Roadlink International Trading Limited outlines its sophisticated deep learning-assisted autonomous driving technology
Shenzhen, China / 20th June, 2020 / - Roadlink International Trading Limited, a China-based technology startup, has officially come out of stealth mode – a transitory period to avoid alerting competitors - and outlined some details of its highly anticipated autonomous driving technology.
The evolved startup is focused on developing autonomous driving technology assisted by deep learning software – a sophisticated form of artificial intelligence. The deep learning software is applied to everything the self-driving car performs, from recognizing obstacles around the vehicle to making intelligible decisions.
Roadlink International Trading Limited, which was founded in 2016, will be launching pilot programs with unnamed partners in the coming months to test the delivery of individuals and goods on fixed routes throughout Hong Kong. The pilot program will use route-based vehicle fleets to demonstrate the technology's capabilities. Eventually, the company will collaborate with automakers to integrate the technology within the vehicle.
"Deep learning software is the finest enabling technology for autonomous vehicles," said Lei Peng, CEO and Founder of Roadlink International Trading Limited. "You mainly hear about all these gadgets on the car: the cameras, sensors, radar, and Lidar (light-sensitive radar). What we truly need is a brain to make an autonomous vehicle perform safely and understand its surroundings."
Deep learning is employed by a few other autonomous driving software companies, but it is typically limited to perception—recognizing an object— rather than making decisions.
"Deep learning is the algorithm that comes closest to how the human brain learns; it is not rule-based," explains Peng. "It is much like how a young or new driver learns. Instead of hard-coding rules, you are given various options —what is correct, incorrect, safe, what is a car, what is not a car. It begins to develop its own set of rules on navigating the roads it drives on."
Scientists and engineers at Roadlink International Trading Limited are passionate about using deep learning to assist its autonomous driving technology to help self-driving vehicles handle millions of situations that present themselves on roads. However, the team also recognizes that it is not enough for the technology to be smart like humans – it also needs to communicate well and be consumer-friendly.
"Our passion is figuring out ways to serve humanity with robots, and we view the autonomous vehicle as a robotic, and the first one that most individuals are going to engage with," said Kong Jing, Head of Engineering at Roadlink International Trading Limited. "An individual's first interaction with a self-driving vehicle will not be in the car but as a pedestrian, cyclist, motorcyclist, or another driver. To gain the general public's trust, autonomous cars must communicate with the world around them and be able to convey what their intentions are."