We use ideas from game theory to define a new notion for an optimal threshold for the number of erroneous responses that occur during the rapid-bit exchange over noisy channels in a distance-bounding protocol. The optimal threshold will ensure that even if an intruder attempts to cause incorrect authentication, the expected loss the verifier suffers will still be lower than when the intruder does not attack. Any rational intruder, who always tries to maximise the verifier’s loss, will not therefore have any incentive to attack the protocol. We then demonstrate how statistical analysis and binary search can be used to locate such a unique and optimal threshold accurately.
Long H. Nguyen