Exoskeletons have been around for quite a while, originally aimed at augmenting soldiers capability (decreasing fatigue and helping in carrying heavy loads) and then progressively used to help patient with walking deficit, paralyses and during the rehab.
Current exoskeletons are good but not as much as designers would have thought: in addition to being bulky and having power needs that constrain their usability they have to undergo through a difficult and time consuming customisation.
Researchers at Carnegie Mellon University, CMU, have come up with an interesting solution: to involve the user, the human, in the training of the exoskeleton, what is called “human in the loop”.
They approached the tuning of the exoskeleton by analysing the metabolic rate of the user. This is an indication of the effort involved in the specific activity and of course the goal of an exoskeleton is to minimise this effort. The metabolic rate can be measure, with quite good precision, by analysing the consumption of oxygen, which, in turns, can be measured by analysing the breathing.
As you can see in the clip, volunteers have been asked to wear a face mask connected to a breath analyser. The data are analysed in real time by an algorithm developed at CMU (this is what makes the exoskeleton smart) and the result is used to try different exoskeleton configuration till a minimum effort is achieved. This leads to the customisation of the exoskeleton from that particular user, in that particular activity. The experiments have shown a decrease of 1/4 of the effort required by a specific activity as consequence of the finely tuning of the exoskeleton.
Obviously, it is not convenient to move around with with a face mask attached to a bulky breath analyser. However, this has been used to refine the algorithm. Researchers feel that now, with the algorithm polished up, it will be possibile to feed it with data derived from the heart beat, which are much easier to acquire, and even from muscle activity. These are not as accurate as measuring the metabolic rate from breathing but it is a good approximation that becomes very good over extended periods of time.
They expect to have this “human in the loop” becoming common in the next decade, improving rehab procedures and resulting in more effective human augmentation, where desired.
It is interesting to notice that the improved technology used in prosthetics coupled with interaction with the human body is leading to a significant improvement of the combined “cyber-entity”. The prosthetic is getting smarter, it is an autonomous system that becomes aware through interaction with the wearer and “learns” -evolve- over time. These are exactly the characteristics we are using to define a Symbiotic Autonomous Systems in our SAS Initiative.