Predicting a signal based on other signals

I want to predict a signal based on other related signals, how would I go about doing this?

My current approach is to do some feature extraction (in the time and frequency domain) on both the ground truth signal and on the input signals. I use the features that I calculated on my input signals to predict the ground truth signal with basic regression models such as RandomForestRegressor or GradientBoostingRegressor models. I've used a rolling window approach with varying step/window sizes on my dataset in order to generate a lot of samples (after a train-test split so there's no overlap) and to look for the ideal window size. In the image below you can see the results of my current GradientBoostingRegressor model for the prediction of one of the features I calculated on the ground truth.

I also thought about building a GAN to generate synthetic time series data, is this realistic?

More practically, I'm using a set of accelerometers to predict forces (measured with force sensors which provide the ground truth) that act upon a bicycle.