RFID data interpolation algorithm based on dynamic probabilistic path-event model

Missing reads occur frequently during RFID (radio frequency identification) data collection, which will reduce the accuracy of query results in RFID applications. To solve this problem, the existing algorithms mainly take primitive RFID readings as granularity and adopt window smooth strategy based on tag historical readings, which may interpolate data that the query doesn't care about and incur inaccuracy when multiple logic areas are involved. In this paper, data are transformed from data level to logic area level as the interpolation granularity. Then three data interpolating algorithms based on the probabilistic path-event model are proposed, where the incoming events are judged and interpolated by mining the sequence correlation of known area events. Furthermore, the factor of time is considered, and thus probabilistic path-event model is developed. Abundant experiments prove the proposed algorithms have different performance advantages in different conditions and are predominant over the existing strategy in redundancy and accuracy.