A Black Box Measurement Method for Reliability of Wireless Communication

Abstract

In the future, connected vehicles and automated factories will use wireless communication. To implement these applications, wireless networks must be reliable. Because the cost of communication failures in such applications is high, reliability should not only be estimated in advance, but also measured directly in practice. To date, no method exists that is able to determine, if a wireless network is 99.999 % reliable. To address this need, we propose a black box test that uses only the success/fail status of transmissions to determine the reliability of wireless communication. The test is based on two assumptions: (1) reliability is time-invariant and (2) transmissions are statistically independent. Because it is crucial for any measurement method that its assumptions hold, we test both assumptions as part of the method. The proposed method is especially suited when access to lower layer information is limited to the information returned by off-the-shelf hardware. The method measures reliability of wireless networks without support and knowledge from the network operator and administrator. In the future, methods such as the proposed one are needed to ensure reliable operation of wireless networks in critical scenarios.