Table 1.

Neural Network architectures and hyperparameters.

Joint Angles
ANN	Layers	Learning Rate	Dropout	Activation
MLP	6000–4000	0.0003	0.5	relu
LSTM	32–32	0.0003	0.7	tanh
CNN	3000–6000	0.00003	0.4	relu
Joint Moments
ANN	Layers	Learning Rate	Dropout	Activation
MLP	3000–1000	0.0003	0.5	relu
LSTM	128–1024	0.0003	0.4	tanh
CNN	2000–4000	0.0001	0.4	relu