Have you tested a Hövding? We have.
Once it is installed in Hövding, we need to validate the AI-created algorithm. This involves doing the same tests again and having stunt persons cycle and fall in every conceivable manner to ensure that the algorithm reacts correctly. When this has been done and any fine adjustments have been made, Hövding is ready for certification.
For a product designed to protect against injury and save lives, the development work must be extremely precise. Over the years of development of Hövding, every component has been analysed and tested in minute detail to guarantee the safety and reliability of the product. Nothing is left to chance in our striving for maximum safety, user-friendliness and comfort. This is why we need to conduct tests. Again and again and again.
Two types of test
The tests Hövding undergoes can broadly be divided into two groups: development tests and certification tests. The development tests are conducted by us at Hövding in our work to create the safest cycle helmet in the world. The finished product then undergoes certification tests in which RISE (Research Institutes of Sweden) checks that Hövding complies with standards and may be marketed.
‘We raise the bar.’
Technical safety and user-friendliness
Heino Wendelrup is our CTO at Hövding. What he doesn’t know about cycle helmet tests is not worth knowing. He and his colleagues carry out technical tests to establish the technical functionality of Hövding. They also employ focus groups of consumers and test riders to analyse user-friendliness, handling and ergonomics. Heino explains here how Hövding’s own tests go further than the certification tests:
– When we carry out our airbag tests, we apply considerably more impacts than in certification tests, and we do them at higher speeds than standard. We don’t just want to meet the requirements. We want to exceed them. So we raise the bar to be on the safe side, and this ultimately benefits the consumer in the form of greater safety.
Airbag tests
Hövding’s airbag is an ingenious design that makes use of the benefits of airbag technology and adapts it to the reality and needs of cyclists. As cyclists are so unprotected in traffic (compared with drivers, for example, who have seat belts, airbags and crumple zones), the cycle airbag needs to provide extremely high protection and perform to the maximum. Which also means that the tests have to be the best. Heino explains:
– The transformation from collar to helmet is a critical moment and therefore we have carried out thousands of tests and inflated thousands of airbags over the years. In practical terms, we film the airbag at a high frame rate from various angles in a controlled laboratory environment so that we can follow the entire process. We do this on both dummies and real people with different head sizes.
The process in which the collar is converted into a helmet takes just a few milliseconds. Using the high-speed films, we ensure that the airbag unfolds efficiently and enfolds the head correctly. When it does so, it is time for the next stage.
Out into reality
After the airbag has been fully tested in the lab, we go out to do some live deployments with stunt persons. This ensures that the airbag also unfolds correctly when the wearer is riding a bike.
‘If you hit the ground at an oblique angle with your head, the injuries are more serious than if your head hits is straight on.’
One of the major advantages of Hövding is its ability to protect the head against oblique impacts against the ground. This is where the airbag is tested to ensure it really does what it promises in this respect. Heino Wendelrup again:
– We test rotational force caused by oblique impacts. Because this is what happens in reality. If you hit the ground at an oblique angle with your head, the injuries are more serious than if your head hits is straight on. Hövding’s airbag creates a flexible mobile zone between the head and the surface that effectively eats up the energy that causes the harmful rotational force. This is what makes Hövding so good, the fact that it deals with the rotational force, whatever the crash is like.
Hövding’s brain is based on artificial intelligence
Parallel to the development of the airbag, just as much time and care was devoted to developing Hövding’s other fundamental component, the algorithm that controls the airbag. The task of the algorithm is to read the cyclist’s patterns of movement. When it recognises a pattern that indicates an accident, the algorithm triggers the airbag at exactly the right moment. But how does the algorithm know when that is?
‘To teach the algorithm, we expose our stunt persons to triggered accidents.’
The answer is Artificial Intelligence ( AI). In brief terms, we feed large quantities of data into a computer which then learns to distinguish between different patterns of movement and creates an algorithm that reacts when an accident happens.
Black ice and other everyday events
To teach the algorithm, we expose our stunt persons to triggered accidents. Heino describes this hard job, in which you need to be tough:
– We insert sticks into the front wheel so they fly over the handlebars. We put concrete barriers in the road. We put soap on the floor so that it is suddenly as slippery as ice and they tumble in all directions. And we drive into them in cars. For higher speeds we use dummies of course. We record all these patterns of movement and feed them into the computer.
However, it is not just patterns of movement in cycling accidents that need to be recorded. It is equally important for the algorithm to understand when there is NO accident. So we use volunteers to cycle for thousands of hours and record their patterns of movement. We also have them do everyday things such as dropping and picking up keys, looking in various directions, bending down to fix something on the bike, etc. All of these events and patterns of movement are fed into the computer, which we subsequently instruct to distinguish the harmless movements from the dangerous ones. Using machine-learning, the computer learns the difference between the two and creates the algorithm that we put in Hövding to control the airbag.
Then we do it all again
Once it is installed in Hövding, we need to validate the AI-created algorithm. This involves doing the same tests again and having stunt persons cycle and fall in every conceivable manner to ensure that the algorithm reacts correctly. When this has been done and any fine adjustments have been made, Hövding is ready for certification.
‘Hövding is not only the safest cycle helmet on the market, but also the smartest.’
– Everyone is talking about how AI can be used in the future in self-driving cars, for example, but we are already using it in Hövding. Over the years, we have collected data from other five thousand simulated accidents and tens of thousands of hours of cycling data. This makes Hövding not only the safest cycle helmet on the market, but also the smartest, says Heino Wendelrup in conclusion.
