constrained model predictive control

In this article, using the constrained predictive controller, the tracking of a moving target by a group of flying robots while controlling the group arrangement is presented. The maintenance and control of the group arrangement is based on position control with the assumption of decentralized architecture and leader-follower communication structure between the agents. Target tracking is done by the group leader and the followers are responsible for tracking the group leader. The leader of the group constantly sends the current position and the estimate of his position in the next steps to the followers to use it to calculate the optimal current position and the next moments. In the following, rewriting the constraints of the problem, such as ensuring that the distance is maintained and that the followers do not collide with the group leader, and the limitations of inputs and their changes on all members of the group, as well as providing an innovative geometric method for crossing obstacles and systematically applying all of them in the control process. has been Simulations are presented for a group of five flying robots that seek to intercept a moving ground target while maintaining the formation of the group. The results show that the controller has a good resistance to the disturbances in the group's movement path and the group interception is carried out in a favorable way. Predicting the movement path of the moving target and using it in the design of the predictive controller also improves and smooths the movement path of the group.

Designing the suitable autopilot for the quadrotors is very important in how the flight vehicle moves and follows the specified reference path. One of the suitable controllers for autopilot design is the predictive controller, which is the most well-known Generalized Predictive Controller method. In the autopilot design, constraints on inputs as well as communication delays must be taken into account, and if these two issues are not addressed in the controller design, the autopilot will not function properly and may even lead to instability. In this paper, a generalized predictive controller with consideration for delayed input data for a quadrotor autopilot is presented. Also, in order to determine the predictive control parameters, the Metaheuristic method, particle swarm optimization, has been used and these parameters have been optimally adjusted. In adjusting the controller parameters, the objective function is used based on the performance indicators such as settling time, peak time, overshoot, and steady-state error. By adjusting the weights of this function, the controller performance indicators can be determined. Also, the simulation results show that the controller performance based on the defined objective function is much improved compared to the cost function based on the integral of the error.