How to convert transfer function to state space function in Neural Networks?

I obtained/extracted using the weights and biases, this transfer function from the Neural Networks toolbox: y1 = b2 +LW*tansig(b1+IW*x)

I want to implement it onto a controller, so I have to change it to a state space function in the form of: [A,B,C,D] = tf2ss(bf,af)

But i have only one equation, here it requires two variables(in the form of coefficients of each)

What do i do?