Your Name
Carlos Andrés Erazo and Juan Felipe Rubiano
Email
juanfeliperzmcs@gmail.com
GitHub Repository Link
https://github.com/Juan-Felipe-Rubiano/AgroMesh
Project Description
AgroMesh: A Mestastic-Powered Distributed Soil Profiling and Irrigation Prediction System
Context
Internet connectivity in colombian agriculture
Agriculture plays a critical role in Colombia's economy. In 2024, more than 5.5 million hectares of crops were registered across the country, with significant concentrations in departments such as Meta (10%), Antioquia (7.8%), Valle del Cauca (6.8%), and Tolima (6.5%). As shown in Figure 1, departments containing more than 100,000 hectares of cultivated land account for a substantial portion of the country's agricultural production.
Despite their agricultural importance, many of these regions still experience limited Internet penetration and inconsistent connectivity. Figure 2 highlights that some of Colombia's most productive agricultural departments do not enjoy full Internet coverage, creating a significant barrier for the adoption of modern digital agriculture technologies.
This lack of connectivity presents a major challenge for farmers who could benefit from real-time monitoring, predictive analytics, and data-driven decision-making.
Figure 1. Colombian departments with more than 100,000 hectares of cultivated land. Most large-scale agricultural regions are concentrated outside major urban centers.
Figure 2. Internet penetration across Colombian departments. Several major agricultural regions still experience limited connectivity, restricting the adoption of conventional IoT solutions.
Irrigation: A critical yet uncertain decision
Large-scale agricultural enterprises often rely on advanced precision agriculture platforms capable of monitoring crops, machinery, and environmental conditions in real time. However, these solutions are typically expensive and inaccessible to small and medium-sized farmers, who represent a significant portion of Colombia's agricultural sector.
Through conversations with local farmers, we identified irrigation management as one of their most significant challenges. Today, irrigation decisions are frequently based on empirical observations, including:
- The visible color of the soil
- Weather conditions observed during the previous days
- The time elapsed since the last irrigation event
While farmers possess invaluable experience and expertise, these methods provide only surface-level information and may hide critical conditions occurring beneath the soil.
For example, a field may appear moist on the surface while deeper soil layers remain dry. In such cases, water may not be reaching the root zone effectively, causing hidden water stress despite recent irrigation. Conversely, excessive irrigation may lead to water waste, nutrient leaching, and reduced soil health.
As a result, irrigation often occurs too early or too late, negatively impacting crop productivity, water efficiency, and long-term soil sustainability.
The Challenge
Developing a practical solution for this problem is difficult due to three fundamental constraints:
- Large agricultural areas
Agricultural plots may extend over dozens or even hundreds of hectares, requiring communication technologies capable of operating over long distances.
- Limited connectivity infrastructure
Traditional IoT solutions often rely on Wi-Fi, cellular networks, or fixed Internet infrastructure. In many rural regions, these services are unavailable, unreliable, or economically impractical.
- Cost and maintenance constraints
Small farmers require solutions that are affordable, energy-efficient, and capable of operating autonomously for extended periods without intervention.
These challenges make conventional smart agriculture deployments difficult to implement at scale.
Why Meshtastic
LoRa technology provides long-range, low-power communication, making it an attractive option for rural environments. However, traditional point-to-point LoRa deployments introduce new challenges:
- Sensors located far from the gateway may be unable to communicate directly
- Expanding coverage often requires additional gateways and infrastructure
- Network redesign may become necessary as new sensors are added
Meshtastic addresses these limitations by building a self-healing mesh network on top of LoRa.
In a Meshtastic network, each node acts not only as a sensor but also as a relay capable of forwarding messages from neighboring nodes. This creates several advantages:
- Extended coverage: Data can travel across multiple hops until reaching a gateway
- Scalability: New nodes can be added without redesigning the network
- Resilience: Communication can continue even if some nodes become unavailable
- Infrastructure independence: Large areas can be covered without cellular networks or extensive communication infrastructure
This architecture is particularly suitable for agricultural environments, where fields are geographically distributed and connectivity is often limited.
The solution: AgroMesh
AgroMesh is a distributed soil intelligence platform designed to help farmers optimize irrigation through continuous monitoring of soil moisture dynamics.
The core of the system is a solar-powered Meshtastic-enabled stake designed for outdoor deployment.
Each stake contains multiple soil moisture sensors positioned at fixed depths along a 50 cm profile, allowing measurements at different layers of the soil. A mechanical insertion guide ensures that all stakes are installed consistently, guaranteeing comparable measurements across different locations.
Unlike conventional moisture sensors that only measure near-surface conditions, AgroMesh captures the vertical movement of water throughout the soil profile.
This enables the system to:
- Monitor water infiltration after irrigation events
- Observe how moisture propagates through the root zone
- Measure soil drying rates over time
_ Identify potential water stress before it becomes visible
- Detect inefficient irrigation practices
Each node transmits its measurements through the Meshtastic mesh network, allowing information to reach the farmer even when sensors are located far from the nearest gateway.
Beyond monitoring: predictive irrigation
AgroMesh is not intended to be just another sensor network.
By combining:
- Multi-depth soil moisture measurements
- Historical irrigation patterns
- Weather forecasts
- Environmental conditions
the platform can estimate soil drying behavior and predict future irrigation requirements.
The collected data is aggregated into spatial soil moisture maps, providing farmers with a real-time representation of water distribution across their fields.
Machine learning models can then analyze these patterns to:
- Predict when irrigation will be required
- Identify areas experiencing abnormal drying
- Detect potential irrigation inefficiencies
- Support data-driven water management decisions
Expected impact
AgroMesh aims to bring precision agriculture capabilities to farmers who currently lack access to expensive commercial solutions.
By combining low-power hardware, Meshtastic mesh networking, distributed sensing, and predictive analytics, the platform enables:
- Reduced water consumption
- Improved irrigation efficiency
- Better crop productivity
- Early detection of soil stress conditions
- Deployment in remote areas with little or no Internet connectivity
Ultimately, AgroMesh transforms irrigation from an empirical practice into a data-driven decision process, empowering farmers with actionable insights while maintaining affordability and scalability.
Tech Stack
Hardware Layer
- Seed Studio Wio-SX1262 Wireless Module
- Seed Studio XIAO Esp32 (TBD)
- Soil moisture sensors
- Solar panel
- Lithium Battery
- (...TBD)
Networking Layer
Backend Layer
Data Layer
- PostgreSQL
- TimescaleDB
- PostGIS
Machine learning Layer
Frontend Layer
A simplified overview of what the architecture of the project would look like:
Message to the Organizers
No response
Your Name
Carlos Andrés Erazo and Juan Felipe Rubiano
Email
juanfeliperzmcs@gmail.com
GitHub Repository Link
https://github.com/Juan-Felipe-Rubiano/AgroMesh
Project Description
AgroMesh: A Mestastic-Powered Distributed Soil Profiling and Irrigation Prediction System
Context
Internet connectivity in colombian agriculture
Agriculture plays a critical role in Colombia's economy. In 2024, more than 5.5 million hectares of crops were registered across the country, with significant concentrations in departments such as Meta (10%), Antioquia (7.8%), Valle del Cauca (6.8%), and Tolima (6.5%). As shown in Figure 1, departments containing more than 100,000 hectares of cultivated land account for a substantial portion of the country's agricultural production.
Despite their agricultural importance, many of these regions still experience limited Internet penetration and inconsistent connectivity. Figure 2 highlights that some of Colombia's most productive agricultural departments do not enjoy full Internet coverage, creating a significant barrier for the adoption of modern digital agriculture technologies.
This lack of connectivity presents a major challenge for farmers who could benefit from real-time monitoring, predictive analytics, and data-driven decision-making.
Figure 1. Colombian departments with more than 100,000 hectares of cultivated land. Most large-scale agricultural regions are concentrated outside major urban centers.
Figure 2. Internet penetration across Colombian departments. Several major agricultural regions still experience limited connectivity, restricting the adoption of conventional IoT solutions.
Irrigation: A critical yet uncertain decision
Large-scale agricultural enterprises often rely on advanced precision agriculture platforms capable of monitoring crops, machinery, and environmental conditions in real time. However, these solutions are typically expensive and inaccessible to small and medium-sized farmers, who represent a significant portion of Colombia's agricultural sector.
Through conversations with local farmers, we identified irrigation management as one of their most significant challenges. Today, irrigation decisions are frequently based on empirical observations, including:
While farmers possess invaluable experience and expertise, these methods provide only surface-level information and may hide critical conditions occurring beneath the soil.
For example, a field may appear moist on the surface while deeper soil layers remain dry. In such cases, water may not be reaching the root zone effectively, causing hidden water stress despite recent irrigation. Conversely, excessive irrigation may lead to water waste, nutrient leaching, and reduced soil health.
As a result, irrigation often occurs too early or too late, negatively impacting crop productivity, water efficiency, and long-term soil sustainability.
The Challenge
Developing a practical solution for this problem is difficult due to three fundamental constraints:
Agricultural plots may extend over dozens or even hundreds of hectares, requiring communication technologies capable of operating over long distances.
Traditional IoT solutions often rely on Wi-Fi, cellular networks, or fixed Internet infrastructure. In many rural regions, these services are unavailable, unreliable, or economically impractical.
Small farmers require solutions that are affordable, energy-efficient, and capable of operating autonomously for extended periods without intervention.
These challenges make conventional smart agriculture deployments difficult to implement at scale.
Why Meshtastic
LoRa technology provides long-range, low-power communication, making it an attractive option for rural environments. However, traditional point-to-point LoRa deployments introduce new challenges:
Meshtastic addresses these limitations by building a self-healing mesh network on top of LoRa.
In a Meshtastic network, each node acts not only as a sensor but also as a relay capable of forwarding messages from neighboring nodes. This creates several advantages:
This architecture is particularly suitable for agricultural environments, where fields are geographically distributed and connectivity is often limited.
The solution: AgroMesh
AgroMesh is a distributed soil intelligence platform designed to help farmers optimize irrigation through continuous monitoring of soil moisture dynamics.
The core of the system is a solar-powered Meshtastic-enabled stake designed for outdoor deployment.
Each stake contains multiple soil moisture sensors positioned at fixed depths along a 50 cm profile, allowing measurements at different layers of the soil. A mechanical insertion guide ensures that all stakes are installed consistently, guaranteeing comparable measurements across different locations.
Unlike conventional moisture sensors that only measure near-surface conditions, AgroMesh captures the vertical movement of water throughout the soil profile.
This enables the system to:
_ Identify potential water stress before it becomes visible
Each node transmits its measurements through the Meshtastic mesh network, allowing information to reach the farmer even when sensors are located far from the nearest gateway.
Beyond monitoring: predictive irrigation
AgroMesh is not intended to be just another sensor network.
By combining:
the platform can estimate soil drying behavior and predict future irrigation requirements.
The collected data is aggregated into spatial soil moisture maps, providing farmers with a real-time representation of water distribution across their fields.
Machine learning models can then analyze these patterns to:
Expected impact
AgroMesh aims to bring precision agriculture capabilities to farmers who currently lack access to expensive commercial solutions.
By combining low-power hardware, Meshtastic mesh networking, distributed sensing, and predictive analytics, the platform enables:
Ultimately, AgroMesh transforms irrigation from an empirical practice into a data-driven decision process, empowering farmers with actionable insights while maintaining affordability and scalability.
Tech Stack
Hardware Layer
Networking Layer
Backend Layer
Data Layer
Machine learning Layer
Frontend Layer
A simplified overview of what the architecture of the project would look like:
Message to the Organizers
No response