Markov Decision Processes to Model Decision-making in Harsh Electromagnetic Environments

Electromagnetic interference (EMI) poses a significant threat to the safe and reliable operation of electronic devices. Implementing data triplication and EMI-detection techniques are effective approaches to improve electromagnetic resilience. Although detector warnings can prompt the system into a safe state, they may also decrease system availability and increase data uncertainty. Nonetheless, prioritising EM resilience is essential for organisations that rely heavily on electronic devices. This paper presents a Markov decision process-based approach to model decision-making in a reverberant electromagnetic environment. To determine the optimal action in each state, the model calculates the state transition probabilities using the warnings generated by the EMI detector. The effectiveness of the decision model is assessed by testing it against various frequencies of EMI and bit rates. The findings suggest a significant decrease of 24.8 0% in the symbol error rate, compared to the baseline scenario. These results demonstrate the value of implementing a decision model in order to reduce errors and improve performance.