How Yandex.Taxi is using automation to detect drowsy and dangerous driversby Paul Sawers
In its two decades in business, Yandex has been called the Russian Google, Amazon, and Spotify, mostly due to the Moscow-based tech giant’s expansive reach into every nook — including online search, music streaming, email, maps and navigation, video, and more. In 2011, Yandex launched a mobile taxi-hailing service called Yandex.Taxi, leading to the inevitable “Uber of Russia” proclamations. Then in 2017, Yandex.Taxi and Uber merged their operations in the region to launch a new joint venture targeting Eastern Europe.
Yandex.Taxi now operates across the Commonwealth of Independent States (CIS), in addition to a handful of markets elsewhere in Europe, the Middle East, and Africa. The company has followed a trajectory similar to Uber’s, insofar as it now also offers food delivery, and in 2018 it launched one of Europe’s first public self-driving taxi services as part of a limited pilot.
But safety has emerged as a focal point in the ride-hailing realm, with concerns around everything from fatigue to driver identity. Back in November, Uber lost its London license — pending another appeal — after regulator Transport for London (TfL) reported a “pattern of failures” and breaches that “placed passengers and their safety at risk.”
One of the issues TfL identified is how easy it is for drivers without background checks to use legitimate drivers’ Uber accounts to pick up passengers. In response, Uber revealed plans to launch facial recognition technology that would require U.K. drivers to verify themselves periodically before rides, similar to what’s done in the U.S. market. Uber has also previously sought to allay fears about fatigue by restricting drivers to 12 hours behind the wheel before they’re forced offline for a six-hour break.
Meanwhile, Yandex has been keeping a close eye on proceedings at Uber and has been developing a slew of technologies to avoid some of the pitfalls encountered by its big-name rival.
Yandex has been quietly developing AI-infused proprietary hardware and software that monitors drivers’ attention levels. While similar technology is being built into fancy new cars, such as the Subaru Legacy 2020, Yandex’s incarnation can be easily retrofitted to any vehicle, and the company hopes to see ride-hailing drivers take advantage of the technology. It’s worth noting that the system is similar to one currently being piloted by Chinese ride-hailing giant Didi.
The company says its SignalQ1 camera looks at 68 points on a driver’s face and — with the help of machine learning — detects when a driver is tiring or distracted. To do so, the system looks at factors such as blinking and yawning and then attributes a sleepiness and distraction score.
The system is currently being tested in a small number of cars in Moscow. For now, the alert is limited to an audible beep — but in the future the camera will link directly with the driver’s Yandex account through the mobile app, meaning the company will be able to take preventative actions if it deems the driver unsafe.
“Whenever the driver gets tired, [they] will be notified and suspended from receiving further [ride] orders until [they get] some rest,” noted Aram Sargsyan, Yandex.Taxi’s regional general manager for EMEA and CIS, at the Move 2020 mobility conference in London this week.
Deploying this kind of technology at scale could be a challenge, given that Yandex.Taxi claims hundreds of thousands of drivers across 18 countries. However, while Yandex does allow drivers with their own cars to operate on its platform, it also works directly with taxi fleets in most of its markets, which could ease the path to wide-scale deployment. “We can work with our partners and find a way to implement [the technology] en masse,” Sargsyan told VentureBeat.
Yandex is also in the early stages of developing a facial recognition system, similar to Uber’s, that identifies who is really behind the wheel. “It’s in the test phase being developed, and we are optimizing it,” Sargsyan said.
Rather than requiring dedicated hardware, Yandex will simply use the camera on the driver’s smartphone, similar to what Uber and Didi are already doing. However, Yandex is going a step further, noting that it is also testing voice-recognition smarts to match the active driver with a registered account.
While Sargsyan didn’t offer any specifics about how prevalent driver identity fraud is in Yandex’s existing markets, he said, “We know that the problem exists.”
The main concern for Yandex is that the various regulators in its almost 20 active markets could begin to take note of the issue. Taking a cue from Uber’s woes in London, the company is working to prevent this practice from escalating into a bigger problem further down the road.
“In the 18 markets we operate in, [regulators] are strict in certain countries, but not as strict as TfL,” Sargsyan said. “So we are not waiting until this becomes a problem; we are trying to solve it now.”
Yandex has also been working on other automated safety technologies, including a speed control system that notifies drivers when they’re driving too fast. In the wake of this launch, speeding dropped 12-fold, according to Sargsyan. And similar to Uber, which has for some time been using telematic data to monitor drivers’ behavior on roads, Yandex also tracks driving styles and said it may suspend drivers who display erratic or aggressive behavior.
As we edge toward the advent of autonomous vehicles, a world Yandex.Taxi very much wants to be a part of having just passed 2 million self-driving miles, there has been a concerted effort to ramp up safety and security efforts — after all, truly self-driving cars are likely still years away from permeating society.
“I would say until we have self-driving cars replacing regular taxi drivers and car-sharing, we will have to implement all possible technologies to improve the safety,” Sargsyan said.