A team of researchers from the robotics lab at Bar Ilan University in Israel is developing a system that will monitor patients under mechanical ventilation and alert if anything goes wrong.
The technology is based on a defense system that detects faults in Unmanned Aerial Vehicles. The work initially began with participation of the Ministry of Defense in order to assist in monitoring problems that affect UAVs, such as ice on the wings or blockage of the sensors with ice.
Lead researcher and head of the robotics lab at Bar Ilan, Professor Gal Kaminka, told Israeli newspaper Haaretz: “There are no simple malfunctions. There are many types of failures that require a human being present to look at the dashboard and see, for example, that the engine is strained or the speed and altitude are decreasing. The Ministry of Defense contacted us and asked if there is a way to create detectors that can automatically alert in case something is not functioning as it should.”
Following the connections between sensors
Dr. Kaminka and his team developed an algorithm that can recognize the connections between various sensors and know that in some heights it is expected that the temperature will drop, or that if the drone changes direction the compass should turn. In case of malfunction, some sensors continue to act normally while others fail – and the ties between them break. This new method identifies when these connections break and can alert that something unusual is going on.
“Then, Dr. Haim Berkenstadt who is an anesthesiologist from Sheba Hospital, approached us,” tells Dr. Kaminka. “Apparently one of the problems that anesthetists have is to monitor the patient during surgery, recovery room or intensive care and understand that something is unusual.”
Kaminka adds that “monitoring provides a record of what is happening with the patient in terms of heart rate and blood oxygen levels, but cannot say that something is wrong. The application uses the same method and begins to recognize when there is an exception and gives the doctor or nurse a warning. It’s not that there are no systems that try to provide an answer to the problem today, but they are not reliable.”
Robotics research became a medical study
This brought Kaminka’s team to partner with Dr. Berkenstadt and Dr. Meir Kalech from the Department of Information Systems Engineering at Ben Gurion University, to turn the system into a device that can assist anesthesiologist. “We developed the method to identify exceptions in medicine, and because there’s a commercial potential we made sure to register it as a patent,” says Kaminka.
“What began as a robotics research is becoming a study in medicine and what started as a technique for Unmanned Aerial Vehicles is now tested on humans. The transition was not trivial. It is very nice when in science you start with a solution for one problem and suddenly you have a direction that is completely different from what you started with.”