Huawei recently became the first mobile device manufacturer in the world to use an AI-powered smartphone to actually drive a car. The company has been developing an algorithm for their ‘RoadReader’ project. The project pushes the boundaries of Huawei’s object recognition technology and puts the learning capabilities, speed and performance of its AI-powered devices to the test.
How is this different to other driverless cars?
A key difference is the technology used itself. While most autonomous cars available at present require specially developed chips by 3rd party technology providers to harness the computing power required, the team at Huawei are using technology already available in its smartphones. Using the processing power of the Kirin 970 found in Huawei’s flagship Mate 10 smartphone, the team have demonstrated its functionality and ability to stand up to even the most advanced technology developed for use in self-driving cars.
Whereas other driverless cars just detect obstacles, Huawei’s views are that it won’t just see obstacles, but will also understand its surroundings and change accordingly. To better understand this, Huawei took a regular Porsche Panamera and converted it into a driverless vehicle.
A number of video cameras mounted to the top of the vehicle captured footage that was then fed to the NPU which used object recognition. In case you didn’t know, the Huawei Mate 10 Pro contains a neural processing unit or NPU, that can perform artificial intelligence tasks faster while also using considerably less power than its regular processing chip.
If an object was detected, the vehicle was programmed to either stop or swerve in order to avoid the object based on the passenger’s preference. The vehicle doesn’t just see, but also understands its surroundings and adapts accordingly. This actually means that it can distinguish between hundreds, perhaps thousands of objects and take action accordingly.
Putting all that power to use
In a video on their official YouTube channel, Huawei has capitalized on this by showcasing the RoadReader project in a video. The almost 1 minute and 40-second video takes key moments of the development of RoadReader and compresses it to highlight the fact that all this was done in just 5 weeks.
Andrew Garrihy, Chief Marketing Officer of Huawei Western Europe stated “Our smartphone is already outstanding at object recognition. We wanted to see if in a short space of time we could teach it to not only drive a car but to use its AI capabilities to see certain objects and be taught to avoid them. If our technology is intelligent enough to achieve this in just 5 weeks – what else can it make possible?”
Huawei will be showcasing the ‘RoadReader’ project and the vehicle’s capabilities at a two-day event at Mobile World Congress in Barcelona (MWC) from 26th to 27th February 2018.