Automatic collision avoidance based on supervised predictive controllers

In general, adaptive predictive control techniques have been applied to systems presenting slow dynamics or large delays. In cases where fast maneuvers are considered, their parameters estimation function and their control computation function must be enhanced to cope with these transient situations. Here a supervision layer, based on fuzzy logic techniques, is introduced to realize the on-line adaptation of identification and control law parameters. This approach is applied to the case of aircraft collision avoidance maneuvers.