Apple Car has another new development. Apple will reportedly introduce machine learning into its electric vehicles to enable cars to make timely autonomous driving decisions in critical situations.
Machine learning belongs to the field of AI technology, mainly through experience and data, the study of computer algorithms. The use of AI-related technologies in Apple Car actually has a premonition. It is understood that at the end of 2020, John Giannandrea, who is in charge of apple AI and intelligent assistant Siri research and development, was appointed as the head of the Apple Car project.
Giannandria, who worked at Google and was involved in Google Assistant, Gmail and search-related businesses, officially joined Apple in 2018 and reported to Apple CEO Tim Cook.
Recently, a newly announced patent reflects the reasons and methods for Apple to use machine learning technology. It all has to do with the fact that the car has to react very quickly during the autonomous driving process.
Figure | Apple applied for related patents (Source: U.S. Patent and Trademark Office)
The current automatic driving, due to the limitations of software and hardware, the analysis and calculation of the external environment of the vehicle is not enough for the car to make timely driving decisions on its own.
Even though the hardware and software of cars are becoming more and more powerful, with fast processors and advanced algorithms, this is still not enough. The car's systems must determine the current state of the surrounding environment at all times to determine "a set of corresponding feasible or proposed actions that can be taken." If the right decisions, such as lane change or avoidance, are not made quickly, they can lead to fatal accidents.
To some extent, machine learning can help cars figure out the optimal course of action from a multitude of possible decisions.
(Source: Pixabay)
It is understood that Apple's patent application describes more than 17,000 words, describing the time and distance of the decision made by the self-driving car.
Apple introduced that when driving on a largely empty straight road, the vehicle may not turn within a few kilometers, and the driving behavior to be evaluated by the car is relatively small; but when the vehicle is at a busy intersection, the number of behaviors that the car needs to evaluate will increase greatly.
Apple also spoke in its patent note about the "complexity of autonomous decision-making." It notes that real-world environments in reality tend to be much more complex than any test site, and that some of the behavior of other cars and pedestrians on the road is almost unpredictable. In the absence of data, it is difficult for autonomous driving systems to make correct and timely decisions.
Using a machine learning model on a vehicle, you get a "corresponding value metric for behavior" that can be used to select the action to be implemented.
Finally, let's talk about Apple Car production. Apple Car can be said to have been in the "rumor", Apple has not officially announced its car project. At first, it was speculated that Apple was only developing a self-driving system and would not build a car itself, but then more and more evidence showed that Apple was considering releasing a private label car.
The Apple Car project has been brewing for many years, here are some reports about Apple cars in recent years.
In 2014, Apple began working on electric vehicles, naming the program Project Titan, with more than 1,000 automotive experts and engineers working on electric vehicles at a secret location near Apple's headquarters.
In 2015, BMW and Tesla engineers joined Apple, and Apple is also recruiting talents related to self-driving car algorithms.
In 2016, Apple acquired a new domain name related to Apple Car and obtained a "Car Test Center" building permit. There are also reports that Apple has a secrecy automotive laboratory in Berlin, Germany.
In 2017, it was reported that Apple was collaborating with China's CATL times on automotive battery research; Apple also released a research paper on lidar (LiDAR) 3D object recognition.
In 2018, Apple submitted a self-driving car application to the California Vehicle Authority and hired dozens of employees from Tesla. Cook also said that Apple is working on self-driving systems, focusing on the software side,
In 2019, Cook said that future products will be "a surprise." Apple has built multiple large cabs in California to work on improving battery technology. At the same time, Apple has also acquired AI startup Drive.ai.
In 2020, a patent shows that Apple is developing Tesla-style induction motors. There are reports that Apple will work with TSMC to develop apple car processors.
Fast forward to 2021, and there has been some substantial progress in the news about Apple's car manufacturing. In January 2021, Apple negotiated with Hyundai to help make the Apple Car, but the negotiations were shelved. In February, analysts said the Apple Car might abandon the steering wheel. In March and April, there were separate reports that Foxconn might assemble Apple Cars, and LG and auto parts supplier Magna could win contracts with Apple Cars. News broke in May and July that the Apple Car's windshield could be used for AR displays, and Apple wanted to make car batteries in the United States. The November report said that apple cars are expected to be released by 2025.
Finally, it is understood that due to the company's leadership and other conflicts and other issues, Apple's car project has changed many times, and even once shelved the car project, but now Apple Car is on the right track.
Now there are rumors that Apple is developing a "fully autonomous self-driving car" that can drive without user intervention. That would go further than other automakers in terms of autonomous driving.
-End-
reference:
https://appleinsider.com/articles/22/02/08/apple-car-needs-machine-learning-to-make-driving-decisions-fast-enough
https://appleinsider.com/inside/Apple-Car