Wireless accelerometer sensor data filtering using Recursive Least Squares adaptive filter

One of the applications of wireless accelerometer sensors is to monitor the human body motion and orientation. Such applications range from medical observations, diagnostics and fall detection to game console and entertainment. In typical wireless accelerometer sensors the received data is affected by the nonlinear environmental noises leading to a decreased resolution and the accuracy of received data. In this paper, a Recursive Least Squares (RLS) adaptive filter is used to filter the three-axial wireless accelerometer sensor received data. We used the Shimmer 9DOF wireless sensor placed on a robotic arm which is created to perform a precise 3D movement and is controlled by a PC. In the proposed RLS filter the input data is defined by the Shimmer accelerometer data and the desired data is defined by the robotic arm orientation. The obtained results from running the RSL filter on the wireless accelerometer sensor data shows a reasonable accuracy compared to the unfiltered data.