Soldered Anywhere logo Soldered Anywhere
Menu
← Back to Editorial Columns
Edge AI and the Global Water Crisis: How WaterLens is Democratizing Water Quality Detection visual summary
Analysis

Edge AI and the Global Water Crisis: How WaterLens is Democratizing Water Quality Detection

By Soldered Anywhere Editorial Team • 11/13/2026

The global water crisis is often framed as a problem of scarcity, but for billions of people, the issue is one of transparency. According to a 2025 report from the WHO and UNICEF, approximately one-quarter of the world’s population currently lacks consistent access to safely managed drinking water. Perhaps more startling is the finding that many people who technically have access to "basic" drinking water services are still consuming water that is not "safely managed"—meaning it may be contaminated, unreliable, or located far from their homes.

In this landscape of uncertainty, decentralized technology is stepping in where centralized infrastructure has failed. Thomas Vikström developed WaterLens, a tool designed to bridge the gap between "basic" access and "safe" consumption. By utilizing Artificial Intelligence at the edge, WaterLens represents a shift in how we monitor environmental health: moving away from expensive, slow laboratory tests toward real-time, localized intelligence.

The Data Gap: Basic vs. Safely Managed Water

To understand the necessity of projects like WaterLens, we must first look at the metrics used by global health organizations. "Basic" water service is defined as an improved source where the collection time is no more than 30 minutes for a round trip. However, "safely managed" water must meet three criteria: it must be accessible on-premises, available when needed, and free from fecal and priority chemical contamination.

The transition from basic to safely managed is the primary hurdle for developing regions. Traditional water testing requires taking samples to a laboratory, which involves logistical challenges, high costs, and significant delays. By the time a community learns their well is contaminated, the health impact has already occurred. This is where the concept of the "Engineering Student’s Toolkit" evolves from academic exercises into life-saving hardware. Aspiring engineers are now looking at how to deploy low-cost sensors to provide the same level of assurance as a professional lab.

For more on the foundational gear needed for such projects, see our guide on The Engineering Student’s Toolkit: Essential Gear for the New Academic Year.

Architecture of WaterLens: Edge AI on Arduino

WaterLens is built on the premise that data is most valuable when it is processed at the source. Using the Arduino ecosystem, Vikström implemented Edge AI (often referred to as TinyML) to analyze water quality parameters.

Unlike traditional IoT devices that simply upload raw data to a server for processing, an Edge AI device performs the heavy lifting locally. This is critical for water monitoring in remote areas for several reasons:

  1. Latency: Immediate feedback can prevent someone from drinking contaminated water.
  2. Connectivity: Many areas lacking safe water also lack reliable high-speed internet.
  3. Privacy and Autonomy: Communities can manage their own data without relying on third-party cloud providers.

The system typically monitors several key indicators, such as Total Dissolved Solids (TDS), pH levels, and Oxidation-Reduction Potential (ORP). While a single sensor reading might not tell the whole story, the AI model can recognize complex patterns—"fingerprints" of specific types of contamination—that a simple threshold alert might miss.

The Role of Sensors in Water Quality Analysis

To build a system like WaterLens, one must understand the variables that define water safety. For many hobbyists and engineers, the first step is establishing a baseline using digital tools. A common starting point is the TDS meter, which measures the concentration of dissolved combined content of all inorganic and organic substances in a liquid.

0-9990ppm 2PC TDS Meter Digital...

While a TDS meter is an excellent tool for a quick check, it is only one piece of the puzzle. To achieve the level of detection WaterLens aims for, a multi-sensor approach is required. This involves measuring:

  • pH Levels: To detect acidity or alkalinity that could indicate chemical runoff.
  • ORP (Oxidation-Reduction Potential): To measure the water's ability to cleanse itself or indicate the presence of sanitizers.
  • Conductivity (EC): Often linked to salinity and mineral content.

For those looking to implement a more permanent, automated solution, integrated monitors can provide a constant stream of data to an AI model or a mobile app.

8 in 1 WiFi Smart Water Quality...

These types of "8-in-1" monitors allow for the continuous data collection necessary to train a machine learning model. By observing how these variables fluctuate over time, an AI can learn to distinguish between natural seasonal changes and dangerous contamination events.

Hacking and Customizing Environmental Hardware

The spirit of WaterLens is deeply rooted in the "maker" philosophy—taking off-the-shelf components and pushing them beyond their intended use. This often involves hardware hacking to increase precision or add functionality. For instance, many budget sensors use standard LCDs that are difficult to read in the field or lack data logging capabilities.

Engineers often find themselves modifying these tools to stream data wirelessly to a central hub for better visualization. This is a common theme in hardware engineering, similar to how one might approach Wireless LCD Streaming for the ANENG AN870, where a standard tool is upgraded for modern data needs. By hacking the output of a standard water tester, a developer can feed that data directly into a TinyML pipeline on an Arduino.

Practical Steps for Localized Testing

While high-tech AI solutions are the future, there is still immense value in "analog" or hybrid testing methods, especially for immediate verification in the field. For those who do not yet have an Edge AI setup, comprehensive test kits provide a wide-spectrum baseline that can be used to verify the accuracy of electronic sensors.

16 in 1 Water Test Kit Strips 10...

These kits can test for 16 different parameters, including lead, fluoride, and bacteria, which are often difficult to detect with low-cost electronic sensors alone. Using these strips in conjunction with a digital monitor allows an engineer to "label" their data—telling the AI exactly what "contaminated" looks like in a specific geographical context.

Building a Monitoring Pipeline

If you are looking to develop a localized water quality project, the workflow generally follows these steps:

  1. Data Collection: Use sensors to gather pH, TDS, and ORP data over several weeks.
  2. Labeling: Use chemical test strips to verify the actual safety of the water during that period.
  3. Training: Feed the sensor data and the "labels" (Safe/Unsafe) into a TinyML platform like Edge Impulse.
  4. Deployment: Upload the trained model back to the Arduino (the "Edge" device).
  5. Monitoring: The device now alerts the user in real-time when sensor patterns match the "Unsafe" label.

The Future of Decentralized Water Security

The WaterLens project is a testament to the power of democratized technology. As the WHO and UNICEF report highlights, the infrastructure for safely managed water is not expanding fast enough to meet global needs. Decentralized, AI-powered tools provide a "stop-gap" that empowers individuals and small communities to take control of their own health.

By moving the intelligence to the edge, we reduce the cost of safety. We no longer need a million-dollar lab in every village; we need a $50 microcontroller, a suite of reliable sensors, and the shared knowledge of the global engineering community.

Whether you are an engineering student building your first toolkit or a seasoned developer hacking hardware for environmental good, the tools to solve the water quality gap are more accessible than ever. The integration of AI and hardware isn't just about making "smarter" gadgets; it's about solving the fundamental challenges of human survival.