Stochastic gain tuning method applied to unmanned space vehicle flight control design

This paper studies the feasibility of applying the stochastic gain tuning method to the flight control system design of space vehicles. Stochastic gain tuning is a form of parameter optimization by which the probability of the flight control system's total mission achievement is maximized. This probability is estimated by applying the Monte Carlo method to the results of a large number of simulated flights. The flight simulation model contains various types of uncertain parameter, the stochastic properties of which are defined a priori. The flight control system requirements are defined based on the flight simulation results, and an optimization algorithm called the mean tracking technique is used to tune the feedback/feedforward gains of the flight control laws, which maximizes the probability of mission achievement. Although stochastic gain tuning requires large computational resources, the recent advent of low-cost, high-performance computers means that it has become feasible and practicable if efficient computational algorithms are employed. This paper demonstrates its feasibility by applying it to the design of the flight control system of a re-entry space vehicle, a low-speed sub-scaled model of which was flight tested in 1996.