BAYESIAN NONLINEAR FILTER DESIGN FOR SPACECRAFT TERMINAL GUIDANCE,
Abstract
The research applies Bayesian methods to the design of a nonlinear filtering algorithm for digital computers. It is generally recognized that nonlinear filters require important approximations for practical realization, and hence the design details are not independent of the application. A sequential digital filtering technique is developed and applied to the important problem of Mars landing terminal guidance. A detailed digital simulation is conducted to examine the performance of the nonlinear filter of this research. The guidance system is simulated in closed loop fashion and the performance of the system using the nonlinear filter is compared with the performance of the same system using the minimum variance linear filter. The results of the simulation study indicate that the nonlinear filter of this research provides improved performance over the linear filter, particularly in initial filter response. Terminal landing is accomplished accurately and efficiently.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1970
- Accession Number
- AD0708124
Entities
People
- Frederick A. Evans
Organizations
- University of California, Los Angeles