Designing nonlinear control system considering pre- and post-dynamics of feedback loop

In practical applications, dynamic physical systems under control can usually be accompanied by different types of associate pre- and post-dynamics, e.g. regarding actuation and measurement units. However, the designed control algorithms for these systems in both the industrial sector and research practices are typically focused on the main plant, neglecting the impacts of these dynamic-laden associate modules, which may have been added provisionally during the control procedure and whose dynamics can affect the closed-loop system behavior by complicating the system nonlinearities while challenging its asymptotic stability. Owing to this crucial role in the system overall performance, the controller design-stage incorporation of associate dynamics has been put into spotlight by this paper. It presents a trajectory tracking control design procedure, integrating sensor and actuator dynamics under measurement uncertainty and input hard constraints. To this end, two different approaches, namely with and without considering sensor dynamics in the control algorithm design, were explored. The Gradient Descent (GD)-based optimal two-mode discrete-time sliding mode tracking control with the searching algorithm devised for reference trajectory redesign was proposed. Then, the results were investigated in terms of tracking error, control energy, and robustness against parametric uncertainty. The sensor combined-based approach accomplishes redesigned trajectory tracking with less error and lower control energy consumption. Despite being more agile, the measurement dynamic-neglected approach posts poor tracking performance, as its position error and control energy curves enfold those of the sensor combined-based approach. The sensor combined-based approach also proves more robust against parametric uncertainty, as it successfully performs under some circumstances that direct the sensor dynamic-neglected approach toward instability. Additionally to demonstrate the performance of the proposed controller, two different control strategies – with and without devising the optimization-based searching algorithm – were compared. Adding the searching algorithm deeply decreases the control system dependence on the pre-tuning of implementation parameters. It also further reduces the tracking error and control energy compared to the second strategy. Finally, in addition to the effectiveness of the controller, results obtained from the comparative simulations conducted between the aforementioned approaches and strategies, show the impact of control design-stage incorporation of sensor dynamics in terms of tracking error, robust stability, and energy efficiency while considering input hard constraints.