Manipulator with\nIntegrated Flexible Tactile Sensing\nArrays for Kiwifruit Ripeness and Size Classification

Fruit grading for ripeness and size is an essential process\nin\nthe supply chain. Incorrect grading can easily lead to spoiled and\ndegraded fruits entering the market, reducing consumers’ confidence\nin purchasing. At the same time, it is easy to cause the fruit supply\nchain to reduce profits, unreasonable resource allocation, and related\npractitioners’ income. The current mainstream machine vision\ngrading and manual grading in the production line have dilemmas such\nas susceptibility to environmental interference, inconsistent grading\nstandards, high cost, and labor shortage. To overcome these problems,\nthis study proposes an integrated flexible tactile sensing array (3\n× 4) manipulator for efficient, stable, low-cost, and accurate\nripeness and size grading of kiwifruit. The flexible sensing manipulator\ngrasps the kiwifruit, detects the hardness of the kiwifruit by relying\non tactile sensing, and determines the ripeness level based on the\nhardness. The size of the kiwifruit is also differentiated according\nto whether there is a significant change in the resistance of the\ntopmost sensing unit of the flexible pressure sensor array. The 0,\n1, 2, 3, 4, and 5 anomalies that may occur in actual production were\ntested and combined with machine learning KNN, SVM, and RF algorithms\nfor data modeling and grading. The results show that the lowest accuracy\nof 0, 1, 2, 3, 4, and 5 possible outliers is 86.67% (KNN), 95.83%\n(SVM), and 92.5% (RF), respectively. KNN has the lowest classification\neffect, and SVM has the best. This study overcomes the drawbacks of\ninefficient destructive detection and unstable manual detection and\nmakes up for the vulnerability of single machine vision to interference\nfrom environmental factors. This study can alleviate the challenges\ncaused by fruit wastage and promote the sustainable production and\nconsumption of the fruit industry chain.