physics-based path planner

[christo]

Currently PathPlanner uses nearest unvisited neighbour to plan vector rendering. Due to the cost of velocity changes, it should choose next waypoint by minimising impulse (via dot product using impulse-momentum theorem and Newton's second law of motion) combined with distance to next point (will hopefully fall out of the former automatically). Minimising both blank time and impulse should produce a more optimal path. Ideally the estimated velocity and acceleration should be configurable for each display driver. Vectrex, cathode ray oscilloscope and laser scanner each have different such parameters.

[christo]

This does not depend on the simulator, although the simulator would enable empirically derived coefficients