Using Simulation Techniques To Improve Skeletal Plans For The Control Of A Vertical, Internal Grinding Machine

This research investigates the use of simulation techniques for the improvement of skeletal plans selected by a planner. These plans are used for control of a vertical internal grinding machine. Plans are selected using a description of the grinding task. These plans reflect the machinist's grinding knowledge. Once selected the plan is instantiated with the proper parameter values. The instantiated plans are passed to the grinding simulation where simulated force sensor readings emulate real-time signals. Feedback from the simulation is used by a knowledge-based system to modify successful plans and to identify failed plans.