This project belongs to the field of automobile and vehicle engineering. As the main transportation of the city, buses and subways have large traffic volumes and high efficiency with large passenger flows, but they are poor in flexibility and have high operating costs. McKinsey has listed ldquo; autonomous driving rdquo; as one of 12 disruptive technologies that will affect the future. In addition to reducing costs, the application of driverless driving technology is also an effective means of achieving high-quality on-demand transportation services in low passenger flow scenarios and solving ldquo; the last mile rdquo; transportation problems. Urban low-speed unmanned systems have received widespread attention as the first choice for commercial operations. Europe has been developing and piloting urban low-speed unmanned systems since the end of the last century. France, the United States and other countries have successively launched driverless minibus services. In recent years, driverless minibuses have also become a research hotspot in our country. High-reliability, high-precision, and low-cost positioning technology is the key to the promotion, application and industrialization of urban low-speed unmanned systems. The project team has been researching driverless technology since 2004. It has been supported by key projects of the National Natural Science Foundation of China, the support plan of the Ministry of Science and Technology, and the European Union's international cooperation. The research work lasted for 12 years and achieved breakthroughs in high-precision positioning of low-speed unmanned systems in cities. A series of key technologies have been made. Vehicle-road collaborative positioning technology-invented a magnetic ruler sensor based on a magnetoresistive sensor array, proposed a magnetic nail curvature map model based on simultaneous positioning and mapping (SLAM), and adopted a magnetic nail lateral positioning technology based on pseudo-Gaussian curves to solve the problem of positioning vulnerability to severe weather such as rain, snow and fog. A magnetic nail positioning system based on vehicle-road collaborative was developed with a positioning accuracy of 1 cm. The installation height of the sensor from the ground has been increased from 4 centimeters to 25 centimeters, and the map building time has been shortened by 10 times, ensuring the reliable operation of the unmanned system in complex road conditions around the clock. Visual SLAM positioning technology-Invented a global positioning technology based on ground texture maps, proposed a longitudinal positioning technology based on terrain maps, a horizontal positioning technology based on end-to-end learning, and a VSLAM+GPS hybrid mapping method based on tight coupling optimization, solving the problem of visual positioning being susceptible to external light interference and blocking dynamic targets. A semantic SLAM positioning system based on a panoramic camera has been developed, achieving high-robustness and low-cost centimeter-level positioning in complex dynamic environments. Lidar SLAM positioning technology-Invented the SLAM+GPS map fusion technology based on factor maps, proposed a semantic point cloud map generation method based on panoramic scene understanding, a semantic point cloud segmentation algorithm based on depth separable maps, and a semantic point cloud registration algorithm based on graph and set theory, solved the problem of poor robustness of GPS signal failure area location and lidar positioning to initial values, and developed a semantic SLAM positioning system based on lidar. It realizes centimeter-level positioning in urban-level complex dynamic environments. This project authorized 52 invention patents and published 101 papers (16 papers were published in IEEE Trans, the top international journal in this field). It has developed three generations of driverless minibuses, participated in more than 10 large-scale exhibitions such as the Shanghai World Expo and the Shanghai Industry Expo, and was reported by nearly 20 media including CCTV and Xinhua Agency. In 2007, it hosted the ldquo; First Asia-Europe Smart Vehicle Symposium rdquo;, and the first generation of driverless minibuses achieved the first public demonstration in China at Shanghai Oriental Green Boat. The second-generation driverless minibus achieved demonstration application in Shenzhen Vanke in 2015. The third-generation driverless minibus developed in 2016 has been actually put into operation in Shanghai, Nanjing, Hangzhou and other places, running more than 30,000 kilometers in total and carrying more than 12,000 passengers. The results have also been promoted and applied in enterprises such as SAIC Motor, Qingfei Intelligent, and Guangdong Vanke. In the past three years, the added output value has been 3.9138 million yuan and the added profit has been 1.0707 million yuan. It broke the technological monopoly of European and American countries, greatly reduced costs, and promoted the commercialization of my country's driverless systems.