Please click T omRobotFollow _ _
Robot
drive.ai
:D
Do you still remember drive.ai founded by Wu Enda ?
As one of his three entrepreneurial practices, after a period of preparation, drive.ai's unmanned vehicle was finally launched. The L4 -level orange car will be deployed and available to the public in Texas starting in July.
1
drive.ai
Early this morning, Andrew Ng held a conference in Frisco, Texas, and gave a keynote speech on autonomous driving and drive.ai's autonomous driving implementation and related technologies
[ppt] https://www.slideshare.net/AndrewNg44/selfdriving-cars-are-here
In his speech, in addition to the development process and technical implementation of autonomous driving, he explained the deployment plan of drive.ai.
At the same time, it also introduced the cooperation model with local parties and the people-oriented safety concept, and used various technologies to protect the safety of people inside and outside the vehicle.
The autonomous driving of drive.ai will be realized in three stages, from the safety driver in the driving seat to the co-pilot's guide , and then to remote control . Step by step to achieve autonomous driving and ensure safe operation.
Let's take a closer look at drive.ai's autonomous driving!
According to the official statement, drive.ai's autonomous driving mainly has three key elements:
Industry-leading AI and deep learning technologies ;
The deployment of autonomous driving through collaboration with public and private partners ;
People-oriented security policy;
In terms of technology , drive.ai said that it has reserved a high-level AI team and an experienced software and hardware architecture team. The team not only includes Wu Enda's graduate students at Stanford but also Wu Enda and his wife, who are technical experts themselves. They use modern AI technology to create a new autonomous driving architecture. At the same time, new algorithms have been developed for indoor perception, motion planning, map components, positioning, etc. The team has also carried out rapid development and iteration in fleet management, mobile app communication and remote assistance systems.
In terms of operation , seamless and smooth operation is achieved through the same cooperation with local government and private enterprises, and the added value of use is improved. As the technical team drive.ai is very aware of the limitations of AI, the team knows how to implement a practical system through the existing system. Through cooperation with the government, they operate autonomous driving operations in a specific area, avoiding the impact of construction and other situations.
The initial deployment will be around HallPark and gradually expand to surrounding areas. Such on-demand deployment can be convenient for everyone around. Commuters can drive around looking for a parking space without having to grab their breakfast in one hand, and it will bring more traffic to stores near the operating site. At the same time, in addition to promoting local economic development, it can also ease traffic and parking pressure, reduce personal driving and reduce people's travel costs.
In terms of safety, drive.ai proposes a human-centered safety concept that takes into account the safety of both occupants inside and outside traffic participants, and emphasizes the importance of communication and public education. The behavior of self-driving cars is not only related to the car itself, but also closely related to the surrounding traffic participants. It is unrealistic to place all safety on the car, and it requires the participation of all traffic participants to achieve a safe traffic environment, which is also the reason for the need for progressive public education. drive.ai works on three fronts:
Eye- catching self-driving appearance allows everyone to recognize the self-driving car;
LED screens for interacting with other traffic participantsand signals;
At the same time, we will cooperate with local governments and schools to increase the public's awareness and understanding of autonomous vehicles, learn to interact with autonomous vehicles in a special way, and try our best to ensure safety.
2
route map
drive.ai laid out a three-step roadmap, with a safety driver in charge of the first phase , but also provided an advanced remote control interface. When a situation arises that requires human intervention, the car will first stop and seek the assistance of a remote operator. Deep learning systems will continue to improve autonomous driving performance over time. At the same time, different from remote driving, the remote control system of drive.ai has been carefully designed considering the network load and delay, and the delay is controlled within 100ms.
In the second stage , the co-pilot safety guide will be responsible for the safe operation of the vehicle. In addition to monitoring the situation of the vehicle, he will also assist pedestrians and other vehicles to communicate with the autonomous vehicle.
Finally , it will enter the remote monitoring stage, thinking that security officers will remotely monitor and assist several autonomous vehicles to achieve large-scale deployment.
3
The future of autonomous driving
Autonomous driving has advantages and disadvantages compared to human drivers. They are always attentive and responsive (<100ms), no blind spots and no fatigue. But they are incapable of understanding the world around them in complex situations. By using geofencing and working together, drive.ai says it can maximize the advantages and avoid the disadvantages of autonomous driving, making it an efficient, safe and valuable mode of transportation. From the DARPA City Car Challenge in 2007 to today, ten years have passed and technology has advanced rapidly. Many dreams of the past are gradually becoming a reality.
Finally, learn from role model Ng!
Interested in learning can continue his deeplearning.ai :
There is also landing.ai where the AI industry has landed :
ref:
https://twitter.com/AndrewYNg
https://medium.com/@andrewng/self-driving-cars-are-here-aea1752b1ad0
https://www.slideshare.net/AndrewNg44/selfdriving-cars-are-here?from_action=save
Image source (from the Internet, if there is any infringement , please contact to delete )
Want to see more robots ～～～ please click---->>>:
The Wind Rises - Indoor Commercial Service Robot
AutoStore Ohmni walks a tightrope
Open Source Robot Robotic Architect
TALOS quadruped machine ANYmal
Pop.Up takes you to fly the MIT cube
IHMC's Open Day Bone Shrinking Robot
TomRobot>: $
Lover Maker Player Creator ~ ~ ~
For more exciting robots , click " Read the original text " below to view
~more than you want~