Doctors have long been using sound to diagnose illness. The tics, gurgles, wheezes and clunks in patients can be an excellent indication of an existing or impending medical issue. A heart murmur, for example sounds a bit like water running through a garden hose, suggesting that disease has roughened the heart valves or arteries that should be investigated.
The practice gained a considerable boost when in 1816, when René Laënnec faced the awkward problem of how to listen to the chest of a young woman who had presented with the symptoms of tuberculosis. To avoid pressing his ear to the curvaceous breast, the coy doctor rolled up pieces of paper. Curiously, history even records that he used 24 sheets.
Laënnec was delighted because as well as saving embarrassment, he was better able to hear the telltale sounds within her chest. After several years of development, we had the first stethoscope, which is derived from the Greek "I see" and "chest". His stethoscope, which he initially called Le Cylindre, was simply a wooden tube. It later it evolved into the iconic device worn by doctors today.
Sadly, Laënnec was to later die of the very disease he was treating - tuberculosis.
Today there are electronic stethoscopes, and medical researchers have created software that can classify lung sounds into five common diagnostic categories.
Then, as any car owner will attest, it's not just doctors who use sound to find faults. You know that sinking feeling when you notice the moans from the gearbox or wheel bearing. A skilled mechanic can often diagnose the faults from those aberrant noises. They also know what sort of noises to expect when they hand you the bill.
Traditionally, a doctor or a mechanic will use their experience to interpret sounds, but increasingly this is becoming possible with computers using complex mathematics. It can be a difficult problem because, in a noisy environment it requires picking out meaningful information out of the background clutter.
The idea of using acoustics to detect faults is being extended to many other fields. The Australian company, Ping Services, is installing 55 devices on turbines at a wind farm in Victoria in a pilot program. The Ping Monitor uses acoustic analysis, machine learning and the Internet of Things to report possible wind turbine blade damage.
It uses an algorithm to rate the health of the turbine based on its acoustic signature, monitoring changes over time. The device is attached using magnets to towers and actively listens to the blades. It can detect blade faults such as pitting or cracks caused by lightning strikes or hail.
Ping Services have partnered with Australian space startup Myriota, a satellite communications company, to transmit data from remote areas.
Response by: Rod Taylor
- The Fuzzy Logic Science Show is 1100am Sundays on 2xx 98.3FM. Send your questions to AskFuzzy@Zoho.com Twitter @FuzzyLogicSci Podcast FuzzyLogicOn2xx.Podbean.com
