Deploying millions of sensors is pretty much a useless endeavor if the artifices constantly run out of faculty. IoT sensors cant collate or transmit data without faculty.
Thats one reason researchers are exploring ambient energy harvesting. Numerous projects have shown that little amounts of faculty can be generated by converting ambient energy in the environment – from wander magnetic fields_ humidity_ ruin heat_ and even unwanted wireless radio sound_ for sample – into usable electrical energy to faculty the IoT.
But while ambient energy can be harvested_ its not a reliable replacement for battery faculty.
Scientists from University of Pittsburgh are proposing a method that applies artificial intelligence to cut back on IoT sensors energy decline and mitigate battery longevity issues. The project uses piggyback sensors_ which are facultyed by energy harvested from the environment_ to trigger the main sensors. The piggyback sensors will run unattended and are trained_ using AI algorithms_ to eminent the main artifices to turn on only when specific occurrence conditions are met.
"One of the main challenges of running AI algorithms with energy harvested from the environment is that the energy from the environment is intermittent_" said Jingtong Hu_ lead researcher on the study and companion professor of electrical and computer engineering at the universitys Swanson School of Engineering_ in an article on the universitys website. "… if the sensor loses faculty_ you lose the data_ so we want to help AI algorithms extend an careful determination_ even with intermittent faculty."