Information gain based sensor search scheduling for low-earth orbit constellation estimation

Abstract

This paper addresses the problem of sensor search scheduling in the complicated space environment faced by the low-earth orbit constellation. Several search scheduling methods based on the commonly used information gain are compared via simulations first. Then a novel search scheduling method in the scenarios of uncertainty observation is proposed based on the global Shannon information gain and beta density based uncertainty model. Simulation results indicate that the beta density model serves a good option for solving the problem of target acquisition in the complicated space environments.

Best Basis Selection Method Using Learning Weights for Face Recognition

“…Non-Negative Least Squares classification (NNLS) and Linear Regression Classification (LRC) were used for this classification test [ 52 – 54 ]. In addition, Information Gain (IG) and Sequential Forward Selection (SFS) were also used for this best principal component selection test [ 55 , 56 ]. These algorithms were suitable for our experiment since the performances of these algorithms were recognized by many researchers.…”