Intel-owned Mobileye says it has cooked up a safety framework for fully autonomous, human-independent, driverless cars – and desperately wants people to notice this before the "inevitable fall" of public interest in driverless tech.
In a formal paper written in August, Mobileye reckons its Responsibility Sensitive Safety model will magic away all the thorny problems, such as "who is responsible for an accident", currently vexing the world's sharpest autonomy tech minds.
The company calls this its "formal model of blame" and says it will improve the safety of driverless cars and be economically viable for world+dog to adopt – though details on that viability are, apparently, reserved for a "future post" on the topic.
A plain-language translation of the full paper is available on the Intel website (PDF, 8 pages). It boasts: "Our proposed model provides a detailed, practical, and efficient solution for designing and validating an AV system that results in drastically improved safety."
The paper itself begins with a portentous warning about the Winter of AI, in which the hype cycle over robots being able to run the world for humanity's benefit collapsed in the 1980s after conspicuously failing to deliver anything it had promised.
"We believe that the development of Autonomous Vehicles (AV) is dangerously moving along a similar path that might end in great disappointment after which further progress will come to a halt for many years to come," wailed lead author Shai Shalev-Shwartz. This might be the case in America but British academics are taking it rather seriously with practical trials.
Mobileye has, it says, partnered with BMW to introduce this model as a "nonexclusive platform" into the German firm's vehicles. The model it proposes boils down to a mathematical formula, according to parent company Intel, and, among its other benefits, allows other driverless technology devs to validate their own software using the model.
Intel bought Mobileye in March for $15bn. The firm specialises in autonomous vehicle technology, including an in-car camera system.
The full paper can be read on academic repository Arxiv as a 30-page PDF. It is neither for the time-challenged nor those who don't have more than a passing familiarity with advanced mathematics. ®