# US 11,298,287 B2 · Systems and methods for a compressed controller for an active exoskeleton

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 11,298,287 B2
- Document ID: US11298287B2
- Status: Granted
- Kind: B2
- Application number: 17/028,761
- Filed: 2020-09-22
- Published: 2021-12-02
- Granted: 2022-04-12
- Earliest claimed: 2020-06-02
- Assignees: Dephy, Inc.
- Applicants: Dephy, Inc.
- Inventors: Luke Mooney; Jean-François Duval
- Web page: https://patentgrep.com/patent/US11298287B2

## CPC

- A61H 3/00: Appliances for aiding patients or disabled persons to walk about (apparatus for helping babies to walk A47D13/04 {; applying electrical currents by contact electrodes for stimulating motor muscles, e.g. walking assistance A61N1/36003})
- A61H 1/0266: {Foot}
- G06N 20/00: Machine learning
- G16H 20/30: relating to physical therapies or activities, e.g. physiotherapy, acupressure or exercising
- G16H 40/63: for local operation
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- G16H 50/70: for mining of medical data, e.g. analysing previous cases of other patients
- A61H 2003/007: {secured to the patient, e.g. with belts}
- A61H 2201/1207: with electric or magnetic drive
- A61H 2201/1215: Rotary drive
- A61H 2201/1642: Holding means therefor
- A61H 2201/165: Wearable interfaces
- A61H 2201/501: connected to external computer devices or networks
- A61H 2201/5012: using the internet
- A61H 2201/5043: Displays
- A61H 2201/5046: Touch screens
- A61H 2201/5061: Force sensors
- A61H 2201/5069: Angle sensors
- A61H 2201/5082: Temperature sensors
- A61H 2201/5084: Acceleration sensors
- A61H 2201/5092: Optical sensor
- A61H: PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY (electrotherapy, magnetotherapy, radiation therapy or ultrasound therapy A61N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA

## Abstract

A system to augment motion via a battery-powered active exoskeleton boot is provided. The system can include a controller and an electric motor that generates torque about an axis of rotation of an ankle joint of the user. The controller can receive sensor data associated with activity of the exoskeleton boot during a first time interval. The controller can determine, based on the sensor data input into a model trained via a machine learning technique associated with one or more users performing one or more physical activities, one or more commands for a second time interval. The controller can transmit the one or more commands generated based on the model to the electric motor to cause the electric motor to generate torque about the axis of rotation of the ankle joint of the user in the second time interval.

## Claims

### Claim 1

A system to augment motion via a battery-powered active exoskeleton boot, comprising:
a shin pad of an exoskeleton boot to couple to a shin of a user below a knee of the user;
one or more housings enclosing i) a controller comprising memory and one or more processors, and ii) an electric motor that generates torque about an axis of rotation of an ankle joint of the user, wherein at least one of the one or more housings is located below the knee of the user and coupled to the shin pad;
a battery holder coupled to the shin pad, the battery holder to receive a battery module;
an output shaft coupled to the electric motor and extending through a bore in a housing of the one or more housings enclosing the electric motor; and
the controller to:
receive sensor data associated with activity of the exoskeleton boot during a first time interval;
input the sensor data into a model trained via a machine learning technique using historical motion capture data associated with one or more users performing one or more physical activities;
generate a candidate torque profile based on the sensor data input into the machine learning model;
modify the model based on a comparison of the candidate torque profile and a target torque profile generated prior to receiving the sensor data;
determine, for a second time interval subsequent to the first time interval, one or more commands to match the candidate torque profile; and
transmit the one or more commands generated based on the model to the electric motor to cause the electric motor to generate torque about the axis of rotation of the ankle joint of the user in the second time interval.

### Claim 2 (depends on 1)

The system of claim 1, comprising the controller to:
receive, via a network, the model from a command modelling system that trains the model based on the historical motion capture data.

### Claim 3 (depends on 1)

The system of claim 1, comprising a command modelling system to:
receive historical video data associated with the one or more users performing the one or more physical activities;
identify, based on historical video information, one or more torque profiles corresponding to the one or more physical activities; and
train, using the machine learning technique and based on the one or more torque profiles, the model to cause the model to output the one or more commands responsive to the sensor data.

### Claim 4 (depends on 1)

The system of claim 1, comprising a command modelling system to:
receive the historical motion capture data comprising historical sensor data;
provide, for display via a display device communicatively coupled to the command modelling system, a visual indication of the historical motion capture data;
receive, via a user interface, an indication of the candidate torque profile corresponding to the visual indication of the historical motion capture data; and
train, using the machine learning technique and based on the indication of the candidate torque profile received via the user interface, the model to cause the model to output the one or more commands responsive to the sensor data.

### Claim 5 (depends on 1)

The system of claim 1, comprising a command modelling system to:
receive the historical motion capture data comprising historical sensor data;
provide, for display via a display device communicatively coupled to the command modelling system, a visual indication of the historical motion capture data;
receive, via a user interface, an indication of a type of physical activity corresponding to the visual indication of the historical motion capture data; and
train, using the machine learning technique and based on the indication of the type of physical activity received via the user interface, the model to cause the model to output the one or more commands responsive to the sensor data.

### Claim 6 (depends on 5)

The system of claim 5, wherein the type of physical activity comprises at least one of: walking, running, standing, standing up, or sitting.

### Claim 7 (depends on 1)

The system of claim 1, wherein the one or more physical activities comprise at least one of steady state activities or transient activities.

### Claim 8 (depends on 1)

The system of claim 1, comprising the controller to:
determine the one or more commands for the second time interval to match the candidate torque profile selected based on the sensor data via the model.

### Claim 9 (depends on 1)

The system of claim 1, comprising a command modelling system to:
receive the historical motion capture data comprising historical sensor data;
receive indications of types of physical activities corresponding to the historical motion capture data; and
train, using a second machine learning technique and based on the indications of types of physical activities corresponding to the historical motion capture data, a second model to generate a second torque profile based on second historical motion capture data.

### Claim 10 (depends on 9)

The system of claim 9, comprising the command modelling system to:
receive the second historical motion capture data;
determine, based on the second model, one or more torque profiles based on the second historical motion capture data; and
train the model based on the determined one or more torque profiles and the second historical motion capture data to cause the model to generate the one or more commands based on the sensor data.

### Claim 11 (depends on 1)

The system of claim 1, wherein the historical motion capture data corresponds to data collected via the exoskeleton boot in a plurality of states comprising: an unpowered state, a partially powered state, and a fully powered state.

### Claim 12 (depends on 1)

The system of claim 1, comprising the controller to:
receive, via a user interface, input from the user prior to the second time interval; and
generate, via the model, the one or more commands based on the input and the sensor data.

### Claim 13 (depends on 1)

The system of claim 1, wherein the sensor data comprises at least one of ankle joint data, inertial measurement unit data, or battery data.

### Claim 14 (depends on 1)

The system of claim 1, wherein the historical motion capture data comprises at least one of inertial measurement unit data, goniometer data, infrared reflector data, or force plate data.

### Claim 15

A method of augmenting motion via a battery-powered active exoskeleton boot, comprising:
providing a shin pad of an exoskeleton boot for coupling to a shin of a user below a knee of the user;
providing one or more housings enclosing i) a controller comprising memory and one or more processors, and ii) an electric motor that generates torque about an axis of rotation of an ankle joint of the user, wherein at least one of the one or more housings is located below the knee of the user and coupled to the shin pad;
providing a battery holder coupled to the shin pad, the battery holder to receive a battery module;
providing an output shaft coupled to the electric motor and extending through a bore in a housing of the one or more housings enclosing the electric motor;
receiving, by the controller, sensor data associated with activity of the exoskeleton boot during a first time interval;
inputting, by the controller, the sensor data into a model trained via a machine learning technique using historical motion capture data associated with one or more users performing one or more physical activities;
generating, by the controller, a candidate torque profile based on the sensor data input into the machine learning model;
modifying, by the controller, the model based on a comparison of the candidate torque profile and a target torque profile generated prior to receiving the sensor data;
determining, by the controller, for a second time interval subsequent to the first time interval, one or more commands to match the candidate torque profile; and
transmitting, by the controller, the one or more commands generated based on the model to the electric motor to cause the electric motor to generate torque about the axis of rotation of the ankle joint of the user in the second time interval.

### Claim 16 (depends on 15)

The method of claim 15, comprising:
receiving, by a command modelling system, historical video data associated with the one or more users performing the one or more physical activities;
identifying, by the command modelling system based on the historical video data, one or more torque profiles corresponding to the one or more physical activities; and
training, by the command modelling system, using the machine learning technique and based on the one or more torque profiles, the model to cause the model to output the one or more commands responsive to the sensor data.

### Claim 17 (depends on 15)

The method of claim 15, comprising:
receiving, by a command modelling system, the historical motion capture data comprising historical sensor data;
providing, by the command modelling system, for display via a display device communicatively coupled to the command modelling system, a visual indication of the historical motion capture data;
receiving, by the command modelling system via a user interface, an indication of the candidate torque profile corresponding to the visual indication of the historical motion capture data; and
training, by the command modelling system, using the machine learning technique and based on the indication of the candidate torque profile received via the user interface, the model to cause the model to output the one or more commands responsive to the sensor data.

### Claim 18 (depends on 15)

The method of claim 15, comprising:
receiving, by a command modelling system, the historical motion capture data comprising historical sensor data;
providing, by the command modelling system, for display via a display device communicatively coupled to the command modelling system, a visual indication of the historical motion capture data;
receiving, by the command modelling system via a user interface, an indication of a type of physical activity corresponding to the visual indication of the historical motion capture data; and
training, by the command modelling system, using the machine learning technique and based on the indication of the type of physical activity received via the user interface, the model to cause the model to output the one or more commands responsive to the sensor data.

### Claim 19 (depends on 15)

The method of claim 15, comprising:
determining, by the controller, the one or more commands for the second time interval to match the candidate torque profile selected based on the sensor data via the model.

### Claim 20 (depends on 15)

The method of claim 15, comprising:
receiving, by the controller via a user interface, input from the user prior to the second time interval; and
generating, by the controller via the model, the one or more commands based on the input and the sensor data.