Bulletin No. 2, 2023

Pooling expertise The Centre, founded in 2020 under the Hong Kong government’s InnoHK initiative, focuses on the research and development of robotics and artificial intelligence technologies for the “future workplace” as well as innovative solutions to pressing problems in the logistics industry, a key pillar of Hong Kong’s economy, such as manpower shortages and soaring labour costs. The University of California, Berkeley contributes research. The Centre runs four research programmes: robust sensing and perception; human-robot collaboration; smart manipulation robots; and unmanned logistics vehicles. They are steered by 20 principal investigators: 15 engineering professors from CUHK and five from UC Berkeley. The head of the Centre, Professor Liu has long established himself in the field of logistics robots, his signature innovation being the world’s leading vision-based autonomous forklift. Hundreds of projects have been deployed around the world by VisionNav Robotics, a global manufacturer of autonomous industrial vehicles he set up with some of his students. Professor Liu has also helped developed China’s largest smart warehouse for Cainiao, a logistics company owned by e-commerce giant Alibaba, in the city of Wuxi. It deploys automated guided vehicles developed by Professor Liu’s team, including smart forklifts and sorting robots. Smart building inspections Apart from the waste sorting start-up, the Centre has just launched another company, CU-Tech Limited, which provides building inspection services using unmanned systems technology, commonly known as drones. Professor Chen Benmei and Professor Chen Xi from the Department of MAE, who developed the system, note that there is a sizeable demand for efficient building inspection services in Hong Kong, where about 60% of private buildings are older than 30 years. Professor Chen Xi says: “Traditionally, the industry needs to erect scaffolding and send workers at heights to identify, with the naked eye or hand- held tools, defects on the exteriors of a high-rise. But drones can do a much more accurate, faster and safer inspection. This will be the key step in collecting data for building owners to decide whether repair works are needed.” The automated inspection and information system can first plan a flight path for the cooperative drones, with algorithms written by the team, to conduct a comprehensive survey of a building’s facades within a shorter period of time than existing systems on the market, which rely on human control. When operating with three drones, the system can save at least two-thirds of the time to conduct a fully autonomous inspection, compared to a single-drone manual operation. (From left) Professors Koushil Sreenath and Mark Mueller from UC Berkeley came to meet their CUHK collaborators in July A solid-waste sorting system developed by SOTA How AI and robotics reshape the world 21 Robots for the workplace