Autonomous driving has become the main function of the models sold in the market, and with the passage of time and the improvement of market competition, autonomous driving has also begun to exert force in multiple use scenarios. On the track of autonomous driving, L4 technology is moving from the laboratory to the fission period of the real world. When Tesla's pure vision solution caused controversy and Waymo's Robotaxi fell into a large-scale dilemma, the Chinese startup Whale Intelligence opened up another front with a unique path - a full-stack self-developed technology based on lidar fusion, and at the same time tore a market gap in the field of passenger cars and unmanned delivery.
Behind this double breakthrough of technology and commercialization, there is a hidden effort of Whale Smart Globalization. From the R&D center in Shenzhen to the unmanned retail vehicles in Silicon Valley, a blueprint for "China's technology output" is unfolding. Through a written interview with Chang Yufei, founder and CEO of Whale Intelligence, we learned about the company's journey and its vision for the future.
Chang Yufei: The end game of autonomous driving is not a breakthrough in a single scenario, but a dynamic adaptation of technology and demand. With the core strategy of "passenger car-level technology driving multi-scenario landing", Whale Intelligent has built a dual-line product system covering people and goods. Our logic is to modularize the L4 algorithm verified by passenger cars and transplant it to vertical scenarios such as unmanned delivery vehicles, which not only ensures technical safety and redundancy, but also reduces commercialization costs. Originally used for positioning passenger cars in complex road conditions, it has now been adapted to unmanned delivery vehicles, and can still achieve centimeter-level accuracy in areas where satellite signals are missing. This technology reuse strategy allows Whale Intelligence to form differentiated competitiveness in R&D efficiency and scenario scalability.
Chang Yufei: The scale of unmanned distribution needs to break through three major thresholds: cost, reliability, and scenario adaptability. At the technical level, we have achieved breakthroughs through three innovations: first, high-precision map and positioning technology, using centimeter-level high-precision acquisition technology to build an environmental perception system, and cooperating with Fixposition to introduce a visually enhanced RTK 4 positioning solution to solve the positioning problems of unmanned vehicles in complex scenes such as tunnels and rainy days, and achieve centimeter-level accuracy in all scenes. The second is the multi-sensor fusion algorithm: based on the core technologies such as pre-fusion hardware design and passenger car-level decision control, the L0-level unmanned distribution system has been developed to support traffic lights, obstacle avoidance, path planning and other functions, which can adapt to the urban open road environment, and in terms of modular unmanned car design, we have also made great efforts to launch multi-functional autonomous driving unmanned vehicles (such as Pea II.), which support unmanned distribution, retail, connection and other scenarios, and adopt modular design for rapid customization.
In terms of commercialization, we are progressing and are landing unmanned retail vehicles in the U.S. market, covering fresh food, coffee and other scenarios. Unmanned delivery vehicles have been successively promoted and applied to California, Texas and other places.
Chang Yufei: The L4 landing of passenger cars follows the principle of "scene hierarchical penetration". We divide it into three stages: the first stage is the priority of closed scenarios: Whale Smart focuses on vertical scenarios such as campus logistics and unmanned distribution, which can achieve all-weather operations. Its multi-functional autonomous driving vehicle has been deployed in multiple regions, covering multiple application scenarios, and the frequency of manual takeover is low. Semi-open scenario expansion: Whale Intelligence has created the L0 passenger car autonomous driving suite, which is an advanced solution with lidar as the core perception device, supplemented by customized cameras, multi-sensor fusion technology, and equipped with high-precision maps and passenger car-level decision-making and control systems. Our achievements are obvious to all: a number of autonomous vehicle solutions have been successfully delivered, bringing customers a rich autonomous driving experience and forming a closed loop of "technology-product-ecology".
常宇飛：從需求側看，無人配送的爆發力更強。歐美市場人力成本高達中國的4-6倍，且城鄉道路標準化程度高，預計2027年全球室外無人配送車市場規模將突破500億美元，年複合增長率達67%。而乘用車領域，特定場景的L4車型將在物流園區、機場等場景率先形成百億級市場。
In the longer term, the technological synergy between the two will give rise to new business formats. For example, the real-time traffic data of unmanned delivery vehicles can feed back the high-precision map update of passenger cars, and the complex scene algorithm of passenger cars can improve the long-tail problem processing ability of delivery vehicles, which is the core barrier of Whale Intelligence.
Chang Yufei: The popularization of technology requires the deep integration of production, education and research. The L4 autonomous driving development and teaching kit for universities and scientific research institutions supports closed-loop testing and human-machine co-driving functions of autonomous driving systems, making it convenient for researchers to quickly get started and verify autonomous driving technology. It integrates multi-sensor fusion hardware (such as lidar, camera, millimeter-wave radar, etc.) to support high-precision mapping and positioning, and has powerful perception and decision-making capabilities.
Through in-depth cooperation with leading autonomous driving companies such as Baidu, Whale's intelligent human-machine co-driving development platform (DTV) has been widely applied in colleges and universities across the country, and has become an important practice carrier in the field of intelligent networked education. The platform not only provides a full-scenario closed-loop solution for autonomous driving, but also supports modular secondary development, providing a highly flexible technical base for university research and teaching. Through the integration of curriculum resources, software and hardware platforms and practical environments, Whale Intelligence has built a three-dimensional training system covering theoretical teaching and engineering practice, helping colleges and universities to build a talent training model that integrates the industry.
At the level of talent accumulation in the industry, intelligent networked education shows three core values: first, through the intelligent networked industry-education integration base jointly built by schools and enterprises, the whole chain of "teaching-training-R&D" is realized, and secondly, the open technology platform accelerates the cultivation of technical skills compound talents, and students can participate in the development of real industrial projects through the platform, which significantly improves the ability of engineering migration and innovation. The deep coupling of the education chain and the industrial chain can effectively shorten the talent supply cycle, achieve a good counterpart employment rate, and reserve strategic talents with cross-border integration capabilities for enterprises. This education model is promoting the intelligent network industry to form a virtuous circle of "talents leading technology and technology feeding education", providing continuous momentum for the high-quality development of the industry.
Chang Yufei: Internationalization is not a simple product output, but a trinity of "technical standards + scenario understanding + ecological collaboration" competition:
In the process of advancing technology, we will first find a way to localize. For the Japanese market, we are actively engaged in local cooperation. By combining the full-stack development capabilities of autonomous driving software and hardware systems with Kudan's strengths in the field of map and navigation technology, the product has been successfully implemented in Japan. At the same time, we have also established a close cooperative relationship with Autoware, the world's largest open source platform for autonomous driving, and Tier IV, a leading Japanese autonomous driving company. We are well aware that the localization and integration of technology will better meet the needs of the local market and help the products achieve a higher level of application and promotion.
Finally, we also need to put compliance first, and set up a dedicated team to study the new regulations in markets such as the European Union and the United States to ensure that products comply with the safety certification system of the target market. At present, a large number of commercial orders have been received at home and abroad. This year, the amount of new orders signed has doubled significantly compared to the same period last year, and the global business is also advancing rapidly in North America, Europe, Latin America and other regions.
AI is not a replacement for autonomous driving, but an amplifier. On the other hand, AI is not only an engine for technological breakthroughs, but also a lever for cost optimization. We will make good use of artificial intelligence, and will also apply new technologies to scenarios such as smart cities and smart logistics to improve business efficiency and user experience.
In Chang Yufei's view, the competition for autonomous driving has shifted from "single-point breakthrough in technology" to "ecosystem construction". Through the two-way empowerment of passenger cars and unmanned delivery, the three-dimensional layout of education and going overseas, Whale Intelligence is trying to drop a key chess piece with both technical depth and commercial breadth in the global chess game of L4 autonomous driving. "While others are still arguing about the pros and cons of lidar and pure vision, we have opened up the European, American, Japanese and South Korean markets with the 'China solution' - this is the global narrative that hard technology companies should have."
Chang Yufei also said that he will not stop keeping up with the pace of AI, and will continue to optimize the ecology in the hardware chain, software model, and talent echelon, based on the local market, and facing global users, which is a road that needs to continue to climb, but Whale Intelligence has the confidence to go on.