Xilinx Enables ADAS, IMS for Smart Car Mass Production Applications

Self-Tech continues to mass-produce highly reliable and cost-effective intelligent driving solutions based on FPGA platform, such as Advanced Driver Assistance System (ADAS), Intelligent Cockpit System (IMS) and other in-vehicle and out-of-vehicle vision-aware intelligent driving solutions.

Ltd. is a leading global provider of intelligent driving solutions and devices, and a pioneer in developing computer vision and sensor fusion technology based on FPGA hardware platform in China, with deep talent and technology reserve and perfect product development tool chain in the field of FPGA-enabled image processing technology and deep learning technology. The company is also one of the drafters of the Chinese national standard “Driver Attention Monitoring System Performance Requirements and Test Methods” (known as DMS), a software partner of Bosch in China, and a certified partner of Toyota Tsusho and Xilinx.

Core technologies of our own

Project Overview

The development of artificial intelligence, 5G and connected car technologies is redefining the safety and human-machine interaction experience of future cars, and the corresponding iterations and deployments of various smart driving technologies are increasingly accelerated.

“ADAS and IMS, as the first two important areas of intelligent driving, are driving in two major directions: “driving automation” and “cockpit intelligence”. ADAS and IMS, as the first two important areas of intelligent driving, are promoting the process of intelligent driving from two directions: “driving automation” and “cockpit intelligence”.

Self-Tech is one of the earliest FPGA companies in China, and one of the earliest certified partners of Xilinx in China for the automotive market. With Xilinx automotive-grade Zynq-7000 SoC series adaptive computing platform, Self-Tech has continued to mass-produce a series of ADAS and IMS in-vehicle and out-of-vehicle vision-aware intelligent driving solutions. The Dynamic Function Translation (DFX) feature of the Xilinx Zynq platform provides Self-Tech with higher performance, lower latency, and the flexibility to load and dynamically switch multiple ADAS functions on a unified computing resource in real time.

Industry Challenges

Around View Monitor (AVM, Around View Monitor) is one of the most penetrated ADAS features in recent years, also known as 360-degree panoramic surround view system. The system uses four camera sensors installed in the front and rear of the vehicle to sense the environmental data outside the vehicle, helping the driver understand the vehicle’s surroundings and providing timely alerts before danger occurs. Compared with a simple front-view or rear-view system, AVM requires real-time processing of 360° data from the body and helps drivers make quick decisions through image stitching, video processing and other technologies.

Traditional AVM systems are usually implemented based on ARM or GPU, for example, NXP’s I.MX6 series chips. These chip platforms applied to AVM systems will face various limitations, such as low efficiency of ARM, high power consumption of GPU, slow start-up time and high cost characteristics, making it both from energy efficiency and cost, can not meet the car manufacturers intelligent driving super large-scale deployment needs.

At the same time, in the field of intelligent cockpit, with the end consumer demand for diversified interaction experience of intelligent cockpit increasingly developed, in the traditional voice, touch interaction outside, based on the intelligent interaction of visual sensors, multi-sensor fusion intelligent interaction is also increasingly loved by consumers. In the face of multiple and continuously changing experience needs, intelligence needs, sensor fusion needs, coupled with the unchanging cost requirements, in the limited cabin environment, the existing ASIC, ASSP, DSP and other fixed or standard processor platforms are difficult to meet all of the above needs in terms of size, performance, cost and other aspects at the same time.

With rich experience in FPGA design, Self-Tech is the first to realize that FPGAs are flexible in deployment, fast in startup, stable, rich in logic resources, and adaptive in provisioning, which can meet the various functions required by AVM and IMS systems. They hope to combine their own world-leading AI technology and outstanding FPGA development capability to create a new generation of AVM and IMS systems for the industry.

Solution

Self-Tech selected the Xilinx automotive-grade (XA) Zynq-7000 SoC family of adaptive computing platforms to build the next-generation AVM and IMS systems.

As a heterogeneous multi-processor hardware processor platform based on FPGAs, the Zynq SoC combines programmable logic (PL) and on-chip processing system (PS) into a single platform that not only has the features of typical processors such as ASICs, ASSPs and DSPs, but also has more features than other processors,” said Dr. Yanfeng Guan, CEO of Self-Tech. . These features, combined with our leading AI deep learning technology and computer vision and sensor fusion technologies, bring valuable differentiation and leadership advantages to build a new generation of AVM and IMS systems.”

In response to customer needs, Self-Tech has developed a highly efficient AVM system based on the mature automotive-grade Zynq-7Z010 platform from Xilinx, which achieves fast start-up, low stitching distortion and high robustness, as well as dynamic viewpoint following and dynamic loading of front and rear auxiliary lines, injecting a fresh impetus to the traditional red sea market of AVM system. Meanwhile, with its leading AI algorithm capability, Self-Tech has developed a cost-effective and rich IMS based on the Zynq-7Z010 and 7Z020 platforms, which covers a wide range of functions, including fatigue detection, attention detection, identity recognition, status recognition, behavior recognition, gesture recognition, emotion recognition, occupant status, rear recognition, etc.

Self-Tech’s new generation AVM and IMS system achieves, compared to the industry’s existing solutions.

Higher performance. A single Zynq chip enables a complete ADAS solution design from environment characterization to sensor fusion and feature extraction, making our solutions avoid the bandwidth bottlenecks and security concerns of all other cross-chip platform solutions, resulting in higher performance and more stable and reliable solutions.

Lower latency. Real-time responsiveness with low latency is critical to driving safety. The unique FPGA parallel processing advantage of the Xilinx SoC allows flexible access to all types of sensors and parallel processing of data from the entire front-end sensor at the nanosecond or microsecond level, resulting in lower system latency relative to either the GPU or CPU.

Lower System Total Cost of Ownership (TCO): By leveraging the rich hardware resources such as BRAM, lookup tables, registers, DSP, etc. in Zynq, we can successfully deploy computationally complex deep learning networks on low-cost Zynq chips to develop cost-effective, versatile and competitive IMS and ADAS systems.

More flexibility. With the reconfigurable feature of FPGA and the deep learning modular design and reconfiguration technology of Self-Tech, the new Self-Tech solution can support users to customize new functions or easily expand sensors according to their needs. At the same time, it also lays a solid foundation for rapid and optimized deployment of new deep neural networks and subsequent expansion of various ADAS system functions.

Smarter and safer: The DFX function of Zynq SoC enables the new solution to load and dynamically transform multiple ADAS functions on a unified computing resource in real time. For example, other driver assistance functions, such as lane departure warning (LDW) and blind zone monitoring (BSD), can be implemented on the new AVM system to make driving smarter and safer.

Self-technology car intelligent product layout

Design Effectiveness

COTECH continues to develop cost-effective ADAS, IMS and other intelligent driving solutions based on FPGA hardware platform. Compared with the traditional I.MX6D solution, our new AVM system reduces the cost by more than 50% compared with the traditional chip (I.MX6D solution) only in the main chip, and achieves the current level of mainstream intelligent automotive electronics solutions in terms of function and performance, which can meet the needs of domestic and international intelligent automotive electronics market.

At present, all kinds of intelligent driving electronic products and solutions developed on FPGA platform have been successfully mass produced in domestic commercial vehicles and passenger cars (China’s largest bus brands, as well as well-known joint venture car brands with annual sales of 2 million, etc.) front-mounted market, and the accumulated shipments have reached tens of thousands of units.

Bus DMS product diagram

Auto AVM product diagram

Redmi K50 Gaming Edition (12GB/128GB/5G Edition) Snapdragon 8 Gen1 Processor, Dual VC Powerful Cooling, 120W Speedy Flash Charging, Magnetic Power Pop-up Shoulder Key

[Distributor] Store

[Product Price] RMB 3599

Redmi G 2021 Dragon Edition (R7 5800H/16GB/512GB/RTX3060) Dragon 5000 Series R7 Processor, 130w RTX3060 6G Solo Display, 144Hz Gaming Screen, Backlit Keyboard

[Distributor] Store

[Product Price] 7299RMB

Redmi Gaming Router AX5400

[Dealer] Store

[Product Price] 599 (Pre-sale)