How to implement intelligent driving solutions on a large scale? Horizon Robotics

Recently, Horizon Robotics, a well-known intelligent driving computing solution provider, held a 2024 Intelligent Driving Technology Product Launch Conference. At the conference, "Intelligent Driving Equality" and "Software and Hardware Integration" were undoubtedly the most frequently used words. For example, Yu Kai, founder and CEO of Horizon Robotics, believes that the real large-scale mass production of high-end intelligent driving functions in the future requires the synergy of software and hardware to promote...

Although the development of my country's intelligent driving field has been positive in recent years, the large-scale implementation of intelligent driving solutions still faces many problems. For example, low- and medium-level intelligent driving solutions are limited by the original automotive electronic and electrical architecture, and cannot reduce costs and meet the diversified needs of car companies. At the same time, the technology of high-level intelligent driving solutions is still immature and cannot play the ultimate effect of experience functions. To solve the above problems, it requires a long time of polishing and integration of products and technologies. The Journey® 6 in-vehicle intelligent computing solution launched by Horizon Robotics obviously provides "answers" to these problems.

It is understood that Horizon's Journey® 6 in-vehicle intelligent computing solutions include six products, including Journey® 6B, Journey® 6L, Journey® 6E, Journey® 6M, Journey® 6H, and Journey® 6P, covering everything from all-in-one devices to domain control, from active safety to full-scenario intelligent driving. More importantly, Horizon's Journey® 6 series of in-vehicle intelligent computing solutions have a unified hardware architecture, a unified tool chain, and a unified software stack, with impressive performance in access capability, processing capability, and integration.

Among them, Horizon Robotics' Journey® 6B in-vehicle intelligent computing solution is aimed at the low-end intelligent driving market, focusing on active safety capabilities. Currently, based on the Journey® 6B in-vehicle intelligent computing solution, Horizon Robotics has jointly released the world's first 17 million high-performance forward-looking perception solution with Sony. The Journey® 6E in-vehicle intelligent computing solution and the Journey® 6M in-vehicle intelligent computing solution are aimed at the mid-end intelligent driving market, with computing power of 80TOPS and 128TOP respectively. Both are committed to filling the gaps in the high-speed NOA and popular urban NOA markets. Not only can they provide high-performance support for cutting-edge algorithms including BEV Transformer and OCC, but they can also greatly reduce the cost of the entire integrated system. At the same time, the efficiency and flexibility of development are greatly improved, the loading efficiency is accelerated, and the development cycle is shortened. At present, the Journey® 6E in-vehicle intelligent computing solution and the Journey® 6M in-vehicle intelligent computing solution have received orders from several mass production cooperation car companies and brands including SAIC Group, Volkswagen Group, and BYD.

In addition, Horizon Robotics has also provided a solution for urban NOA, namely the Journey® 6P in-vehicle intelligent computing solution + SuperDrive™ full-scenario intelligent driving solution. From the perspective of positioning, the Journey® 6P in-vehicle intelligent computing solution is designed for urban NOA. It adopts the new BPU Nash architecture. At the same time, the 18-core Arm Cortex-A78E CPU can provide 410K DMIPS computing power, and integrates a GPU with 200G FLOPS computing power, supporting 3D image rendering; the full-function MCU provides 10K DMIPS ASIL-D computing power, and a single Journey 6P can support full-stack computing tasks such as perception, planning decision-making, and control. The SuperDrive™ full-scenario intelligent driving software solution supports end-to-end perception architecture and interactive game algorithms for dynamic, static, and OCC occupied networks. According to Horizon Robotics, the end-to-end perception architecture that integrates dynamic, static, and occupancy networks is an effective means to accurately restore the objective physical world. Under this architecture, the occlusion recall rate can be increased by 70%, the number of dynamic code lines can be reduced by 90%, and the network load can be reduced by 50%. At the same time, the SuperDrive™ full-scenario intelligent driving software solution can also solve the current problems of imperfect and unstable urban NOA in a more anthropomorphic way, and can take into account the scene pass rate, traffic efficiency, and behavior anthropomorphism in any road environment. Based on the Journey® 6 in-vehicle intelligent computing solution + SuperDrive™ full-scenario intelligent driving solution. Horizon Robotics has created a high-end intelligent driving model room, striving to improve R&D and delivery efficiency through the flexible opening of full-stack technical capabilities, and accelerate the industry's move towards the end of high-end intelligent driving.

For Horizon Robotics, "software and hardware integration" is not a sudden inspiration, but a development strategy that the company is implementing step by step. Of course, the development of high-level intelligent driving cannot be achieved overnight, and Horizon Robotics still has a long way to go...