Automatic Fruit Picking: An Entry-level, in Scale, Approach for Cyber Physical Systems Understanding

Abstract

The growing demand for food supplies, caused by the global population growth, and the continuously diminishing natural resources are posing many serious challenges. Furthermore, the economic and social uncertainty and the restrictions incurred by the recent COVID-19 pandemic are favouring the research for further fruit harvesting alternatives. The modernization of agriculture, involving interoperation among networking, machine learning, robotics, and big data entities seems to provide a promising direction to tackle the abovementioned problems. These systems, officially called cyber physical ones, can have a key role in the digital transformation of the agricultural sector, provided that people getting involved, from students and scientists of agricultural engineering to farmers and consulting professionals, are capable of understanding and using the diverse set of innovative applications belonging in this area. In this regard, taking advantage of the wide availability of cheap electronic components and suitable programming environments, modern educational practices can be further updated to provide tailored agricultural engineering solutions. The work presented herein describes the efforts to utilize a small robotic arm to identify and pick small fruits, assisted by a smart camera and a voice-commanded module. Despite the selection of low cost commercial off-the-self components and the small size of the final implementation, the main challenges towards the realization of a full-scale system, suitable for farm use, are fluently outlined while additional evaluation results are provided, from the educational point of view, as well.