An Intelligent System for Fruit Harvesting Robot applied to Vertical Green Walls
Journal
2022 International Automatic Control Conference (CACS)
Part Of
2022 International Automatic Control Conference, CACS 2022
Start Page
1
End Page
6
ISBN (of the container)
978-166549646-9
Date Issued
2022-11-03
Author(s)
Zhan, Hong-Qi
Abstract
Issues related to environmental sustainability have been listed as important development goals in countries around the world with climate changes in these years. There have been successful cases of urban greening in many cities. This paper proposes an intelligent system for fruit harvesting robots applied to vertical green walls. In this study, we divide the intelligent system into two parts. The first one is the identification and grading of fruits on the green walls. Combining deep learning and computer vision techniques, we trained an object detection model which divided the fruits into three categories: unripe, ripe, and overripe. Ripe fruits were considered the harvesting target in this study. The next part is the design of the optimal harvesting path for the robot. The different harvesting orders of each fruit will affect the total moving distance of the harvesting robot, then the terrible traffic flow of the robot will lead to low efficiency of the overall harvesting task. This could be seen as a traveling salesman problem. Therefore, an optimal path algorithm based on self-organizing map was developed to find the shortest path which goes through every ripe fruit for the harvesting robot to follow. According to our experimental results, the mean average precision of the ripe fruit detection model is 93.85%. We executed the optimal path algorithm in the berlin52, xqf131, and qa194 instances in TSPLIB, VLSI TSPs, and National TSPs, the resulting path length was within 6% different from the best-known solution. © 2022 IEEE.
Subjects
Harvesting Robot
Intelligent System
Object Detection
Optimal Path Planning
Self-Organizing map
Vertical Green Wall
Publisher
IEEE
Type
conference paper