Implementation of fruit plucking robot in apple harvesting: A review

This paper presents a review of the implementation of the fruit-plucking robot in apple harvesting. Robotic apple harvesting has several advantages, particularly in reducing labor costs and shortage, improving efficiency, real-time data collection, and monitoring. The joint efforts of academics from all around the world have resulted in a large number of prototypes being developed since the French did the initial study on the apple-harvesting robot >30 years ago. Due to its ineffective harvesting abilities, the prototype apple harvesting robot remains in the experimental research stage. With the use of artificial intelligence, the related research has ushered in a significant improvement that offers advantages as well as challenges for researchers studying robotic apple harvesters. Early prototypes from 1985 have a recognition rate of around 80 % and harvesting time in a range of tens of seconds per apple. Modern robots have now improved these metrics. This review presents different techniques of robotic apple harvesting, their literature review, composition, and structure. The research findings for apple harvesting were then methodically compiled into the following three categories: quick and accurate target fruit recognition and positioning, robotic hand design for safe grasping of fruit, and application of intelligent computing theory in apple harvesting robots. Different methods and techniques used for different fruit recognition and harvesting are also presented in this review in tabular form. After that, there are some challenges of robotic fruit harvesting, and finally, there are future recommendations for future research in the field.