Intelligent Breeding Robot Achieves Automated Pollination for Greenhouse Crops

A team from the Institute of Automation, Chinese Academy of Sciences (CAS), in collaboration with the Institute of Genetics and Developmental Biology, CAS, has developed a fully automated method and robotic system to accelerate breeding and seed production processes in modern greenhouses. The research, titled "Engineering tomato floral morphology facilitates robotization of cross-pollination and speed breeding," was published online in the journal Cell on August 11, 2025. This breakthrough represents a deep integration of biological technology, artificial intelligence, and robotics, establishing a new intelligent breeding paradigm—referred to as BAR (Biological Technology + AI + Robotics). The innovation overcomes longstanding bottlenecks in traditional hybrid breeding and seed production, significantly reducing costs, shortening breeding cycles, and improving efficiency. The achievement marks China’s leadership in building a closed-loop intelligent robotic breeding technology system and demonstrates the transformative potential of "AI for Science" in advancing new-quality productivity and reshaping the future of agricultural biotechnology. For thousands of years, humans have domesticated over 400,000 plant species, selecting just over 40 for large-scale cultivation in greenhouses and fields—these form the foundation of global food security. In modern greenhouse crop production, the breeding process typically involves four stages: seed, management, growth, and harvest. However, for both cross-pollinating crops (such as cucurbits and crucifers) and self-pollinating crops (like solanaceous and gramineous plants), the breeding phase still heavily relies on manual labor to carefully remove anthers and transfer pollen to the stigma. This manual hybridization process is extremely labor-intensive, time-consuming, and requires high precision—making it the primary obstacle to full automation in greenhouse breeding. To address this challenge, the research team developed advanced methods for precise perception of minute targets under pollination constraints, compliant manipulation of delicate reproductive structures, and accurate positioning of robots on greenhouse tracks. For self-pollinating crops, they further engineered a male-sterile line with exposed stigmas using molecular breeding techniques. Combining these AI-driven robotic and molecular approaches, the team designed and deployed an intelligent breeding robot at the Beijing Shounong Cuihu Industrial Base North Zone Breeding Demonstration Site (Figure 2). The robot achieved a stigma recognition accuracy of 85.1%, with an average pollination time of just 13 seconds per flower. During a single autonomous patrol, the system achieved a success rate of 77.6% ± 9.4% in successful pollination, enabling continuous, round-the-clock operation to ensure consistent fruit set. In terms of pollination efficiency, the robot matched human performance for cross-pollinating crops, while significantly outperforming human labor for self-pollinating crops. This demonstrates strong potential for accelerating the development of climate-resilient crop varieties, enhancing efficiency, and reducing costs. The study provides a scalable technological blueprint for precision agriculture. Its “biological design–machine adaptation” co-optimization strategy offers a smart solution to food security challenges in the face of climate change. Future applications may extend to automated phenotyping, intelligent harvesting, and other stages of agricultural automation. All robot components are now domestically produced, highlighting strong self-reliance and broad application prospects. Minghao Yang, Deputy Researcher at the Institute of Automation, CAS, is a co-first author and led the AI and robotics development. Graduate students Sun Yangchang (2021), Lyu Hongchang (2023), Wang Jinyang (2025), and interns Xiao Jun, Qi Jingda, Liu Anqi, and Xiao Zhigang contributed to robotic vision, dexterous manipulation, system integration, and autonomous navigation in greenhouses. Cao Xu, Researcher at the Institute of Genetics and Developmental Biology, CAS, is the corresponding author. Co-first authors Yue Xie and Tinghao Zhang led the molecular breeding design. Hua Han, Researcher at the Institute of Automation, CAS, and Professor Tao Jianhua from Tsinghua University also made significant contributions. Support was provided by Li Xinxu and Li Shushan from Shounong Cuihu Base. The work was supported by the CAS Huai’rui Brain Cognitive Functional Atlas and Bionic Intelligence Interdisciplinary Research Platform, the National Natural Science Foundation of China’s Beijing-Tianjin-Hebei Basic Research Cooperation Project, the CAS Strategic Priority B-Class Project, and the Guangxi Key R&D Program. Paper references: Yue Xie, Tinghao Zhang, Minghao Yang, Hongchang Lyu, Yupan Zou, Yangchang Sun, Jun Xiao, Wenzhao Lian, Jianhua Tao, Hua Han, and Cao Xu#, "Engineering tomato floral morphology facilitates robotization of cross-pollination and speed breeding," Cell, 2025. Minghao Yang, Hongchang Lv, Yongjia Zhao, Yangcheng Sun, Hang Pan, Qi Sun, Jinlong Chen, Hongbo Yang, "Delivery of Pollen to Forsythia Flower Pistils Autonomously and Precisely Using a Robot Arm," Computers and Electronics in Agriculture, 214, 2023, pp. 108274–108287. DOI: https://doi.org/10.1016/j.cell.2025.07.028 Link: http://dl.acm.org/doi/10.1016/j.compag.2023.108274