So far, arrangements have been tested in rats and later versions are coming out for dogs, Lavela says. In a demo video that the WIRED can see exclusively, the scientist uses a cane to capture air samples in four different Petri dishes, each with different odors. The cane sends the smell to a pod with a mouse mounted on the nose -computer interface through a tube through a warning sound. A few seconds after the animal smells, the fragrance information is sent to the phone on the pod. The mobile app displays the name of the compound with the smell of animal and the quality score that takes into account the accuracy and concentration of molecules.
Currently, Canaery’s rat prototypes can detect not only metampletamine, cocaine and pentanil, but also fires accelerators and lax powder used for ammunition.
In mammals, the nose and the brain work together to detect the odor. When the odor molecule enters the nose hole, it is binded to the olfactory receptor. Humans have about 450 kinds of olfactory receptors and twice as many dogs. All odors stimulate other combinations of receptor types to produce unique electrical signals. This signal is transmitted to the olfactory bulb to be processed. LAVELLA compares the olfactory bulb surface to the board. When it smells, the rectangle is turned on with a specific pattern on the checkerboard.
Canaery uses AI software to recognize these patterns and associate with odors. After transplanting the array, scientists expose animals to the smell to train the AI model. LAVELLA says this software can be trained in about three sessions. During this session, scientists offer more than 24 samples of the same odor as animals. Later, the animal is exposed to odor to verify the AI model.
The current array, which is transplanted into a demo rat, has 128 electrodes that capture the nerve signal from the olfactory bulb. Researchers at Lawrence Livermore National Laboratory are studying new arrangements with 767 electrodes to capture more information. Lavela said, “The next -generation device can have greater performance in the field for complex background odors and chaotic steams in the air.
The smell of decoding is not a new effort anyway. Researchers have been studying “e-NOSE” technology to detect smells for the past 40 years. The device uses a chemical sensor to convert the odor molecule to an electrical signal, then analyzes it by the pattern recognition system to identify the odor sauce. But these devices have historically detected small smells.
Joel Mainland, a non -profit researcher at Philadelphia’s non -profit research institute, said, “Animal is a wise way to solve the problem because animals can do something that can not get the current sensor.
