Optimal Sensor Position: Exploring the Interface Between the User and Sensor in Activity Recognition System
Chengshuo Xia · 2021
Human activity recognition systems combined with machine learning normally serve users based on the fixed sensor position. Uniform sensor position normally cannot satisfy the user’s demand according to different conditions. In this paper, we recognized the sensor position as an interface between the user and sensor system. We designed the optimization scheme to generate the best sensor position for activity recognition system. The user can indicate his/her preferred or disliked position and sensor numbers and the proposed optimization evaluates which position or positions combination can generate best accuracy under user’s preference. With the experiment, the proposed scheme can be employed to discover the optimal position to help the HAR system in a simple and customized way.