Researchers Boast of an IoT Breakthrough with New Sustainable Solar Cells and Power Management AI

Researchers from Newcastle University, the Technical University of Munich, and the Swedish University Network (Sunet) have created high-efficiency yet sustainably-built solar cells that, they say, are ideal for the Internet of Things (IoT) — and have demonstrated their pairing with a neural network for improved energy management.

"Our research marks an important step towards making IoT devices more sustainable and energy-efficient," claims principal investigator and project lead Marina Freitag, PhD, of the work. "By combining innovative photovoltaic cells with intelligent energy management techniques, we are paving the way for a multitude of new device implementations that will have far-reaching applications in various industries."

The team's solar cell design uses dye sensitization of a copper electrolyte, using exclusively non-toxic materials which the researchers say are fully sustainable, and provides what they say is an "unprecedented" power conversion efficiency of 38 percent at 1,000 lux light intensity. That's key to the panel's success, as it's designed to be used on indoor devices — where it would be harvesting artificial light, rather than sunlight — as well as in the great outdoors.

"IoT technology, encompassing the growing number of electronic devices connected to the internet, will create energy savings of more than 1.6 petawatt-hours (PWh) each year — equaling the electricity needed to power more than 136.5 million homes," the team claims. "Nonetheless, the devices themselves will raise the worldwide energy demand by 34 terawatt-hours (TWh) by 2030. It is therefore crucial to (i) employ local energy harvesters for continuous power supply, (ii) reduce electronic waste by using sustainable materials and avoiding batteries and (iii) minimize the energy cost of computation and data transfer."

Being able to harvest ambient light efficiently is one thing, but the team has something else to offer too: a long short-term memory (LSTM) neural network model which, they say, can adjust the computational load of IoT sensors in order to provide dynamic energy management — both reducing the absolutely power requirements of the overall network and lowering congestion on the communications infrastructure. To prove it, the team successfully deployed the model and new cells on a test system driven by a FireBeetle board built around an Espressif ESP32 microcontroller.

"Our demonstration of dynamic energy management on light-powered wireless sensors paves the way for a multitude of device implementations," the team claims. "The sensors devices could, pending sufficient energy availability, pre-process sensed data and infer conclusions; therein, the energy required for network communication can be reduced. The synergy of artificial intelligence and ambient light as power source appear is poised to enable the next generation of IoT devices."

The researchers' work has been published under open-access terms in the journal Chemical Science.