The entire automobile industry seems to be focused on one thing only: a successful 100% autonomous car. The autonomous vehicle technology has come a very long way through the years. However, every once in a while, we hear about an autonomous vehicle that was in some kind of accident. There is another issue under consideration; dynamics of the humans taking control of the self-driving cars. IBM has patented a new technology to determine the conditions, which decide about the control of the vehicle, whether the human driver should control it or not.
The cognitive system licensed by IBM will make driving decisions by using the indicators of human fatigue and emotional state. Sensors in the car will monitor the driver for various physiological factors like heart rate, gaze direction, and attention level. If this system finds symptoms of distraction or another anomaly, it will decide to take control of your car. In addition to the monitoring of human driver, the system will also keep an eye on the technical aspects and mechanical function of the car. If the system finds that a person must control this situation or obstacle, it will prompt the human to take over the control again.
The envisioned technology is the ‘third intelligence’ that will keep an eye on both, self-driving system and the human driver. If there is a minor fault in the car, the system will prompt the user to take control. It happens after it performs an analysis to check if the human is prepared or not. If both the driver and car are unfit for driving, it would automatically slow down the car to stop it at some safe location nearby. It can also learn from the traffic patterns and accident history of a vehicle to ensure that it has enough awareness of its surroundings. James Kozloski, who is a master inventor with IBM Research, explained:
“What we are doing is envisioning a self-driving vehicle that is able to assess the readiness and risk associated with a human taking control of the vehicle, given some anomaly on board.”
In the aftermath of a potential accident, a person may be in shock and unable to drive properly. The system will read the symptoms of distraction or distress and drive itself until the human driver calms down. Associate research professor at Carnegie Mellon University Aaron Steinfeld says that the transition from a computer to human is quite complicated. When transitioning from the driver to the AI, it can be as simple as pressing a button, but when the control is reversed, it becomes a tad bit complex.
In a scenario like this, it is quite hard for the AI to figure out how active a person is. “If you’re in a traffic jam, you kind of zone off, and barely pay attention—you from the outside look very similar to someone who’s paying attention. This is why it’s a hard problem,” says Steinfeld.
As far as the trust on AI and its safety is concerned, Waymo, which was previously a Google autonomous car project, has a lower crash-per-mile rate than conventional vehicles. If something goes wrong, humans are always present there to take over the control of the vehicle.
IBM has patented the technology already. Now, all we can do is to wait and see how long it will take to implement this system on an autonomous car successfully.