Research progress of the kiwifruit operation robots and prospects for fully intelligent production

Kiwifruit is recognized as a favorite fruit crop in China, due to the significant economic and nutritional value. However, the large-scale production has still remained largely on the manual labor, particularly in the key operations, such as the pollination, pruning, thinning, and harvesting. Efficiency and quality are often required in its rapid industrial expansion in recent years, due to the highly labor-intensive and time-sensitive. It is a pressing need for the labor-saving, precise, and reliable alternatives in industry, with the increasing cost of seasonal labor. Among them, the agricultural robots have received growing attention in kiwifruit production. As a representative direction of smart agriculture, the agricultural robots can typically integrate the machine vision, intelligent decision-making, and automated execution. The broad application prospects can also improve the precision for the standardized management of the orchards throughout the year. Multi-degree-of-freedom manipulators, and environment-aware perception modules can also suitable for the complex environment of field orchard. This study aims to systematically review the recent technologies in the kiwifruit robots, particularly on their system structures, perception strategies, motion planning, and end-effector designs. The robot enabling technology was also examined to clarify the current research landscape and development bottlenecks in intelligent horticulture. Special emphasis was placed on the robotic applications during pollination and harvesting. The system performance was evaluated to fully meet the short operation windows and high accuracy requirements. Meanwhile, the intelligent robotic equipment was integrated into the broader context of fully intelligent kiwifruit production (such as bud and young fruit thing, as well as branch pruning). The entire production cycle was involved the sensing, decision-making, execution, and feedback. The selective operation techniques were further enhanced the production efficiency. The targeted interventions were employed to manage only the necessary fruits, flowers, or buds. This selective process was optimized the resource use for the overall productivity without excessive input. Furthermore, a comprehensive framework was also proposed to adapt the intelligent technologies in real-world orchards, including the spatial compatibility between robotic equipment and orchard infrastructure, biological coordination with the phenological development of crops, and the modularity in equipment design to accommodate diverse operational scenarios. The need was highlighted to coordinate the planning in orchard, the standardization in cultivation practices, and equipment innovation. An integrated production was fully met the demands of modern fruit farming. Representative case studies and experimental results were presented to assess the feasibility, efficiency, and practical challenges after robotic deployment in kiwifruit production. Strategic guidance can also provide for the future innovation, standardization, and industrial application, ultimately supporting the transition toward a fully automated, efficient, and sustainable kiwifruit production.