A memory of motion for visual predictive control tasks
This work addresses efficiency in visual predictive control for robotics, but it is incremental as it applies standard regression techniques to a known bottleneck.
The paper tackles the problem of efficiently achieving visual predictive control tasks by using a precomputed memory of motion to provide warm-starts and waypoints, enabling high performance with limited computational time in simulations and experiments with a 7-axis manipulator.
This paper addresses the problem of efficiently achieving visual predictive control tasks. To this end, a memory of motion, containing a set of trajectories built off-line, is used for leveraging precomputation and dealing with difficult visual tasks. Standard regression techniques, such as k-nearest neighbors and Gaussian process regression, are used to query the memory and provide on-line a warm-start and a way point to the control optimization process. The proposed technique allows the control scheme to achieve high performance and, at the same time, keep the computational time limited. Simulation and experimental results, carried out with a 7-axis manipulator, show the effectiveness of the approach.