Exoskeleton data labeling system and method
Inventors
Swift, Timothy Alan • Cox, Nicolas • Kemper, Kevin Conrad
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
An exoskeleton network. The exoskeleton network includes an exoskeleton system that has one or more sensors and a memory; a user device that is local to the exoskeleton system and that operably communicates with the exoskeleton system; and a classification server that operably communicates with at least one of the exoskeleton system and the user device. The exoskeleton network performs feature extraction on sensor data obtained from the one or more sensors to generate feature-extracted sensor data and performs label derivation on the feature-extracted sensor data to generate labeled sensor data.
Core Innovation
An exoskeleton network is described that includes a wearable pneumatic exoskeleton system with a plurality of pneumatic actuators and a pneumatic system configured to introduce pneumatic fluid to actuate the plurality of pneumatic actuators. The wearable pneumatic exoskeleton system includes an exoskeleton computing device with a plurality of sensors, a memory storing at least a classification program, and a processor executing the classification program, and the processor controls the pneumatic system based at least in part on classifications generated by the classification program of sensor data obtained from the plurality of sensors.
The architecture includes a user device local to the wearable pneumatic exoskeleton system and operably communicating with the wearable pneumatic exoskeleton system. The wearable pneumatic exoskeleton system processes at least a portion of the sensor data to generate processed sensor data, sends the processed sensor data to the user device, and the user device performs feature extraction on the processed sensor data to generate feature-extracted sensor data for transmission to a classification server remote from the user device and wearable pneumatic exoskeleton system.
The classification server performs label derivation on the feature-extracted sensor data to generate labeled sensor data, and generates updated classification program data based at least in part on the labeled sensor data. The classification server sends the updated classification program data to the user device, and the user device sends the updated classification program data to the wearable pneumatic exoskeleton system, which replaces a current classification program in memory with the updated classification program embodied in the updated classification program data.
The subject matter addresses reconsideration of classification and refinement of classification programs using sensor data from before and after an interest time, rather than fixed finite-state-machine approaches for maneuver and intent recognition. Label derivation and updated classification programs support reconsidering classification of one or more maneuvers after the one or more maneuvers have already been classified by the exoskeleton system, and support automatic update cycles and tuning adaptation speed based on user confirmation or rejection as described in the summary content.
Claims Coverage
The provided independent claims define an exoskeleton network with multi-level processing and update loop. Three inventive features are present: pneumatic exoskeleton actuation controlled by classifications, partitioned processing with feature extraction on a local user device and label derivation at a remote server, and replacement of a current classification program in exoskeleton memory with an updated classification program that supports reconsideration of maneuver classification after initial classification.
Wearable pneumatic exoskeleton actuation controlled by classifications
A wearable pneumatic exoskeleton system including a plurality of pneumatic actuators associated with body parts of a user and a pneumatic system configured to introduce pneumatic fluid to the plurality of pneumatic actuators; an exoskeleton computing device with a plurality of sensors, memory storing at least a classification program, and a processor executing the classification program that controls the pneumatic system based at least in part on classifications generated by the classification program of sensor data obtained from the plurality of sensors.
Local feature extraction with remote label derivation and update loop
Processing at the wearable pneumatic exoskeleton system to generate processed sensor data, sending the processed sensor data to a local user device, performing feature extraction on the processed sensor data to generate feature-extracted sensor data, sending the feature-extracted sensor data to a remote classification server, performing label derivation to generate labeled sensor data, generating updated classification program data based at least in part on the labeled sensor data, sending the updated classification program data to the user device, and sending the updated classification program data to the wearable pneumatic exoskeleton system.
Replacing current classification program with updated classification program for reconsideration
Replacing, by the wearable pneumatic exoskeleton system, in the memory of the wearable pneumatic exoskeleton system, a current classification program with an updated classification program embodied in the updated classification program data received from the user device.
Reconsidering classification of maneuvers using labeled sensor data
Performing feature extraction on sensor data to generate feature-extracted sensor data; performing, by the classification server, label derivation on the feature-extracted sensor data to generate labeled sensor data; and replacing in the memory of the exoskeleton system a current classification program with an updated classification program generated by the classification server based at least in part on the labeled sensor data and reconsideration of classification of one or more maneuvers after the one or more maneuvers have already been classified by the exoskeleton system.
Across the independent claims, the inventive coverage centers on an exoskeleton network that uses sensor data to drive a classification program controlling exoskeleton actuation, partitions processing with feature extraction on a local user device and label derivation at a remote classification server, and updates the classification program stored in the exoskeleton memory. Additional independent-claim language explicitly ties updated program generation to reconsideration of previously classified maneuvers and limits roles and communication topology involving remote servers and local connectivity as described in the claim texts.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Not explicitly described in patent.
Interested in licensing this patent?