Carmakers are racing to meet new high-tech safety standards, such as automatic emergency braking systems, that are expected to take effect across Europe and the United States next year.
But while such systems are expected to save thousands of lives, consumers and insurers worry they will also push up the price of new vehicles and the cost of repairs, because they rely on expensive equipment like sensors and computer processors.
With cars becoming increasingly dependent on computing power, the cost of the onboard computers is even more material to the ticket price than the walnut and leather finish. Sensors built into once-cheap parts like bumpers to allow for automatic collision avoidance or self-parking have raised repair costs from $300 to $1,500, insurers say.
Israeli startup Brodmann17 hopes that its cheaper and hyper-efficient software will lower the cost of the sophisticated artificial intelligence and machine-learning applications behind these new driving features, allowing them to be fitted to more non-luxury vehicles.
“The technology is groundbreaking in terms of efficiency,” said Adi Pinhas, co-founder and CEO of Brodmann17, which is based in Tel Aviv. It can reduce the price of features like dashboard cameras that monitor driving by as much as 50 percent, making them more affordable to consumers, Pinhas said.
“These features today are mostly available in the premium-model vehicles,” such as BMW and Tesla, he said. “But more and more manufacturers are looking to bring this technology to the mass market.”
Brodmann17’s main breakthrough is a new type of computer algorithm, or computational process, that minimizes the number of calculations made on the data flooding in from cameras and other sensors on the car.
This new software allows artificial intelligence and deep-learning programs to run on cheaper standard computer processors, like those in mobile phones and laptops, rather than on expensive specialized processors usually required for data-heavy applications.
“The idea is that it is possible to develop neural networks in a way that they can reuse calculations from the past,” Pinhas explained. “The later stages in the neural network can reuse calculations from the earlier stages. This is how we are saving significant computing power and power consumption.”
The technology also reduces the amount of heat produced by processors, saving both power consumption and cooling costs.
For now, the company is focusing on the automotive industry, where AI and machine learning are on the rise, powering features like automatic emergency braking, keeping cars in their lanes and self-parking. The market for these features, known officially as Advanced Driver Assistance Systems (ADAS), is growing at a rate of more than 21% a year and is expected to be worth about $189 billion by 2026, according to Allied Market Research.
Investors in the company, which is currently raising a Series A round, include Sony Innovation Fund, Samsung NEXT, Maniv Mobility, lool Ventures and OurCrowd. Before starting Brodmann17, Pinhas founded the successful computer-vision startups Vigilant Technology and JustVisual.
“The car is really becoming a computer,” said Yuval Engelstein, mobility sector analyst at Start-Up Nation Central, which tracks tech companies in Israel, home to a growing automotive technology sector. “But you need really fast computing power because this is the brain of the car, and it has to work in real time, and in better than real time. This technology has to become cheaper in order to scale.”
Brodmann17 has formed partnerships with automotive manufacturers and automotive parts companies to install its technology into new vehicles. It can also be retrofitted when fleets are updated with new safety features. Pinhas expects the first cars using Brodmann17’s software to reach the market in 2022, and said several vehicle fleet management companies have already added features using the technology.
The company says its software can be used with most existing hardware, including cameras and other sensors, so it can be easily integrated into the manufacturing process.
Carmakers “are looking for software solutions. The main challenge in the market today is to bridge between how much it costs – and it’s expensive – and how much people are willing to pay for these technologies,” Pinhas said.
Meanwhile, looming deadlines for advanced automatic safety features, including voluntary industry standards in the United States and regulatory requirements in the European Union, are raising demand for better and cheaper solutions.
“There is continuous pressure, there is regulation,” Pinhas said. “Above all, people like these features that are promoting safety, comfort and fuel efficiency. This is forcing the companies to integrate more solutions.”
Learn more about Brodmann17 HERE.