Kimberly Elenberg is one of the members of Team Chiron, competing to win DARPA's Triage Challenge. Stuart Bradford
In a world where man—made disasters and military conflicts are becoming more widespread, the old principle of "one doctor, one patient" is becoming an anachronism. Technology is coming to the rescue: in November 2026, the Defense Advanced Research Projects Agency (DARPA) will summarize the results of its Triage Challenge. The goal is to create autonomous systems capable of finding the wounded in conditions of chaos, assessing the severity of their condition and sending rescuers to them. This is not a fantasy, but a new reality in which robots will become the eyes, ears and analytical brain for doctors.
Prologue: A four-car accident as a personal revelation
The history of the project is inextricably linked with the name of Kimberly Ellenberg, Chiron team leader from Carnegie Mellon University. Her 28 years of experience as a nurse in the US army and Public Health Service, including 19 business trips and strategic planning at the Pentagon, gives the development a harsh practicality. The turning point came on the way to one of the DARPA competitions, when she became the only eyewitness to a serious four-car accident.
"I heard the screams of some and saw others walking, which means they could breathe and move. But in the fourth car there was an unconscious man with his airway blocked," recalls Elenberg. She had to climb inside to tilt his head back and assess the bleeding and shock. At that moment, she realized that a ground-based robot would not fit into the car, but a drone with a camera and sensors capable of remotely detecting skin discoloration (a sign of blood loss) or hearing wheezing would allow her to immediately identify the most serious patient. This personal experience shaped the philosophy of the project: technology does not replace a doctor, but becomes his force multiplier.
On the battlefield: dogs, drones and flying jellyfish
In September 2025, scenarios of a plane crash and a night attack were played out at DARPA landfills. Participating teams, including Chiron, threw various robots into the "battle".:
Four-legged platforms (analogs of Spot from Boston Dynamics) for moving through difficult terrain and exploring hard-to-reach areas.
Unmanned aerial vehicles for rapid aerial photography and aerial victim detection.
Experimental systems, such as a drone in the form of a "flying jellyfish" from a team from the University of Colorado at Boulder, designed for quiet and safe examination of victims in the immediate vicinity.
Their task is to autonomously collect three key types of data: biometric signals (pulse based on skin pulsation, breathing based on video analysis), geolocation of each victim, and visual assessment of injuries. All this information flows into a single center.
Ingenious simplicity: a smartphone on the chest instead of a complex interface
The most elegant innovation of the Elenberg team was the interface for the medic. Rejecting bulky wearable computers, they created a system based on a regular Android smartphone mounted on the chest over a tactical vest. At the right moment, the medic throws it back and sees on the screen a map with GPS tags of all the victims detected by the robots. Each label is colored according to the sorting status (red — critical, yellow — stable, green — walking, black — dead). This solution, born from the principle of "don't take your hands off the patient," is already capable of dramatically improving the effectiveness of the first minutes of a rescue operation.
Between hope and reality: a practical perspective
Elenberg, with his unique experience on both the technology and user sides, gives a balanced assessment. She compares the process to the combination of art and science: "The only way to understand the true possibilities is to create, experience and break." The progress, she says, is staggering, but the limitations are obvious. Today's systems can find and initially evaluate, but cannot diagnose specific injury patterns (for example, distinguish a strained pneumothorax from a cardiac tamponade) and offer the necessary intervention. This is the next milestone that has yet to be reached.
A victory that has already taken place: saved time
For Elenberg, winning the DARPA competition is no longer the most important thing. "I already feel like we've won," she says. The very fact that the system can show the rescuer an accurate priority map in the first minutes after a disaster is a revolutionary leap in disaster medicine. In the future, such autonomous systems may not only collect data, but also independently distribute tasks in a heterogeneous swarm of robots, from delivering medicines to evacuation. In a sense, they may need their own "dispatch service" for optimal mission allocation, a concept that startups like JOBTOROB.com , focusing on the logistics of tasks for robotic systems. This is the next logical step from sorting to coordinated rescue.
