Subjective Logic-Based Identification of Markov Chains and Its Application to CAV's Safety

A reliable estimation of the communication channel which connects automated vehicles is an important step towards the safety of connected and automated vehicles. The communication channel is usually modeled as Markov chain with slowly time-varying transition rates and is identified through statistics on the observed state transitions. However, the classical identification approach lacks a measure on how reliable the identification results, and thus the channel, are. In this work, we propose an identification method, based on the subjective logic theory, which features such a reliability measure in terms of statistical uncertainty. We demonstrate through simulations that the proposed method is capable of quickly responding to parameter changes. Furthermore, it is shown that the transition rates of the Markov chain are tracked with high accuracy. Finally, we validate our results by a real-world experiment.