Assessing Benefits of Coordination on Safety in Automated Highway Systems

In this report, we present a methodology for assessing the benefits of different vehicle coordination strategies on the safety of a platoon during emergency braking. One can say that a coordinated braking strategy B is more beneficial than a strategy A, if strategy B leads to a larger reduction in the probability of a collision, the expected number of collisions, and the expected relative velocity at impact as compared to strategy A. We consider an emergency braking scenario, in which the lead vehicle brakes at its maximum capability and the following vehicles brake while obeying a vehicle following control law. The sequence of maximum deceleration of vehicles in the platoon is assumed to be a sequence of independent and identically distributed random variables; this distribution is assumed to be discrete and known. Due to coordination, however, the “effective” deceleration of a following vehicle may not necessarily be its maximum value. The problem of assessing the benefits of coordination can be formulated as three subproblems: the first subproblem deals with determining the probability distribution of the “effective” deceleration of following vehicles during emergency braking. It is intuitive that the smaller the variance of this distribution, the greater the safety benefits are. The second subproblem deals with determining the probability of an intervehicular collision, the expected number of collisions and the expected relative velocity at impact as a function of the difference in the braking capabilities of successive vehicles in a platoon. The probability of an intervehicular collision and the expected number of primary collisions are computed via a Markov chain. Here, the asymptotic behavior (as the size of the string increases indefinitely) of the probability of an intervehicular collision and the expected number of primary collisions is computed. The third subproblem deals with conducting Monte Carlo simulation to