Bernd John, Group Head of R&D and Technology at Brigade Electronics plc, discusses how AI is helping to enhance road and worksite safety and puts to rest some of the common misconceptions surrounding its ability and use.
What is AI?
Artificial intelligence is the development of computer systems to be able to perform tasks that normally need human input, such as decision-making and visual perception.
If the principles of AI are based on learning and thinking like humans, could this cause potential danger to real-life humans on roads and worksites?
AI works by learning from data sources and comes up with an answer based on this information, also called machine learning. In automotive technology, we are utilising meticulously validated data sources which in our work is primarily used for image detection. The AI for a vehicle electronics which is ‘pre-trained’ before it is installed. It is programmed to operate in a very specific way, for example to detect particular situations or vulnerable road user (VRU) – such as pedestrian or cyclist – and therefore it can’t begin to develop a life of its own and start to decide things for itself. This is essential making it safer – for Brigade’s products, it’s ideal because they help to make informed detections for the driver or operator in potentially difficult situations.
How is AI helping to improve road and worksite safety?
AI has enormous potential to improve safety on roads and worksites, but, as I mentioned above, it’s dependent on pre-learned scenarios and precision. It draws on its data sources and makes a detection decision based on this information, so any AI used in automotive technology, such as we produce at Brigade, has to use validated data sources which have been rigorously checked. It’s very secure as it’s been pre-trained and programmed to act in a specific way. For example, a camera will have been programmed to respond in a particular way if a cyclist veers in front of a vehicle. The pre-programming means it will make the same decision every time because it is a predetermined pattern. This eliminates detection decision-making which could potentially lead to a catastrophic outcome. The machine isn’t defined by emotion, but by pre defined pattern, so it’s more decisive. When it needs updating, we can change the current algorithm with a new refined one.
Should we be concerned about how much AI is taking over?
In the vehicle safety industry, AI should be able to avert for potential collisions, driver assistance systems, enable self-drive vehicles, intelligent traffic management systems, and all based on data-driven decision-making. The potential for safer roads is huge. Currently we’re using AI as a supportive system and, provided the checks are in place, it is a reliable method of enhancing safety, which has to be a positive step. In our industry, I would be more concerned if products came to market which hadn’t been tested as exhaustively as Brigade’s. That’s why it’s so important to buy from a trustworthy source.
Is AI trustworthy – can it fail or get things wrong?
In the vehicle safety, this is arguably one of the biggest concerns. We’ve all read about chatbots giving misleading, legally questionable and even extremely dangerous advice to people because the bot has collated its response from random or unvalidated sources. AI training is similar to human learning – the information it bases its response on has to be credible and true. If it isn’t then issues can arise, and of course it’s absolutely essential to test AI in the real world, because there will always be situations nobody has envisaged. A good example of this is a site worker in a hi-vis vest. The glare from it can create a halo if the AI is not trained on such it will be impossible to get a proper detection ensured. There are cases with the sun is wiping out an image and can bring a vision-based AI system to its limit. Such keeps getting improved due to the advances in the camera technologies as well as with fusion with other sensor technologies e. g. radar. On the whole though, the better the information that are provided to an AI system, the more you can get out of it.
What are the benefits of using AI road safety systems compared to traditional devices?
Traditional detection systems measure physical dimensions but don’t classify what something actually is. They measure it and draw a conclusion as a result. With AI you can classify an object, such as a cyclist, a pedestrian, a moving car or a safety barrier. People can respond randomly to sudden incidents; a machine will always come up with a pre-defined reaction which is usually safer – but it depends on the intensity of the training! And if something goes wrong there will always be the issue of liability. Is an incident the responsibility of the driver, the manufacturer, the AI training material, a combination of the three or none of them?
In my opinion the best compromise for an application at the moment is a combination of AI and to combine image data with traditional radar detection technology, but this may change with advances in AI and further more adding a 3rd sensor technology into the mix.
Brigade has introduced a range of AI products to its portfolio. Can you tell us about these, how they work and how they improve safety?
We’ve been at the forefront of road safety for over 40 years, and we obviously wanted to investigate the use of AI to see if it can enhance our products further. Our first venture into AI technology was a camera which can detect pedestrians using AI with state-of-the-art machine learning. Our CAREYE® Safety Angle Turning Assistant uses AI to evaluate received images accurately in the cameras fitted to vehicles and can calculate when nearby people or objects come critical. Its success inspired us to expand our range of AI products.
The cameras in our new generation of active blind-spot detection system have larger detection areas than the previous ones, recognise humans using AI and warn the driver of a vehicle visibly and/or audibly about a possible hazard. The system has an HD or higher resolution vison sensor with all processing power embedded in camera assembly, and the detection range stretches from the very front of the vehicle to several metres out. The camera’s computer uses machine learning to improve detection rates continually using a thoroughly tested state-of-the-art algorithm.
It goes without saying that all Brigade products are fully pre-trained, and our engineers have spent long working hours ensuring that our systems can cope with any situation in the shortest possible reaction time. Every practicable scenario has been tested; it’s not just a question of identifying a human. The human in question could be six foot four or a child at three foot. They could be standing, running, wearing a large hat, be on a forklift, pushing a wheelbarrow, sitting down, riding a bike – we even had to make an initial adjustment to allow for a human lying down. Another scenario is to train AI to recognise speed differentials and track objects; for example, it needed to learn that a cyclist and other vehicle in relation to own vehicle movement.
Will we see even more advances with AI in the future and what will these be?
Over the years we’ve gone from vehicles giving us warnings, such as petrol getting low to the active assistance of detection cameras. Cameras can currently help us with keeping in lane on a motorway, detecting vulnerable road users and warning us of blind spots. Fully automated cars are already becoming a reality and this market will be one to watch. Further advances in AI are inevitable, but if they continue to prevent death and injury on roads then here at Brigade, we’ll do everything we can to channel them securely and responsibly.
The nature of engineering is that it never stands still – we can always make improvements on what we’ve already achieved!
For further information about Brigade and our range of commercial vehicle safety products, please contact us.