I’ve recently been working with the DataDrivenDiffEq.jl package to implement system identification on non linear systems and have run into a bit of a road block. The model I’m trying to identify takes two control inputs and although there is plenty of documentation showing how to pass a single control input to the ContinuousDataDrivenProblem function I cannot find any that describes how one would pass two or more inputs. Does anyone know how
Managed to fix the issue by passing the control values as a matrix of control inputs of the form C x N where C is the number of separate control signals and N is the number of total inputs to the system e.g.: if you had a system with 2 control inputs that ran for 40 seconds and was sampled every 0.01 seconds C would be 2 and N would be 4000. Couldn’t get the ContinuousDataDrivenProblem function to take a function that generated multiple outputs as it kept generating arrays of the type any instead of float64.
1 Like