Although Google's self-driving cars have safely driven more than 10,000 miles and tested a large number of different traffic environments, Ford has also launched a self-driving project and released its first self-driving prototype. But it is still too early to usher in large-scale mass production.
So, what exactly is the bottleneck of driverless technology?
Previously, there was widespread discussion in the industry that high-quality sensors were too expensive, resulting in high costs and inability to be mass-produced. But it seems that from the perspective of sensors, the biggest challenge is that all current cars rely on their own detectors to sense all the surrounding information, but they cannot truly communicate with other cars.
Some communication channels between cars have not yet been opened, and this technology is not yet mature. Frankly speaking, the important prerequisite for this is that the standards must be consistent between vehicles. But this is a huge challenge. We are constantly working with other manufacturers, including North America and Europe, to find ways to spread this standard.
Generally speaking, technology is still the primary bottleneck restricting the development of this industry. According to Professor Yao Danya from the School of Information Science and Technology of Tsinghua University, driverless cars can currently only operate according to predetermined procedures. Once they exceed the range, they will be unable to do anything.
Take Google's driverless technology as an example. It has a built-in driverless system developed by it, which can combine sensors, radars, cameras, laser rangefinders, etc. to achieve autonomous driving. However, Google’s self-driving cars are still inseparable from human control.
Similarly, although Ford cars currently have the function of detecting surrounding road conditions and automatically implementing emergency braking, and can also automatically park and accept voice-activated commands, they cannot automatically transport you to and from get off work.
In addition, driverless cars face many unknown factors. Ford's Alan Mulally once admitted in an interview that self-driving cars will be disturbed by heavy fog, rain and snow, and there is no way to know whether there is construction on the road ahead because this information does not exist in the navigator.
Currently, Ford’s test content is aimed at verifying the reliability of the self-driving system. Unlike the self-driving system developed by Google, Ford uses pre-programmed technology combined with cameras to monitor the environment around the car. , and then handed over to the control system for processing.
The unmanned driving system needs to judge concrete, gravel roads, speed bumps, rough gravel, mud pits and metal barriers on its own. But even if the vehicle has complete functions of autonomous driving, the safety of the vehicle can be greatly improved through sensors and real-time data. When encountering an emergency, manual processing is still required.
Compared with other companies testing driverless technology, I believe that Ford's great advantage is that Ford already has many cars, many radars, and various detectors. For example, he said that on the new Focus, there are front, back, and rear, which can help the car do some basic automatic parking. It will automatically help you brake in traffic jams. There are more than a dozen different technologies in all aspects.
Although it is not a direct leap to driverless driving, these intelligent driving technologies are slowly developing in the direction of driverless driving. Laughing. In his view, autonomous driving still needs a long process. By 2020, we hope to achieve driverless driving, as well as cars that can communicate with each other, automatically group together, and share information between cars and surrounding facilities.
News