Constraint-based design and simulation of visual predictive architectures

In this paper we present a new constraint-based design of visual predictive architectures. An analytical study over the properties of a local model based predictor and its open loop behavior is detailed. Constraint-based reference trajectory generator and unconstrained optimization block are the new components of the visual predictive control strategy. Simulations are employed to validate the predictive technique and also to reveal the performance of the servoing system.