion profiles were generated via RoboDK and recorded with industrial-grade STWIN SensorTile IMUs placed in different locations. Signals were processed through gravity compensation, projection onto principal directions, and alignment with the robot’s kinematic frame. The spectral analysis revealed a dominant frequency at 6.7 Hz and an amplitude of 0.012 g, in line with physiological tremor ranges. A consistent secondary component at 2.5 Hz was also identified, attributed to repeatable system dynamics. Leakage between joint movements was minimal, with cross-talk coefficients below 4%. The platform allows reproducible and decoupled evaluation of IMUs under controlled conditions, supporting sensor development for early detection of neurodegenerative disorders.