Development of IoT-Enabled Smart Greenhouse for Smallholder Farms

Thesis Abstract

The agricultural sector in sub-Saharan Africa faces persistently low productivity, climate variability, and reliance on manual farming practices, which collectively constrain smallholder farmers' capacity to optimize yields and resource use. While Internet of Things (IoT) technologies offer real-time environmental monitoring and automated control, their adoption remains limited due to affordability, data literacy, and integration challenges with existing farm routines. This study aims to develop and evaluate an IoT-enabled smart greenhouse system designed for smallholder farms, with the objective of enhancing crop yield, resource-use efficiency, and decision-making support. Specific objectives include (i) designing a modular, cost-effective IoT architecture comprising environmental sensors, actuated climate control, and a low-power gateway; (ii) assessing the impact of automated irrigation, temperature, and humidity regulation on tomato productivity under smallholder conditions; (iii) evaluating system usability and farmer adoption factors using a Technology Acceptance Model (TAM) adapted for agricultural ICT; (iv) estimating economic viability through partial budgeting and return on investment under varying market and climate scenarios; and (v) developing a scalable deployment framework for replication in similar agro-ecologies. A mixed-methods research design combines a quasi-experimental field trial with qualitative stakeholder insights. The population comprises 60 smallholder tomato plots across three rural communities, with 40 plots allocated to the intervention and 20 to a control group following conventional practices. Data collection instruments include (i) sensor-driven data loggers (temperature, relative humidity, soil moisture, solar radiation, and CO2) with a 15-minute sampling interval; (ii) automated actuators for irrigation, shading, and ventilation controlled by a microcontroller-based gateway; (iii) crop yield records and input use data collected at harvest; (iv) structured surveys and interviews to capture farmer perceptions, usability, and perceived usefulness; and (v) cost records for a comprehensive economic analysis. Validity and reliability are addressed through instrument calibration, pilot testing, and Cronbach’s alpha assessment for survey scales. Data analysis employs a combination of regression analysis to quantify the effect of the IoT system on yield and water-use efficiency, ANOVA to compare treatment and control groups across growing cycles, time-series analysis for environmental variables, and thematic analysis for qualitative interviews. Theoretical grounding draws on the Technology Acceptance Model (TAM) to frame usability and adoption, and the Diffusion of Innovations (DOI) theory to understand factors shaping scalable uptake. A systems-logic model will link sensor data to decision rules and crop outcomes, with a cost-benefit analysis projecting net present value and payback period under three climate scenarios. Expected findings indicate that the IoT-enabled greenhouse reduces irrigation water use by 25–40%, stabilizes canopy temperature within optimal ranges, and increases marketable tomato yield by 15–25% compared with conventional farming, while lowering input waste and labor demands. Sensor data quality and automated control are anticipated to improve crop vigor during heat stress events, with measurable gains in water-use efficiency (kg of fruit per liter of water). Qualitative results are expected to reveal key drivers and barriers to adoption, including perceived reliability, ease of use, and training needs, informing refinements to the user interface and local support structures. The study contributes to knowledge by integrating an affordable, modular IoT greenhouse platform with rigorous field evaluation and an explicit replication framework for resource-limited settings. It advances practical understanding of how ICT-driven agricultural interventions translate into measurable agronomic and economic benefits for smallholders, while identifying socio-technical conditions that enable sustainable uptake. The main conclusion anticipated is that a carefully designed, low-cost IoT greenhouse can meaningfully improve productivity and resource efficiency for smallholder tomato farming without imposing prohibitive upfront or operating costs. Recommendations include scalable co-design with farmer communities to tailor interfaces and alerts, policy guidance on subsidizing sensor hardware for low-income users, training programs to build ICT literacy, and the development of open-standards protocols to facilitate interoperability with existing farming tools. Suggested avenues for further research encompass long-term field trials across diverse crops and agro-ecologies, optimization of energy harvesting for off-grid deployments, and exploration of machine-learning-driven decision support for adaptive management under climate volatility.