Design and Implementation of Internet of Things (IoT) Platform Targeted for Smallholder Farmers: From Nepal Perspective

Abstract

Nepal, a lower-middle-income country in South Asia, predominantly features smallholder farming communities operating on modest land holdings. These smallholders often adhere to traditional farming methods, relying on familial labor, which has become increasingly inefficient in contemporary agricultural landscapes. To enhance their productivity and efficiency, smallholder farmers require affordable and accessible Internet of Things (IoT)-based systems. However, the prevailing IoT solutions in the market primarily cater to large-scale commercial enterprises, rendering them unsuitable for the specific needs and constraints faced by smallholder farmers. In response to this gap, we have introduced a cost-effective, customizable, scalable, and dependable IoT platform tailored expressly for smallholder farmers. This platform empowers them to visualize, monitor, and control real-time data pertaining to their crops, livestock, and other agricultural assets. To ascertain the efficacy and suitability of our proposed platform, we conducted a comparative analysis with existing counterparts such as Blynk IoT and ThingSpeak IoT, evaluating their respective features and application services against standard requirements. Additionally, we subjected our platform to rigorous server load testing, assessing crucial performance parameters including throughput, response time, user capacity, and data sampling rates. Over an observation period spanning an average of 339 days, our platform successfully processed and stored a substantial volume of data, encompassing 817,633 sensor messages, averaging 2412 messages per day, with a cumulative storage size of 14,238.28 KB. Extrapolating from these results, it is noteworthy that an A0 instance with 20 GB of cloud space can adequately accommodate 200 users at a rate of 100 MB per user, which is adequate for the smallholder needs. Furthermore, the purposed platform was deployed inside a polyhouse to perform off-season grafting of citrus plants. The achieved success rate of 84% closely approached the success rate of 90–95% observed during on-season grafting. These empirical findings, coupled with the extensive data gathered during our research, underscore the reliability and performance of our proposed IoT platform for smallholder farmers.

Evaluating the Impact of Controlled Ultraviolet Light Intensities on the Growth of Kale Using IoT-Based Systems

“…Additionally, concerns arise concerning data privacy and security, given the potentially sensitive nature of agricultural data and the potential ramifications for farm competitiveness or security [52]. Moreover, in remote or rural regions, the reliance on uninterrupted power supply and internet connectivity can pose a substantial obstacle, diminishing the dependability of IoT solutions [53]. As shown in Table 2, aspects of concern about implementing the IoT in agriculture encompassed a more comprehensive array of issues that could potentially impact the effectiveness and triumph of IoT systems.…”

“…Network stability and connectivity [53] Stable network connections are critical to ensure continuous data transmission and system reliability, especially in remote or rural areas where connectivity issues occur.…”

Internet of Things for Crop Farming: A Review of Technologies and Applications

“…Modern weed removal methods blend artificial intelligence and Internet of Things technology for automation. An automated weed removal system involves advanced image processing to process the image of the site/field and algorithms for feature extractions such as Stepwise Linear Discriminant Analysis (SWLDA), Principal Component Analysis (PCA), and Linear Discriminant Analysis (LDA) [48,49]. Classifiers such as Support Vector Machines (SVM), Artificial Neural Networks (ANN), and Convolutional Neural Network (CNN) are used to classify or categorize weeds from the desired plants [46][47][48][49][50][51][52][53].…”

“…An automated weed removal system involves advanced image processing to process the image of the site/field and algorithms for feature extractions such as Stepwise Linear Discriminant Analysis (SWLDA), Principal Component Analysis (PCA), and Linear Discriminant Analysis (LDA) [48,49]. Classifiers such as Support Vector Machines (SVM), Artificial Neural Networks (ANN), and Convolutional Neural Network (CNN) are used to classify or categorize weeds from the desired plants [46][47][48][49][50][51][52][53]. The IoT platform is made up of processor (i.e., Arduino UNO and Raspberry), fitted with a camera, RGB and IR imaging sensors to capture images from the site, and a sprayer to spray herbicide on the detected weed.…”

Agro-Investments among Small Farm Business Entrepreneurs in the Era of the Fourth Industrial Revolution: A Case in the Mpumalanga Province, South Africa

“…A study about the application of technologies such as the internet of things (IoT) among small farmers in Nepal for the optimal growth of mushrooms indicates the real context in which the fourth industrial revolution technologies can be applicable (Lamsal et al 2023). Additionally, a study by Furuholt and Matotay (2011) on the developmental contribution of mobile phones across the agricultural value chain in rural Africa discovered that smart mobile phones boosted the profits of farmers by 15% in Tanzania.…”