Tesla, Inc has unveiled Cybercab, a driver-assisted self-driving robotaxi, reported Reuters, adding that the robotaxi is yet to be road-ready.
As reported, Tesla’s robotaxi strategy relies solely on digital imaging, the computer-equivalent of vision, by using a combination of cameras the way humans use their eyes, supplemented by an artificial intelligence (AI) element called the end-to-end machine learning that can instantly process the images into driving decisions.
Tesla’s technology still needs a human driver to operate safely, and Tesla plans to use the same approach to develop fully autonomous robotaxis.
The robotaxi was introduced at a time when Tesla’s main competitors have dispatched robotaxi fleets to operate in select cities that are heavily mapped.
Tesla’s competitors, including Alphabet’s Waymo, Amazon’s Zoox, General Motors’ Cruise and a host of Chinese firms, use the same imaging and AI-processed technology, but typically add on layers of overlapping systems and sensors such as radar, lidar and sophisticated mapping to ensure safety and get regulatory approval for their driverless vehicles.
Tesla’s strategy is simpler, and much cheaper, but without the layered technologies used by its peers, Tesla’s system might not be adequate to tackle rare driving scenarios that cannot be anticipated by self-driving systems and human engineers.
The electric vehicle maker also faces another major challenge related to the end-to-end AI technology, which is described as a “black box” as it is nearly impossible for engineers to see what went wrong when the AI misbehaves and causes an accident. The inability to precisely identify and analyse such failures makes it difficult to safeguard against them.
End-to-end artificial intelligence involves training a computer to make decisions directly from raw data, with no intermediate steps requiring additional engineering or programming.
Several studies have shown that computer vision is highly accurate but still fails to recognise objects about 3% of the time, reported Reuters quoting Missy Cummings, a robotics and AI professor at George Mason University.
