Colleen Josephson, Manikanta Kotaru, Keith Winstein, Sachin Katti, Ranveer Chandra
9 min
Agriculture accounts for the majority of global freshwater consumption, yet many farms lack efficient soil moisture sensing due to the high cost and maintenance requirements of traditional in situ equipment. Standard commercial moisture sensors require manual installation of probes, data loggers, and cabling, often making them vulnerable to farm machinery and prohibitively expensive for large-scale deployment. While remote sensing from satellites or aircraft offers a wireless alternative, it currently suffers from low resolution and poor depth accuracy compared to ground-based sensors.
The researchers introduce a hybrid sensing method that combines the precision of in situ measurement with the convenience of wireless remote sensing. The system consists of two parts: a portable, commodity impulse-based radar placed above ground, and a low-cost electronic backscatter tag buried at a known depth. The tag uses a simple square-wave modulation scheme to reflect the radar signals, creating a distinct signal that the radar can easily isolate from background noise and ground clutter. By measuring the Time of Flight (ToF) of the radar signal to the underground tag and back, the system calculates the dielectric constant of the surrounding soil, which serves as a highly accurate proxy for volumetric water content.
The team tested their approach against oven-dried ground-truth soil samples and standard commercial capacitive sensors across various soil types, including sandy clay loam and silt loam. The proposed system achieved a 90th-percentile error of just 3.4%, matching the performance of expensive, state-of-the-art commercial sensors. Furthermore, the backscatter tag consumes minimal power, enabling an estimated battery life of nearly three years, while the system effectively functions at depths relevant to crop root zones.
By decoupling the sensing element from the expensive readout electronics, this approach significantly reduces the cost per measurement point, potentially enabling widespread moisture monitoring across large, multi-acre farms. The authors suggest that this system could be further scaled by mounting the radar on autonomous farm equipment or low-flying drones, paving the way for ubiquitous, data-driven irrigation management.
Despite decades of research confirming the benefits, most farms do not incorporate soil moisture sensing into their irrigation practices. Soil moisture sensing can be broken into two broad approaches, both of which have drawbacks. In situ sensors are installed in the ground, tend to be difficult to deploy and maintain, and have high costs. Remote-sensing based approaches use radars to infer soil moisture from surface reflection properties. While completely wireless, remote sensing suffers from lower resolution and accuracy compared to in situ sensing. We propose a hybrid approach that combines the advantages of both. This paper introduces the idea of using inexpensive in situ backscatter tags with above-ground radars, which enables completely wireless soil moisture measurements with high-accuracy and high-resolution. Our key idea is introducing a simple, power efficient modulation scheme that enables commodity radars to easily detect and range the underground tag. We have benchmarked our approach against oven-based, industry-standard ground-truth measurements and demonstrated that, at a realistic depth and across several types of soil, we achieve a 90th percentile error of 3.4%, which is the same accuracy as state-of-the-art in situ sensors. We also demonstrate that our approach works with similar accuracy at a real farm.
Alex: That makes the known depth part of the measurement, not just installation metadata. But finding a small tag underground sounds difficult. Doesn't everything around it reflect too?
Sam: That's the central engineering problem. Radar calls those unwanted reflections clutter. Dirt, rocks, and water surround the tag, and their reflections can overwhelm it. A stationary reflector doesn't necessarily stand out from its stationary surroundings.
Alex: So what makes this tag recognizable without adding complicated electronics underground?
Sam: Square-wave modulation: the tag repeatedly switches between reflection states. That creates a changing return against a mostly unchanging background. To the radar's processing, the stationary tag appears to have motion, separating its signal from the clutter.
Alex: So the apparent motion is a label for finding the tag, not a claim that anything underground is moving. Can you make the cancellation concrete?
Sam: Imagine successive radar recordings. The background reflection stays the same, while the tag's contribution changes sign. Subtract the recordings, and the unchanged background cancels while the tag signal reinforces itself. The paper develops that principle into processing that locates the tag and measures its travel time.
Alex: Earlier systems also used buried radio tags. What is new beyond choosing radar as the reader?
Sam: Earlier radio-frequency identification approaches used signal strength or the power needed to get a response. A related WiFi system used differences between antenna signals. The authors report that those approaches had limited sensing depth and poorer accuracy. Their contribution combines precise travel-time measurements with a simple tag that commodity radar can detect.
Alex: Those comparisons depend on how accuracy is defined. What did this study use as its reference?
Sam: In the laboratory, it used oven-based measurements. You take a known volume of soil, weigh it, dry it, and weigh it again. That establishes volumetric water content: the fraction of the wet soil's volume occupied by water. They also compared their system with a commercial capacitive sensor, which estimates moisture through electrical properties.
Alex: And did they test more than one convenient soil? The signal has to penetrate the material, so soil composition could matter.
Sam: They tested three agricultural soil types, spanning sandier, silty, and clay-rich loams. The commercial sensor received soil-specific calibration. The researchers gradually added water to initially dry soil, mixing it throughout the container so samples could represent the moisture being measured.
Alex: What did that comparison show, and how deep was the buried tag?
Sam: The moisture tests used a burial depth of thirty centimetres. Against the oven reference, the radar-tag system's average error was about one-point-four percentage points of water content. That was essentially the same as the commercial sensor.
Alex: An average could hide difficult conditions. How did the larger errors compare?
Sam: For both systems, ninety percent of errors were no greater than three-point-four percentage points. The largest radar-tag error occurred in saturated sandy clay loam, where it underestimated moisture. Each plotted result summarized repeated measurements, and radar captures lasted tens of seconds rather than being instantaneous.
Alex: That supports laboratory accuracy. What happened when they left the mixed containers and tried an actual farm?
Sam: At a local organic farm, they compared the buried tag with commercial sensors near the surface and at the tag's depth. The radar measurements tracked their average within two percentage points. But there were no oven-based reference measurements outdoors, so this wasn't the same validation as the laboratory test.
Alex: The field comparison supports feasibility, then, rather than independently establishing the same absolute accuracy. What about the claim that it can reach deeper roots?
Sam: The paper separately demonstrates tag detection deeper underground and models still greater reach. But detecting a tag is not the same as validating moisture accuracy there. The reported moisture evaluation stays at the burial depth we've discussed. Likewise, collecting readings from a low-flying drone is a proposed possibility, not a demonstrated deployment.
Alex: Cable-free also doesn't mean maintenance-free. What powers the tag, and how firm are the economic claims?
Sam: The prototype uses a battery, not harvested radio power. Its lifetime is projected at nearly three years. The low tag price also assumes a production run, rather than representing an established deployment cost. The authors still call for large-scale, long-term farm studies.
Alex: Is there a measurement assumption that those longer deployments would particularly need to watch?
Sam: Burial depth. The authors show that erosion can make readings underestimate moisture because the assumed soil path is wrong. They propose linking tags at fixed separation to measure the soil between them. But that would lose immediate feedback after watering, because water must first reach that buried region.
Alex: Given those boundaries, who should read the full paper, and where should they start?
Sam: Researchers in agricultural sensing, radar, and low-power wireless systems should start with the tag-detection design and the evaluation. Look closely at the oven-reference comparison, then the farm test and erosion analysis. For deployment work, keep projected cost and lifetime separate from demonstrated accuracy. Everyone else can remember this: the buried device doesn't have to measure moisture; it can make the soil's effect on a radio signal measurable.
Alex: That's the distinction to carry into the full read.