Efficiency-Aware Two-Stage Hybrid Processing Method for Tactile Data Outliers

Tactile communication has garnered widespread attention due to its advantage of providing users with immersive experience. Processing tactile data outliers, which are caused by factors such as operator hand tremors, is a necessary step for improving the quality of tactile communication. However, the efficiency of the existing data-based outlier processing methods needs to be improved, as extensive computations consume time. Therefore, this paper proposes a two-stage hybrid method for processing tactile data outliers. Specifically, for the first stage, we first propose to rapidly detect abnormal subsequences by utilizing the correlation among data, and then filter the abnormal points in the abnormal subsequences based on the local outlier factor algorithm. For the second stage, we construct linear regression models to predict and correct these outliers. To verify the effectiveness of the proposed method, we implement a prototype of the tactile communication system. The experimental results show that the detection accuracy is more than 97.0%, and the coefficient of determination grows by 1.94- fold.