Agbesi et al. (2026) Development and performance evaluation of a low-cost sensor-based automated drip irrigation system for small-scale farming

Identification

• Journal: Irrigation Science
• Year: 2026
• Date: 2026-03-26
• Authors: Wisdom Eyram Kwame Agbesi, Justin H. Chepete, T. Mpuisang, Dominic Aboagye
• DOI: 10.1007/s00271-026-01109-1

Research Groups

• Department of Agricultural and Biosystems Engineering, Botswana University of Agriculture and Natural Resources, Gaborone, Botswana
• Department of Mechanical Engineering, School of Engineering, Cape Coast, Ghana
• Department of Chemical Engineering, Universität Rovira i Virgili, Tarragona, Spain

Short Summary

This study developed and evaluated a low-cost, sensor-based automated drip irrigation system for small-scale farming, demonstrating reliable performance in fine-textured soils (clay) compared to a commercial sensor, while identifying limitations in coarse-textured soils due to compaction and probe-soil contact issues.

Objective

• To develop an automated microcontroller/sensor-based drip irrigation system.
• To evaluate its performance as an automated drip irrigation method (ASISM) for irrigating crops grown in clay, sandy-loam, and sandy soils.

Study Configuration

• Spatial Scale: Laboratory setting using PVC columns (100 mm diameter × 400 mm height) filled with compacted clay, sandy-loam, and sand soils. Root zone depths for sensor placement were 57 mm, 114 mm, and 229 mm. Water storage tanks with volumes of 12.064 L and 23.6 L were used.
• Temporal Scale: Soil samples were oven-dried for 24 hours for calibration. Sensor readings were recorded after a 30-second stabilization period. The experiment was replicated three times for each soil type. Average irrigation durations ranged from 18.9 minutes to 25.8 minutes per event.

Methodology and Data

• Models used:
  • Custom calibration equations (linear or exponential regression) relating sensor frequency output (Hz) to volumetric soil moisture content (%) for NE555 soil moisture sensors.
  • Custom calibration equations relating millivolt (mV) output to volumetric soil moisture content (%) for the MPM160 reference sensor.
  • Custom calibration equations relating sensor frequency output (Hz) to water column depth (mm) for NE555 water level sensors.
  • Statistical analysis: Paired sample t-test (alpha = 0.05), Root Mean Square Error (RMSE), Mean Bias Error (MBE), Analysis of Variance (ANOVA), and Tukey simultaneous mean comparison.
• Data sources:
  • In-house developed NE555-based soil moisture and water-level sensors (frequency output).
  • Commercial reference soil moisture sensor: MPM160 (Irrigation Crop Technologies, Australia) (millivolt output).
  • Laboratory measurements: Soil textural analysis (hydrometer method), gravimetric water content (oven drying at 105 °C for 24 hours), bulk density.
  • System logs: Irrigation duration, water application volumes.

Main Results

• The developed NE555 sensor-based system showed strong agreement with the commercial MPM160 reference sensor in clay soils:
  • At irrigation start: RMSE = 1.9%, MBE = -0.69% (volumetric moisture content), with no significant difference (p > 0.05).
  • At irrigation termination: RMSE = 5.2%, MBE = -1.9% (volumetric moisture content), with no significant difference (p > 0.05).
• In sandy loam, the NE555 sensor performed fairly at irrigation termination (RMSE = 8.4%) but showed unreliability at low-moisture levels (MBE = -2.4%) at irrigation start, with significant differences (p < 0.05) compared to MPM160.
• In sand, significant differences (p < 0.05) were observed between NE555 and MPM160 sensors at both irrigation start (RMSE = 2.3%, MBE = -1.6%) and termination (RMSE = 6.4%, MBE = -5.6%), attributed to compaction limitations and probe-soil contact issues.
• Average irrigation durations were 18.9 minutes for clay, 25.8 minutes for sandy loam, and 19.4 minutes for sand.
• Corresponding average water volumes applied were 1280 mL for clay, 1073 mL for sandy loam, and 896 mL for sand.
• The developed system cost USD 108, significantly lower than the standard MPM160 sensor (approximately USD 1,300).

Contributions

• Development of a low-cost, sensor-based automated drip irrigation system tailored for small-scale farmers, operating offline without internet connectivity and requiring minimal technical expertise.
• Comprehensive performance evaluation of the system across different soil textures (clay, sandy loam, sand) against a commercial reference sensor.
• Demonstrated reliable performance of the low-cost NE555 sensors in fine-textured soils (clay), validating their applicability for precision irrigation in such environments.
• Identified and quantified limitations of the low-cost sensors in coarse-textured soils (sandy loam and sand), specifically due to compaction variability and probe-soil contact constraints, providing a basis for future research.
• Offers a practical and affordable solution to enhance water-use efficiency and food security in water-scarce regions, particularly for smallholder farming systems.

Funding

The author's MSc studies, which led to this research, were supported by the “Mobility for Engineering and Technology Graduates in Africa (METEGA)” project, funded by the European Union under the Intra-ACP Academic Mobility Call. The research itself received no specific grant from any funding agency.

Citation

@article{Agbesi2026Development,
  author = {Agbesi, Wisdom Eyram Kwame and Chepete, Justin H. and Mpuisang, T. and Aboagye, Dominic},
  title = {Development and performance evaluation of a low-cost sensor-based automated drip irrigation system for small-scale farming},
  journal = {Irrigation Science},
  year = {2026},
  doi = {10.1007/s00271-026-01109-1},
  url = {https://doi.org/10.1007/s00271-026-01109-1}
}