Data-driven testing methodology for RFID systems

A radio-frequency identification (RFID) system including hardware and software may be updated from time to time after first time deployment. To ensure the reliability of the system, extensive tests are required. However, enumerating all test cases is infeasible, especially when the tests involve time-consuming hardware operations. To solve this problem, we propose a testing methodology for RFID systems which does not enumerate all test cases but rather those which are representative. A clustering method is adopted in selecting representative test cases. Although a small number of selected test cases are run, we can still obtain a relatively high bug detection rate compared with running the enumerated test cases. Our extensive experiments show the efficiency and effectiveness of our testing methodology.