next up previous
Next: Implementation Up: Experiments Previous: Cart-Pole Swing-Up Task

Simulation

For all the simulations and the training data set additive Gaussian observation noise with $ \sigma=0.001$ and Gaussian process noise with $ \sigma=0.001$ were used. For the performance comparison between NMPC and OIC, the length of the control horizon was set to 40 time steps corresponding to 2 seconds of system's time. The simulations were run for 70 time steps corresponding to 3.5 seconds of system's time to ensure that the controller was able to stabilise the pole.




Tapani Raiko 2006-08-24