Ng’s Driverless car, Drive.ai, will offer free rides in July.

It’s not just a milestone for Texas, it marks the first time the public will be able to ride in self-driving cars on public roads, and it’s a major step forward for the autonomous driving industry.

The project will be launched in July 2018 for a six-month pilot period and will be limited to The retail, entertainment and office areas of Frisco, with drop-off and pick-up locations fixed at HALL Park and The Star. There are plans to expand to Frisco station.

The drive. ai service is being developed in partnership with local Governments, including Frisco TMA TMA and the North Central Texas Council of Governments (NCTCOG).

Frisco TMA is known to be a public-private partnership. Dedicated to providing last-mile transportation innovations to the growing population of Frisco, Texas. Frisco TMA Partners, including The City of Frisco, Hall Group, Frisco Station Partners, The Star, Denton County Transportation Authority and others, managed The project.

At the same time, Ng, who is a director of Drive.ai, also published an open letter explaining the details of the service, its plans and the development of the autonomous driving industry.

The letter:

Dear friends,

Drive.ai will begin offering self-driving cars for public use in Frisco, Texas, in July 2018.

Self-driving cars are no longer the AI of the future. They are here, and will soon make transportation cheaper and more convenient.

The drive. ai team has been working closely with local partners to ensure that our vehicle deployment takes place safely and delivers real value to users.

Drive.ai’s ability to provide self-driving services to the public depends on three factors:

Self-driving technology remains challenging. It requires a very experienced AI team and a complex software and hardware architecture.

Drive.ai has a strong technical team. The founders include my group at Stanford and many of Carol Reiley’s AI graduate students. The team, made up of deep learning professionals, designed an autonomous driving architecture using modern artificial intelligence from the ground up.

In addition, by developing a complete software stack for autonomous vehicle internal awareness, motion planning, mapping, localization, fleet management software, mobile applications, communications, and our Remote Select remote assistance system, the team was able to quickly resolve any dependencies between the systems.

Autonomous vehicles work with governments and the private sector to operate in designated areas to ensure safety, stability and bring real value to users.

As an experienced AI team, Drive. AI has a clear understanding of the limitations of AI. The team knows how to build realistic solutions within the limitations of current technology.

We know, for example, that none of the driverless car teams have been able to develop a reliable route map for moving vehicles based on the waving hands of construction workers; Computer vision is not perfect. So the collaboration that we’re working with the government and the private sector is limited to areas where we can find other ways for construction workers to communicate with our fleet operations teams.

Special thanks to Mayor Jeff Cheney of Frisco, TMA of Frisco and Michael Morris of NCTCOG (North Central Texas Council of Governments) for working with us. We will initially run a six-month trial on The HALL Park to Entertainment/Retail (The Star) route, with plans to expand to Frisco Station.

Local deployment of autonomous driving services will bring convenience to individuals and businesses, reduce traffic congestion and reduce the need for parking lots. We are also working to address the inaccessibility of remote areas and improve connectivity with existing bus routes. Well thought out deployment of autonomous vehicles can ease traffic pressure and reduce private car use, thus reducing urban transport costs.

The autonomous driving industry must put human safety at the root, ensure the safety of both inside and outside the vehicle, and focus on communication and community education.

The safety of a vehicle depends not only on the vehicle itself, but also on the behavior of people around it. It is unwise to rely solely on AI for security. Instead, operators must consider outsiders, which is why we focus on community education and training.

To ensure safety is the responsibility of every operator, we believe the autonomous driving industry should take the following measures:

• Autonomous vehicles should be visually striking, and surveys show that orange is the color most recognizable to pedestrians and drivers.

• Driverless vehicles cannot communicate with pedestrians and drivers, so we equip four external screens to communicate the vehicle’s “intention to drive” to pedestrians and other drivers on the road, and put information signs on the entire route of the autonomous vehicle.

• Autonomous vehicle companies should work with local governments to strengthen public education and training, and increase public awareness of autonomous vehicles through media, special signs, dedicated pickup and drop-off locations.

In the first phase, Drive.ai will deploy safety driver vehicles in Texas, where our “Remote Selection” technology also improves safety and ride comfort. Suppose, for example, that our vehicle is at an intersection or in a difficult situation. If it determines that it needs human control to stay safe, it will stop and ask a remote operator for input to determine what to do next. Over time, our deep learning system will learn from these cases and improve them automatically. Unlike “remote driving,” the “Remote Selection” mode, in which the operator takes direct control of the car, is designed to address issues such as network delays and temporary network outages, even minor edge situations such as old data automatically lapses or requests lag 100 milliseconds.

In the second phase, drive. ai will operate as a remotely selected operator “companion” (rather than a safety driver) after passing road safety tests. The companion will sit in the passenger seat, assist passengers and monitor operations, but will not take over instantaneously.

In the final phase, the vehicle is fully autonomous, with only passengers in the vehicle, assisted remotely by a remote selection operator. A remote selection operator will be able to monitor multiple vehicles, enabling scale operations.

We still have a lot of work to do, but the future of autonomous driving is there.

Self-driving cars have different advantages and disadvantages compared to human drivers. They are always attentive, have reaction times of less than 100 milliseconds, and have no blind spots. On the other hand, they don’t understand complex situations, such as construction workers using hand signals to communicate with them. By choosing where to operate and working with partners, we can leverage the strengths of autonomous vehicles and avoid their weaknesses. With the implementation of these strategies, the autonomous driving industry will provide people with safe and valuable transportation services.

I remember attending the DARPA Urban Challenge in 2007 and seeing the great work of Stanford university, CMU, and many other pioneering driverless car teams. Our work is built on these rich results.

Ten years later, I’m excited about how far autonomous cars have come.

To learn more about drive.ai autonomous driving, visit Drive.ai

(https://www.drive.ai/).

Wu En da

Drive.ai was founded in 2015 by graduate students working at Stanford University’s ARTIFICIAL Intelligence Lab, which is run by Andrew Ng, a renowned AI expert and current chairman of drive. ai’s board of directors. He was a former chief scientist at Chinese tech giant Baidu and helped Jeff build the Google Brain project.

Since its inception, Drive.ai has had a number of notable moves, including board changes, a partnership with Lyft, and the release of its first self-driving car video. But this is easily the company’s most significant move to date, demonstrating its real-world technological advances and prowess.

Drive.ai stands out among startups by applying deep neural networks to self-driving cars. Deep neural networks allow computers to identify patterns in data through a series of connected networks. Most companies developing autonomous driving technology use deep neural networks to deal with specific problems, but they use machine learning algorithms rather than deep neural networks in broader systems.

Drive.ai uses the deep web to perform driverless cars’ functions, including object recognition, motion planning and making decisions based on what they see.

Proponents of deep learning argue that these algorithms are ideal because they are similar to how humans learn. But other companies confirm that the deep Web is not yet ready and will require more research and computing power to be safe and reliable in autonomous vehicles.