AdaptablePrediction

Illustration of how to use nonlinear recursive least square to adapt prediction models online

Require Matlab2019a or above

Run adaptable_prediction.mlx to visualize trajectory prediction results under different circumstances. Tunable parameters:

1. system parameter
2. prediction horizon
3. forgetting factor

Nonlinear prediction model

x(k+1) = f(a(k), x(k))

Noninear recursive least square

[To be completed]

References

https://en.wikipedia.org/wiki/Recursive_least_squares_filter

https://en.wikipedia.org/wiki/Extended_Kalman_filter