An easily attachable measurement system of joystick angle in a power wheelchair using IMUs for maneuvering logger.

Monitoring joystick operations in power wheelchairs (PWCs) is promising for investigating user-wheelchair interaction and providing quantitative measures to assess the user's driving performance. In this paper, an add-on measurement system, Power Wheelchair Maneuvering Logger (PWhML), is developed to provide an easy-to-implement and cost-effective solution for monitoring the user's joystick operations in PWCs. The proposed system uses two compact inertial measurement units (IMUs), which are respectively attached to the joystick and wheelchair armrest for movement sensing. A coordinate transformation-based method is proposed to estimate the joystick operating angles using the acceleration data measured by the attached IMUs. The accuracy of the proposed method was thoroughly evaluated under different conditions. The evaluation trials in a stationary PWC reported a mean absolute error (MAE) of 0.59° in the forward/backward direction and 0.64° in the leftward/rightward direction, validating the established geometry model for coordinate transformation. The subsequent driving experiments on outdoor test courses demonstrated the effectiveness and robustness of the proposed method in various terrain conditions (MAE of less than 3°). A clustering analysis based on the t-distributed stochastic neighborhood embedding method correctly categorized different driving activities using the joystick operating angles measured by PWhML. These results indicate that integrating the developed PWhML into PWCs can facilitate a quantitative measurement of the user's driving behavior, providing valuable insights to identify careless operation patterns and help PWC users to improve driving performance.