Computing power of v2x
On January 7, Qualcomm officially released its new autopilot computing platform, snapdragonride. The platform includes security system level chip, security accelerator and autopilot software stack, which can meet the computing requirements of the highest L4 / L5 autopilot. Qualcomm announced that the platform will be delivered to automobile manufacturers and first tier suppliers for preliminary development in the first half of 2020, and the vehicles carrying the platform are expected to be put into mass proction in 2023
interpretation: as early as 2012, Qualcomm entered the automotive chip field. Up to now, the cumulative amount of orders from Qualcomm in the automotive sector has reached $7 billion, up from $6.5 billion two months ago
however, the previous chips were all communication and entertainment chips. At the press conference, Qualcomm disclosed that it has been more than four years since it launched the autopilot chip. In 2017, Qualcomm obtained the license of automatic driving test in California. The purpose of this test is to verify its chip
currently, one of the biggest pain points of autopilot chips is that it is difficult to achieve the balance between computing power and power consumption. From the beginning, Qualcomm decided to make use of its accumulation in the mobile phone chip instry to make high computing power and low power ASIC chips. Snapdragonride, the L4 level automatic driving computing platform released by Qualcomm, is equipped with two self-developed automatic driving processors ml (ASIC), GPU and CPU
snapdragonride platform can achieve the highest computing power of 700tops with 130W power consumption. If this plan can be realized, then Qualcomm's automatic driving scheme is quite competitive in terms of computing power. For example, compared with the 144tops / 72W of Tesla hardware3.0, the energy efficiency advantage of Qualcomm's procts is obvious
among the mass proction schemes, Huawei's mdc600 has the strongest computing power, with 352tops computing power and 352w power consumption; Next is NVIDIA's drivepegasus, with 320tops computing power and 500W power consumption
according to the plan, the computing power of agxorin, the computing platform of Orin based on the next autopilot chip, can reach 2000tops, but the power consumption can also reach 750W, which can hardly be applied to mass proction vehicles
in contrast, Qualcomm's autopilot computing platform does not need a fan or liquid cooling system, but uses a simpler passive air cooling system to achieve higher reliability
snapdragonride platform also reces the number of chips and turns them into L2-L3 automatic driving solutions with computing power of 60-120tops or ADAS solutions with computing power of 30tops. At this CES, Qualcomm and GM reached a cooperation on ADAS business
in terms of business model, Qualcomm's approach is relatively open. It can provide full stack solutions, or only sell chips, algorithms and sensor solutions, which can be handled by customers themselves
in the future, v2x chip will also be one of the important revenue sources of Qualcomm in the automotive business
in addition, through its car to cloudservice, Qualcomm also provides technical support for car companies in cloud computing and Ota. It is worth mentioning that its car to cloudsoftsku supports the security upgrade of chipsets
this article comes from the author of car home, which does not represent the standpoint of car home
network communication technology
to realize
intelligent connected vehicle
v2x
function.
V2V represents a car to car
V2P represents a car to person
V2I represents a car to Internet
if you have any help, please accept it, thank you
2. Wireless communication technology of Internet of things
Buick figured it out
if you don't talk about autopilot in a week, you will feel out of date. However, in addition to lidar, camera and radar, neural network and algorithm are a subject to be created, and there is no ready-made formula to follow. So if you really want to find an autopilot path that can make you feel at ease, the v2x that has been around for some time is really reliable, Even though there is a bit of pragmatism
but it can be predicted that it is in order, and the sense of order is a strong hint of the future. Buick has also released some in car intelligent technologies, such as richer app, upgrading in details, speech recognition and feedback speed, which were not the strengths of joint venture vehicles before, and app changes with scene and time, maybe playing a rock music for you when you are sleepy
this time, the v2x blueprint of future automatic driving built by him is the key. He leads to some unknowns, which may also be a multiple-choice question. In the controllable variable but with a sense of order, and the uncontrollable but increasingly independent traffic situation, how do you choose
this article comes from the author of car home, which does not represent the standpoint of car home
