Virtual Sensor for Mobile Sensor Data Cleaning

Missing data is a common data uncertainty occurring in mobile wireless sensor networks. Node mobility prevents existing data cleaning techniques designed for stationary counterpart to properly estimate the missing data. We propose a novel method to clean the missing data in mobile sensor environments. Our method employs computation and memory resources at the base station to establish virtual static sensors collaborating with traditional adaptive filters in prediction model. With certain constraints given below, implementing the proposed algorithm cleans more than 80% of missing data, an increase in performance as compared to existing cleaning methods without any additional hardware implementations.