My autobiography is definitely not a tragedy! -
Chapter 702 - 437: Rise of the Empire_2
Chapter 702: Chapter 437: Rise of the Empire_2
This configuration... is not just low, it’s way too low...
As a car manufacturer, he was extremely familiar with the structure of new energy vehicles. Without delving as deep as dissection, he could basically tell most of it at a glance.
But... not to mention the lidar, he hadn’t even seen any sign of millimeter-wave radar installation.
So how exactly did they achieve this automatic parking?
It couldn’t possibly be just visual, could it?
He couldn’t help but glance at Mr. Bao, his eyes a bit peculiar.
He didn’t speak right then and simply settled into the driver’s seat.
"Bajie, could you please tell me the processor specifications of this car?"
Unexpectedly, his first question didn’t test the voice interaction intelligence but was about probing the car’s specifications.
A lot of people in the audience didn’t understand why, but his two companions nodded repeatedly.
Xiao Peng, you just hit the nail on the head!
[Hello, this car primarily uses chips from Intel Company for crucial computational support. This includes Intel Atom processors, Intel Movidius vision processors, and Intel ALTAI wireless communication modules.
In addition, it is also equipped with MCU control chips by Renesas Electronics and communication modules by Ericsson.]
That’s it?!
Xiao Peng glanced at Mr. Bao, his eyes growing even more bizarre.
With this crap configuration, you want to do assisted driving? What a joke!
If this works, that would not only require "pure vision" but also damn "low computing power"!
The Mao Dou series needed an addition of millimeter-wave radar and advanced chips...
This guy... he couldn’t actually be making false propaganda at the launch event, could he?
All these journalists around, hello!
Do you really think the government doesn’t crack down on entrepreneurs?
"Heh! I didn’t know that such low computing power could also accomplish assisted driving features. Could you explain how you achieved this?"
[In autonomous driving, ’automatic parking’ is considered a low complexity, simple scenario. Moreover, as you mentioned, our computational configuration is not advanced in the market; therefore, it can only complete such scenarios at low speed.
This, to some extent, compensates for the hardware insufficiency.]
"Uh..."
Low speed compensating for low computing power, logically seems alright, but it doesn’t hit the mark.
"So, this car has achieved the ’pure vision’ parking technology?"
Xiao Peng emphasized the words "pure vision".
Upon hearing this, the audience couldn’t miss Xiao Peng’s intention. It was not merely about experiencing a "talking car"; he was outright probing the secrets of their "automatic parking"!
[Yes, your understanding is correct!]
"Wow—" An immediate gasp rose from the audience.
Unexpectedly, this "intelligent voice assistant" had just admitted it on the spot!
"Heh!"
Considering that it was indeed their official launch event, Xiao Peng refrained from scoffing aloud right then and instead turned to look at Mr. Bao beside him.
"Mr. Bao, do you also agree with the view of your... ’intelligent assistant’?"
"Err..."
Mr. Bao touched his nose.
Initially, I was just like you, skeptical, thinking the other side was daydreaming.
However, when their technology literally smacks you in the face, denial isn’t an option...
"Mr. He, as I just mentioned, this ’Tian Xuan’ system isn’t our product, it’s provided through technical support by a strategic partner.
So, regarding technical details, I definitely don’t know as much as an expert like you.
Maybe... you should continue asking?" he suggested, pointing at the car.
Xiao Peng twitched the corner of his mouth.
Keep asking my ass! The boss can’t even spin more yarn and has to let the pros handle it now, right?
Fine! I want to see what kind of story you can spin now!
"This... Bajie, could you explain how you solve the detection and tracking of obstacles in complex parking environments with low computing power?"
[Using deep learning, incorporating classic SSD and YOLO algorithms, but with significant optimizations.]
"Eh?" Xiao Peng didn’t expect such a professional response.
"So for SSD, what specific optimizations did you implement?" he asked, not quite believing.
[The main optimization was in the convolutional parts of VGG16, allowing it to use pre-trained VGG16 network weights and to perform transfer learning on other datasets, thus improving the model’s versatility. After testing, it shows high detection speed and impressive detection accuracy in single-shot object detection tasks.
The specific optimization process is:
$ L_{Cls}=\frac{1}{N}\sum^{N}_{i}L_{Cls,i}$...]
Xiao Peng: "..."
Everyone: "..."
As the car speakers belted out various complicated mathematical symbols and formulas that sounded like the Heavenly Book, everyone was stunned.
Hey, hey! When asked about what optimizations were made, a brief introduction would have sufficed; there’s no need to recite all the technical details and optimization formulas on the spot, right?
Doesn’t your company have any technical confidentiality regulations?
Mengzi didn’t understand it, Confucius didn’t understand it, Laozi also didn’t understand it.
Quite a few journalists were inwardly cursing.
I just came to write a news piece about the "last dirge of a traditional new energy car enterprise", not expecting to attend an "advanced autonomous driving technology release event"?
No mental preparation at all, huh!
Meanwhile, Xiao Peng was thinking...
Should I just take some notes?
No way around it, the algorithm formulas sound increasingly convincing...
"Stop, stop, stop!" he quickly shouted.
The somewhat challenging torrent of symbols and formulas finally halted.
If you find any errors (non-standard content, ads redirect, broken links, etc..), Please let us know so we can fix it as soon as possible.
Report