Exoskeletons can help people whose freedom of movement is limited. In practice, controlling these exoskeletons is one of the biggest challenges. A new system puts artificial intelligence (AI) and allows an exoskeleton to think independently with this.
This is a system being developed by researchers at the Waterloo Artificial Intelligence Institute (Waterloo.ai) at the US University of Waterloo. Exoskeletons are not new. A lot is happening in this field in the Netherlands, among other countries. Consider paraplegic patient Ruben de Sain, who was able to regain his own steps in 2015 thanks to an exoskeleton. De Sain had to return the exoskeleton after a trial period. However, he liked the exoskeleton so much that Ruben started a crowdfunding campaign in 2016 to buy his own exoskeleton. The exoskeleton has a price tag of 80,000 to 100,000 euros.
Limitations
However, current exoskeletons have their limitations. For instance, an exoskeleton is equipped with all kinds of motors that basically act as muscles. In practice, users control these motors manually. This is not only inconvenient, but also cognitively demanding. For instance, every time users want to move their leg, they have to consciously instruct them to do so.
The new system the University of Waterloo is working on should provide a solution. In this system, users do not control the exoskeleton manually, but this process is automated. In practice, this means that the exoskeleton anticipates both the movements of its user and the immediate environment in which it is located. For example, sensors allow the system to recognise stairs and doors, among other things, and adjust its movements accordingly.
Autonomous exoskeletons that walk by themselves
"Our control approach does not necessarily require human thought," says Brokoslaw Laschowski, who leads the ExoNet project at the University of Waterloo. Laschowski is working on the project under the direction of professor John McPhee, the Canada Research Chair in Biomechatronic System Dynamics. "Similar to autonomous vehicles that steer themselves they design autonomous exoskeletons that walk themselves," he said.
Such an exoskeleton consumes energy. The energy efficiency of the system is therefore a key focus. Among other things, the researchers are looking at using human movements to generate electricity and thus charge the built-in batteries. Initial experiments show that exoskeleton if worn by healthy young adults can generate 19 joules per walking movement, while sitting and standing up generates 26 joules. The researchers still want to map how much energy is generated if seniors or rehabilitation patients wear the exoskeleton.
In the next phase, researchers focus on sending instructions to the exoskeleton's motors. This should allow the exoskeleton to climb stairs or avoid obstacles, for example, without user input.
Author: Wouter Hoeffnagel
