Using Swarm Intelligence and Bayesian Inference for Aircraft Interrogation

Abstract

Sensor management deals with the efficient resource allocation to meet mission objectives of the application, air traffic control. A schedule for the sensors is constructed, which simultaneously meets the measurement accuracy and update rate, while minimizing the transmissions from the sensors. Bayesian inference is used to determine management requirements for individual aircraft. Particle swarm optimization, a technique modeled after swarming insects to solve multi-objective optimization problems efficiently, is used to form, test, and search potential sensor schedules for a final schedule solution that maximizes the fitness function. The fitness function is composed of weighted performance estimates enabling quantitative comparisons among schedules. A Bayesian network optimizes the values of the weights as well as the performance requirements to get the best overall performance possible from the entire sensor network system.

Density estimation using a new dimension adaptive particle swarm optimization algorithm

“…The binaryheadered PSO (Veeramachaneni et al 2007) has been applied to the air traffic scheduling problem (Kamath et al 2008), and is adapted in this paper for density estimation. The other two algorithms based on PSO, the exhaustive PSO and the discrete-headered PSO, are newly proposed in this paper to solve the unknown dimension problems.…”