At IBM InterConnect, IBM CEO Ginni Rometty talked about the three principles of augmented intelligence in this AI era: service to mankind, transparency, and skills.
She also discussed how IBM Watson can make us a better version of ourselves. There’s no better example of this than Honda R&D’s Driver Coaching System prototype that helps new as well as older drivers learn how to spot and avoid potentially dangerous road situations.
Honda R&D realized that both Japan and US markets have the same issues: an aging population and a growing share of young drivers. For both countries, it’s these two groups who are most likely to be in deadly auto accidents.
Honda R&D has analyzed the behaviors, skills and judgments that take place in real time as experienced drivers successfully encounter dangerous situations. By understanding the behaviors of very skilled drivers — how they gauge and react to risk — they can apply that to all drivers. The interest is in elderly and new drivers especially prone to accidents and how lives can be saved through actively coaching these drivers.
Good driving deconstructed
Let’s assume an experienced driver can spot danger a few seconds faster than a novice driver. With the Driver Coaching System, it can give that extra reaction time to drivers of all experience levels. In addition, it will support a new driver who faces the anxiety of operating a car. Watson acts as a safe driving coach by engaging in an encouraging conversation as the new driver builds confidence in driving and learns to spot dangerous situations.
I love to drive. I own a fairly exotic car and often “open it up” to feel the raw exhilaration of speed, power and control. My mind clears and focuses solely on the drive and the extremes of the car: the calculus of the curves, the feel of acceleration, timing of shifting and braking. It takes concentration, awareness, risk evaluation, and reaction. I have experience. But does that make me a good driver?
Honda R&D has deconstructed what it takes to be a good driver.
How does it work?
The Driver Coaching System is continually monitoring the driving situation or “the scene” and evaluates the speed (overall and relative to other cars), distance to surrounding objects, adherence to the lane, and braking times and distances. Watson uses that information to offer real-time guidance and advice in Japanese.
It’s also gauging the driver and classifying the driver’s skill and mental state based on changing behaviors and conditions.
I learned the hard way that I’m not a “good driver” all the time. I wrecked a rental car in an accident leaving an airport in an unfamiliar city. In that situation, I was more like a novice driver with the anxiety of driving an unfamiliar car in an unfamiliar place.
The Driver Coaching System is detecting whether a driver is driving outside their norms and is perhaps anxious, distracted or tired. With this information, the prototype can classify the driver’s current state: whether that’s normal, they’re attentive or inattentive, and driving conservatively or aggressively. It adjusts its coaching to fit the driver’s state of mind.
The scene and the driver’s behavior determine the coaching Watson provides to the driver. The goal is for Watson’s coaching to be timely, friendly, supportive and welcomed. Watson is coaching the driver in Japanese.
How is Watson’s Japanese?
Japanese is a high-context language. In Japanese, meaning is expressed between the lines. The listener has to have the background knowledge of multiple dimensions to grasp the intended meaning of what’ being said. As Watson continues to improve how it speaks and understands Japanese, it has to truly appreciate and apply this cultural context. This is quite different from how Watson originally learned English.
“A conversation with Watson is getting more accurate and showing improved understanding of human intention,” said Yoshimitsu Akuta, chief engineer, Honda R&D. “I look forward to more possibilities with the context of the Japanese language. I think both English and Japanese speakers will be excited to have conversations with Watson as their friend in the car.”
For the 26th year in a row, IBM led US companies in patents in 2018, with engineers, researchers, scientists and designers racking up a record 9,100 patents. Of those, more than 2,000 were related to cloud computing, “including a patent on a specialized monitor for unikernel-based virtual machines that takes a lightweight approach to improving […]
There are two million cars in Ecuador. Only 20 to 26 percent of them are insured. There’s no requirement to have car insurance in Ecuador and there’s a predominant attitude that “nothing will happen.” The average value of an auto insurance policy in Ecuador is $700, a considerable sum given that the basic monthly salary […]
For both drivers and companies that create technologies to keep drivers safer, time is everything. At AirWire Technologies, we have been blown away by how quickly the IBM Garage helped bring our solution to market. AirWire created the Connected Car Internet of Things (IoT) device, which plugs into a vehicle’s onboard diagnostic port and provides […]