Advancement and field evaluation of a dual-arm apple harvesting robot

Authors: ZHU, KEYI - Michigan State University; LAMMERS, KYLE - Michigan State University; ZHANG, KAIXIANG - Michigan State University; ARUNACHALAM, CHAARAN - Michigan State University; BHATTACHARYA, SIDHARTHA - Michigan State University; LI, JIAJIA - Michigan State University; Lu, Renfu; LI, ZHAOJIAN - Michigan State University

Submitted to: Smart Agricultural Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/19/2025
Publication Date: 8/28/2025
Citation: Zhu, K., Lammers, K., Zhang, K., Arunachalam, C., Bhattacharya, S., Li, J., Lu, R., Li, Z. 2025. Advancement and field evaluation of a dual-arm apple harvesting robot. Smart Agricultural Technology. 12. Article 101343. https://doi.org/10.1016/j.atech.2025.101343.
DOI: https://doi.org/10.1016/j.atech.2025.101343

Interpretive Summary: Fruit harvesting still relies on manual labor and is the largest cost for apple production in the U.S. Rising costs and shortages of skilled manual labor are top concerns for apple growers. Labor-reducing harvest automation technologies are thus urgently needed. Despite considerable research and development efforts in recent years, the existing robotic harvesting systems still fall short of meeting grower expectations in terms of performance, cost effectiveness, and robustness in working in complex orchard environments. This paper reports on the development and evaluation of an improved dual-arm apple harvesting robot. The robot incorporates a centralized vacuum system, along with smart sensors, and a dual-arm coordination strategy, for optimizing the picking activities of the two arms to enhance harvesting efficiency and system reliability. A robust apple detection and localization pipeline was implemented for enhanced picking performance. Field evaluations were conducted in two commercial orchards with different tree structures. The robotic system achieved successful picking rates of 80.7% and 79.7%, respectively, with an average picking speed of 5.97 seconds per fruit by each arm, which represents 28% improvement, compared with a single-arm robot. With further improvements, the dual-arm harvesting robot can provide a commercially viable solution to automated harvesting of apples, thus helping U.S. fruit growers mitigate the rising production cost and increasing shortage of labor.

Technical Abstract: Currently, apple harvesting still relies on manual labor, which is the largest cost component in apple production. Automated harvesting has thus attracted increasing attention in recent years. However, the existing robotic harvesting systems still fall short in terms of harvesting performance, cost effectiveness, and robustness. This paper reports on the development and evaluation of an improved dual-arm apple harvesting robot. The system integrates a Time-of-Flight camera, two 4-degree-of-freedom robotic arms, a centralized vacuum system, and a fruit gathering module. During harvesting, suction force is dynamically assigned to either arm via the centralized vacuum system, enabling efficient apple detachment, while significantly reducing power consumption and noise. A platform movement mechanism was incorporated in the system to enable both horizontal and vertical adjustments, thus enhancing the robot’s dexterity and adaptability to varying canopy structures. A robust apple localization pipeline that combines a foundation-model-based detector, pixel-wise segmentation, and clustering-based depth estimation, was implemented to improve the robot’s picking performance. Additionally, pressure sensors were integrated into the vacuum system, and a novel dual-arm coordination strategy was introduced to effectively respond to harvest failures based on sensor feedback, further improving picking efficiency. Field evaluations were conducted in two commercial orchards with different tree structures in Michigan, USA. The system achieved successful picking rates of 80.7% and 79.7%, respectively, with an average picking cycle time of 5.97 seconds. The proposed coordination strategy reduced harvest time by 28%, compared with a single-arm baseline. The dual-arm harvesting robot enhances the reliability and efficiency of apple picking. With further improvements in hardware and software, the harvesting robot is promising for fully autonomous operation and future commercialization to support the apple industry.