Modified arithmetic optimization algorithm for drones measurements and tracks assignment problem
Neural Computing and Applications
Seghrouchni, Amal El Fallah
This paper presents efforts to solve the multi-track measurement assignment problem in drone detection and tracking. In many cases, several radars are collectively used to track drones efficiently, generating measurements and several tracks under different circumstances. In this work, several measurements are simulated during a time frame accompanied by the generation of several tracks using the Linear Kalman Filter. The focus is on finding an optimum measurements/track assignment for the simulated measurements and track values. The measurements and track generation are simulated using Stone Soup software. On the other hand, the optimization of the problem is implemented using several evolutionary-based metaheuristic algorithms. This optimization problem is known to be computationally explosive, especially if long time frames are considered. In particular, a new modified method based on the Arithmetic Optimization Algorithm is proposed. The optimization is applied to a formulated cost function that considers uncertainty, false alarms, and existing clutters. Simulations and comparisons show the ability of those evolutionary-based algorithms to solve this kind of problem efficiently. The proposed method obtained promising results compared to other comparative methods used to solve this drone’s measurements and track assignment problem.