Performance Evaluation of AI, Machine Learning, and IoT-Based Smart Agriculture Systems for Environmental Sustainability

Abstract

Agriculture is historically under a lot of pressure, mainly due to the lack of water, climate, soil, over use of fertilizers and pesticides, energy for farm operations and low use of resources in agricultural production. Farming practices can lead to low crop yields and have a large negative impact on the environment. The adoption of artificial intelligence (AI), machine learning (ML) and Internet of Things (IoT) systems and technologies in agriculture have the potential to monitor, predict and make decisions in real time. This article is to compare and evaluate the performance of AI, ML and IoT-based smart agricultural systems in monitoring, prediction and decision-making. Data from field sensors, weather stations, crop monitoring and automated irrigation systems are incorporated in the performance evaluation of smart agricultural systems. The use of ML in agricultural crop yield prediction, plant disease detection, automatic irrigation systems and soil condition monitoring is evaluated. The accuracy, precision, recall, F1-score, irrigation water use efficiency, water saving rate, energy consumption, crop yield, latency and cost of smart agricultural systems are some of the indicators that can be used to evaluate their performance. This article proposes an evaluation framework that can be used to assess the performance of smart agricultural systems based on their technical characteristics and their impact on agriculture and the environment. The main advantage of this article is that it can support farmers, researchers and policy makers to understand how smart agriculture can help reduce the wasted use of resources, increase crop production and promote more sustainable agricultural practices.