Section: New Results
Model Predictive Control: distributed formulations and collision avoidance problems
Participant : Sorin Olaru.
In  , the mixt integer techniques have been analysed in the distributed model predictive control context, underlining the dependence of collision avoidance mechanism on the obstacle modeling and susequently on their treatement inside optimization-based control techniques as MPC (model predictive control). On the same topic of adversary consraints, a geometrical conditions has been establised in  for the local stabilization of a linear dynamics on a boundary of a forbidden region in the state space.
The theoretical developments from the last two years on the MPC design for multi-agent control problem led to the succesful application of receding horizon flight control for trajectory tracking of autonomous aerial vehicles  . In the same line or research, the predictive control for trajectory tracking and decentralized navigation of multi-agent formations has been presented in  .
In  a Characterization of the Relative Positioning of Mobile Agents for Full Sensorial Coverage in an Augmented Space with Obstacles is presented in view of a MPC control design.
A predictive control-based algorithm for path following of autonomous aerial vehicles has been proposed in  to improve the previous trajectory tracking mechanism. The ultimate goal oof both schemes is to avoid the real-time infeasibility problems in MPC.
The distributed predictive control mechanisms have been used for the control of a four interconnected tanks benchmark  , proving the versatility of a nonlinear Distributed MPC technique previously proposed by A. Grancharova.
In  the distributed Model Predictive Control of Leader-Follower Systems has been studied using an interior point method with efficient computations leading to simple tuning mechanisms for the cost functions and the terminal sets of the local MPC sub-problems.