A nonlinear predictive control scheme, which does not require an exact solution of optimal control problem is proposed. The proper asumptions are formulated and the stability of the closed loop system is proved. These results are used for analysis of proposed algorithm robustness. Robust stability of the alghorithm is proved in the case of suboptimal solutions of the optimal control problem in the presence of bounded disturbances. Proposed schme is applied to the pendulum-cart nonlinear control system with disturbances.
A version of Pontryagin's Maximum Principle for optimal control problems with final state and control time constraints is presented. Sufficient conditions for normality of Lagrange multipliers are given. Considerations are illustrated with examples.
Paper describes state estimator for linear systems with quantized measurement. When quantization interval tends to zero estimator became Kalman-Bucy filter. If measurement noise intensity is low relative to quantization interval, then estimation error could be several times smaller than estimation error of Kalman-Bucy filter. Considerations are illustrated with examples.
