| Uncontrolled direct fire rockets exhibit high impact point dispersion, even at relatively short range, and as such have been employed as area weapons on the battlefield. In order to reduce the dispersion of a direct fire rocket, feedback control is employed to fire short-duration solid rocket pulses mounted near the nose of the projectile and acting perpendicular to the projectile axis of symmetry.; The feedback law is developed by first determining a piece wise linear model of the projectile swerving motion, subsequently using this linear model to predict the projectile impact point both with and without control, and using the results to command pulses at appropriate times to drive the impact point closer to the specified target.; Candidate optimal control laws are formed using rules based on decision grids, and a global control strategy search algorithm. The global search control law proves to be prohibitively computationally expensive for on-line implementation. The performance of the baseline control law is found to be comparable to the rule based and global search optimal control laws.; The control gains of the baseline control law are optimized in the presence of parametric plant uncertainty using a Monte Carlo simulation. Performance of the system in the presence of parametric plant uncertainty using the optimized gains is deemed comparable to performance of the baseline controller with no plant uncertainty. The level of uncertainty of several plant parameters is varied in order to compare robustness of the controller when optimized with uncertainty viz. without uncertainty. |
