When adopting a deep learning model for embodied agents it is required that the model
structure be optimized for specific tasks and operational conditions. Such optimization can …