Researchers have developed a soft, wearable robot that can learn the precise movements of the person wearing the kit to maximize the help the robot can provide. The aim is to provide better assistance to people suffering from impaired movement of the shoulder, arm or hands. 

The research, which was published in the magazine Nature Communications, was carried out by a joint team of researchers from the John A. Paulson School of Engineering and Applied Sciences (SEAS), together with physician-scientists at Massachusetts General Hospital and Harvard Medical School. 

Harvard bioengineers have been working on a soft, wearable robot for several years. The goal was a robot that could provide assistance to help people with upper body injuries, or neuro degenerative diseases such as Motor Neurone Disease (MND), or people recovering from strokes.  One difficulty has provide to be that people move in their own individual ways, so it's very hard to design a wearable robot that can work for different people. 

The latest research makes use of machine learning alongside the robotics to personalize the movements by using sensors that track motion and pressure to work out what the user is trying to do. This is based on a vest with sensors.

The team developed a machine learning (ML) intention detection model to infer arm motion from kinematic and human-robot interaction information captured using IMUs and custom soft compression sensors, respectively. An IMU, Inertial Measurement Unit, is an electronic device that measures and reports a body's specific force, angular rate, and sometimes the magnetic field surrounding the body.

The feedback was then used to control an inflatable actuator under the user's impaired arm that can be inflated and deflated to provide the mechanical assistance to the limb.  

Alongside the machine learning model, the research team included a model that estimates the minimum pressure needed to support the arm during movement. This overcomes the problem of providing too much assistance, making it hard for the user to then lower their arm. It also makes the help feel more natural and less intrusive for tasks such as eating and drinking.

The combination of the two models means the robot can increase or decrease the level of assistance based on what it has learned about how that user usually moves.

The soft wearable robot improved joint function, decreased trunk compensation, and increased hand path efficiency across participants in the trial for functional tasks.

The team acknowledges that this is just the start, saying that despite promising results, the proposed method and implementation have several limitations, replying on the users having sufficient intact upper limb joint function to perform movements for training the wearable devices. In addition, the method focused on assisting only the shoulder joint; however, individuals with severe impairments oftentimes have difficulty using multiple joints. Another limitation is that the current version of the robot cannot be independently donned by individuals with upper limb impairments. However, the results are promising. 

More Information

Personalized ML-based wearable robot control improves impaired arm function

Related Articles

A World First For Humanoid Robots

Robots Run Half Marathon - At Snail's Pace

Kawasaki Unveils Robot Horse

Spot With AI - The New Robotics

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin.

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info