Exo and Go

Exoskeletons are an emerging assistive technology designed to enhance mobility and physical capabilities for individuals with disabilities, or those engaged in physically demanding tasks. These wearable robotic devices mimic the structure and function of the human skeleton, providing support, strength, and endurance to the wearer. By augmenting natural movements, exoskeletons offer significant potential in improving quality of life and enabling greater independence.

Some exoskeletons focus on enhancing strength and endurance, enabling users to lift heavy objects or sustain physical exertion for longer periods. Others are geared towards mobility assistance, aiding individuals with walking, climbing stairs, or navigating rough terrain. Advanced exoskeletons may even incorporate intelligent control systems and sensors to adapt to the user's movements in real-time, providing a seamless and intuitive experience.

But despite their promising capabilities, exoskeletons have not yet seen much use outside of research laboratories and other specialized settings. One major reason for this is the complexity associated with calibrating such a device for a new user. Frequently, these systems also require adjustments for each activity the wearer engages in. This level of customization is necessary to ensure optimal performance and safety but can be time-consuming and cumbersome for both users and caregivers.

As a result, simpler, but less effective solutions continue to dominate the assistive technology landscape. Engineers at the Georgia Institute of Technology and MIT have an idea that could make exoskeletons much more practical in the future, however. They have developed a universal control system for partial-assist hip exoskeletons that adapts them to each individual user without a lengthy calibration process. Moreover, by focusing on the user’s motions rather than the types of activities they are engaged in, it was demonstrated that there is no need to switch between modes when the wearer transitions from walking, for example, to climbing stairs. The controller handles the shift in assistance needs automatically.

In order to build this universal controller, the researchers turned to deep learning. The model was trained on a large volume of force and motion-capture data that was collected in the team’s lab. It was collected from a variety of individuals wearing a hip exoskeleton while they walked, climbed ramps and stairs, and also as they transitioned between these activities. The diversity of data that was collected allowed the model to learn how to handle a wide variety of scenarios, and how to efficiently help different people.

The system was assessed in a series of experiments to better understand how it would work for real-world users. It was found that when paired with a hip exoskeleton, the controller reduced the metabolic and biomechanical effort of its users. In short, they used less energy to complete a unit of work, and their joints got a break as well.

In addition to assisting those with disabilities, this device could also give a boost to those with physically demanding jobs, reducing the risk of musculoskeletal injury. There are also opportunities to utilize this technology for rehabilitation — one day, it could help stroke patients to relearn how to walk, for example.

Advances in technology can make a tremendous impact on the lives of those living with disabilities. And in today’s world, where inexpensive and powerful computing and sensing systems are widely available, one does not need the backing of a major research university or other large organization to start innovating. Do you have an idea that might make a positive difference in the lives of others? If so, you will want to check out Hackster’s BUILD2GETHER 2.0 Inclusive Innovation Challenge. There are some big prizes for the best projects, and if you hurry, there is still time to snag some free hardware!