Position validation strategies using partially observable Markov decision processes
October 16, 2011
The collision avoidance system that is currently deployed worldwide relies upon radar beacon surveillance. With its broad deployment over the next decade, aviation surveillance based on Automatic Dependent Surveillance-Broadcast (ADS-B) reports may reduce the need for frequent beacon interrogation over the communication channel, but there is a risk of ADS-B providing erroneous data to the collision avoidance system, resulting in a potential collision. Hence, there is a need to use beacon interrogation to periodically validate ADS-B position reports. Various threshold-based validation strategies based on proximity and closure rate have been suggested to reduce channel congestion while maintaining the reliability of the collision avoidance system. This paper shows how to model the problem of deciding when to validate ADS-B reports as a partially observable Markov decision process, and it explains how to solve for the optimal validation strategy. The effectiveness of this approach is demonstrated in simulation.