The Moving Mass Actuated Projectile Attitude Control and Estimation Using a Predictive Controller and Estimator Based on the Particle Filter

In this paper, two particle filters are proposed to estimate the state and control the attitude of a projectile with a moving mass actuator. Since the dynamic equations of a flying object with a moving mass actuator are nonlinear, a nonlinear estimator and controller are used. The first filter takes the angle measurement and mass displacement along with noise as observations and uses them to estimate the angle, the rate of angle, the mass displacement, and the velocity of the mass. In this research, the nonlinear predictive control problem becomes a dynamic optimization problem and then, at each time step, the best control signal is calculated online using the second particle filter in the finite horizon and applied to the system. The cost function used for the first filter is the error of the values measured and calculated by each particle. The cost function in the predictive control problem for each particle is the error of the angle tracking and the control effort. The simulation results for the performed scenarios show that the algorithm proposed for solving the nonlinear problem of controlling the attitude of a moving mass actuated projectile has a good performance in the presence of Gaussian and non-Gaussian noises. Also, due to the random nature of the problem, statistical analysis of the results is performed using Monte Carlo simulation.