Linear Covariance-Based Optimal Sensor Selection for GN&C System Using Second-Order Cone Programming

Abstract

A novel optimal sensor selection approach is developed in this paper. The key innovation of this work is in formulating the stochastic optimal sensor selection problem as a second-order convex program. The approach quickly determines the required optimal sensor specifications that meet mission navigation and trajectory dispersion requirements with the lowest sensor cost. The proposed approach combines linear covariance analysis with convex optimization to describe and solve the optimal sensor selection problem. First, the trajectory dispersion of the closed-loop guidance, navigation, and control (GN&C) system based on sensor specifications is modeled using linear covariance analysis theory. Then, the linear covariance propagation and update equations are used to formulate an optimal sensor selection problem using the Kronecker product. Second-order cone programming with successive approximation techniques are used to solve the established problem. Finally, a simple nonlinear closed-loop GN&C system is investigated, and the capabilities of the proposed approach are demonstrated. The simulations show that the optimal sensor selection problem can be described and solved efficiently using the proposed approach.