A smart pen prototype with adaptive algorithms for stabilizing handwriting tremor signals in Parkinson's disease

This study presents a smart pen prototype designed to dynamically mitigate hand tremors, thereby enhancing writing quality and user comfort for individuals with conditions such as Parkinson's disease. The device employs an accelerometer for real-time tremor detection, a microcontroller for rapid data processing, and a vibration motor to counteract tremor effects. Adaptive algorithms-including Fx-LMS, Fx-NLMS, a combined Fx-LMS/NLMS approach, RLS, and the Kalman Filter-were evaluated using signals from the NewHandPD dataset. Simulation results revealed that although the RLS algorithm achieved the lowest mean square error, the Kalman Filter converged approximately eight times faster, a finding that was confirmed through microcontroller tests and further validated on an orbital shaking table under constant and variable tremor conditions. These outcomes underscore the potential of the Kalman Filter as a non-invasive, adaptive solution for real-time tremor mitigation in assistive writing devices. Future improvements may include integrating additional sensors and further optimizing microcontroller performance to enhance overall adaptability and accuracy.