Transfer learning of deep learning system and method for task agnostic wearable robot control and human monitoring

A method for configuring a neural-network controller for a robotic or prosthetic device by:

• receiving biomechanical data acquired by a set of first sensors acquired for a first robotic or prosthetic device;
• generating simulated sensor data using a translator network to simulate sensors for configuring at least one neural network of a target device, wherein the target device has a set of second sensors for a second robotic or prosthetic device, wherein the first sensors and the second sensors are different types, wherein the simulated sensor data is generated using the received biomechanical data using physics-based simulation software, where the simulation is performed at body segments for positions, forces, and/or velocities constrained or defined on the same body segment for the target device to provide paired data of (i) the sensor data of the target device and (ii) the biomechanical data, and wherein the paired data is used for training of the at least one neural network;
• and
• training the at least one neural network, wherein the neural network is configured to map the simulated data associated with the biomechanical data to real data associated with the sensor data.

A system comprising:

• a processor;
• and
• a memory having instructions stored thereon for configuring a controller, wherein execution of the instructions by the processor causes the processor to:
• receive biomechanical data;
• generate simulated sensor data using a translator network to simulate sensors for a transferred device using the received biomechanical data generated using physics-based simulation software, where the simulation is performed at body segments for positions, for forces, and/or for velocities constrained or defined on the same body segment for a target device to provide paired data of (i) the sensor data of the target device and (ii) the biomechanical data, wherein the paired data is used for training of at least one neural network;
• and
• train the at least one neural network, wherein the neural network is configured to map the simulated data associated with the biomechanical data to real data associated with the sensor data.

A non-transitory computer-readable medium having instructions stored thereon, a processor is configured to execute the instructions, which include:

• receiving biomechanical data;
• generating simulated sensor data using a translator network to simulate sensors for a transferred device using the received biomechanical data generated using physics-based simulation software, where the simulation is performed at body segments for positions, for forces, and/or for velocities constrained or defined on the same body segment for a target device to provide paired data of (i) the sensor data of the target device and (ii) the biomechanical data, wherein the paired data is used for training of at least one neural network;
• and
• training the at least one neural network, wherein the neural network is configured to map the simulated data associated with the biomechanical data to real data associated with the sensor data.