Inclusion of the Least Square Method in the MCMC Algorithm for Flight Vehicle Parameter Estimation

The present study investigates the utilization of Bayesian inference in the context of flight vehicle parameter estimation. While conventional methods such as least squares and maximum likelihood provide point estimates, this approach em-ploys Markov Chain Monte Carlo (MCMC) sampling to yield the parameter distribution. Further, it incorporates the least squares method to accelerate MCMC convergence and ensure that the sampling process leads to its expected value. The proposed method is evaluated based on real flight test data and benchmarked against established system identification techniques. While achieving comparable results, it offers the distinct advantage of providing parameter estimates as probability distributions, capturing valuable uncertainty information.