Danish scientists build a radar that tells bee and wasp species apart

Identifying pollinator species has traditionally relied on a destructive and expensive method: researchers catch the insects, kill them and examine them in detail under a microscope. The approach delivers accurate identification, but it harms populations. European researchers have now found a way to tell bees, wasps and other pollinators apart without hurting them — using radar.

How the radar recognizes insects

Scientists from the Technological University of Denmark (DTU) and Trinity College Dublin built a system based on millimeter-wave radio signals. The core idea is simple: every insect species beats its wings in its own way, and those movements produce unique micro-Doppler signatures — distinctive patterns in the radar return. Think of it as a fingerprint, but for wings.

For a long time, conventional radar was only used to track large swarms of migrating insects at high altitude, such as locusts or butterflies on migration. But a single small insect flying low over a flower produces an extremely weak signal. Adam Narbudowicz, who led the research at DTU, explains that it was impossible to detect such a faint signal by looking at data from a single moment in time.

The solution was signal integration: instead of analyzing a single instant, the system accumulates and processes data over a longer period and extracts enough information from it to make an identification. The team focused on how an insect's wingbeats generate micro-Doppler signatures — subtle changes in the reflected radar signal caused by microscopic movements.

"When we look at the raw signals, it's hard to pick up all the fine details.

But with machine learning, we can tell them apart," — Adam Narbudowicz.

Testing and results

The scientists trained a machine learning model on five pollinator species: honeybees, ground bees and several types of wasps. The experiment took place on the Trinity College Dublin campus. Each insect was placed in a small plastic cylinder above a millimeter-wave antenna, its radar signature was recorded, and it was then released unharmed.

The model analyzed more than 70 different features of each insect's radar return, including wingbeat frequency, the rate of change in motion and signal amplitude.

The results are impressive:

• 85% accuracy in identifying a specific insect species
• 96% accuracy in broader classification — distinguishing the bee group from wasps
• 70+ features analyzed in each radar return
• Accuracy improves from 75% with 0.1 seconds of observation to 84% with 1 second

Practical applications

The researchers propose building trap-like devices that insects would fly into naturally; the system would analyze them in flight, and they would then fly out unharmed. That opens up a range of practical uses.

Monitoring pollinator populations is critical for agriculture — bees pollinate about a third of the food we eat. The system could also track crop pests and detect invasive species before they spread.

The radio waves the system uses are completely safe — their power is far below any potentially harmful level. That is a sharp contrast with traditional traps, which often use cyanide or other poisons.

What it means

Moving from killing insects to non-invasive monitoring is a major step for entomology and biodiversity conservation. The scientists' next goal is to develop a portable version for field use and to build a global database of radar signatures for all known pollinators.

Such a database would allow an insect to be identified instantly from its flight pattern. By adding environmental data, researchers could track not only species composition but also behavioral changes — for example, anomalous patterns in wingbeat frequency that signal population stress or disease.