Intelligent packaging for food continuously generates informative digital/analog content about the contained products during their entire life span, thus becoming one of the enabling elements of the modern data-driven economy. Packaging shells for fruits, augmented with low-cost wireless sensors for the automatic estimation of the ripening grade, can reduce waste, optimize shelf exposure, suggest when produce should be consumed, and engage customers through enhanced user experiences. Radio-frequency identification (RFID) with sensorless, low-cost labels, empowered with electromagnetic-based intelligence and automatic classification tools, may stimulate the widespread diffusion of this technology. Focusing on avocados, this article presents an experimental characterization of RFID's complex permittivity along with ripening and a near-field numerical model of a passive RFID interrogation system with tagged fruit, aimed at extracting the variation of electromagnetic metrics of the RFID link during ripening. The results are used to design and fabricate an RFID totem for avocado monitoring that, coupled with a properly trained binary tree classifier, is capable of recognizing up to three ripening levels of packaged fruits, with an overall accuracy higher than 85% even if the task is executed by unskilled operators.

中文翻译：

---

基于射频识别的智能包装：热带水果成熟的电磁分类

食品智能包装在其整个生命周期内不断生成有关所含产品的信息数字/模拟内容，从而成为现代数据驱动经济的支持要素之一。水果包装外壳，配备低成本无线传感器以自动估计成熟等级，可以减少浪费，优化货架曝光，建议何时应该消费产品，并通过增强的用户体验吸引客户。带有无传感器、低成本标签的射频识别 (RFID) 加上基于电磁的智能和自动分类工具，可能会刺激这项技术的广泛传播。本文以鳄梨为重点，介绍了 RFID 的实验表征 s 复介电常数与成熟以及带有标记水果的无源 RFID 询问系统的近场数值模型，旨在提取成熟过程中 RFID 链路的电磁指标的变化。结果用于设计和制造用于鳄梨监测的 RFID 图腾，该图腾与经过适当训练的二叉树分类器相结合，能够识别多达三个成熟水平的包装水果，即使执行任务，总体准确度也高于 85%由不熟练的操作员执行。