Backscatter brainwave could make IoT comms even more energy efficient
How does sub-0.6 mW sound?
Boffins in South Korea claim to have developed an energy-efficient system for low-power Internet of Things (IoT) applications that uses "backscattering" to harvest energy from a wireless signal for its communications.
Researchers at Pusan National University say they developed a backscatter communication system that is 40 percent more efficient than conventional methods, and could be useful in devices such as IoT sensors for applications in smart homes and wearables.
A paper describing their findings was published in Volume 11, Issue 12 of the IEEE Internet of Things Journal on June 15.
Backscatter communication refers to a system that uses energy from a transmitted wireless signal to modulate and reflect signals back to their origin. It has been likened to the way RFID (radio frequency identification) tags work, as well as contactless payment cards, which use energy from the reader device to function.
The technology itself is not new, and The Register has covered it in the past.
According to the researchers, modulation schemes such as Quadrature Amplitude Modulation (QAM) are typically chosen for this to achieve relatively high data rates and low bit error rates.
But the optimal modulation method depends on the reflection coefficient, which defines the reflected wave with respect to the incident wave. Discrepancies between simulations and real-world measurements make it difficult to accurately predict the optimal reflection coefficient for the job.
The Pusan team says they got around this problem by turning to AI. Specifically, they say they used transfer learning, a machine learning technique where a model trained on one task is repurposed for a related job.
This involved pre-training an artificial neural network (ANN) using simulated input bias voltages to train it on load modulator behavior across varying voltage conditions. The learning gleaned from this step was then used in a main training step, where the ANN was trained using experimental data to predict the reflection coefficients.
The results allowed them to select optimal 4 and 16-QAM schemes by aligning predicted reflection coefficients with specific points in the QAM constellation. This allows for energy-efficient data transmission, below 0.6 mW, which they claim is lower than conventional wireless systems.
- Imagination licenses RISC-V CPU cores for smart TVs, IoT, embedded stuff
- Here's another thing AI can do: Spark a boom in edge infrastructure spending
- South Korea goes out on a limb to manage forests with AI, satellites
- The S in IoT stands for security. You'll never secure all the Things
Also part of the solution is a 2 × 2 × 2 MIMO transceiver system with two transmit and two receive antennas for different polarizations that enhances signal reception.
The researchers tested the resulting system in the 5.725-5.875 GHz C-band of the Industrial, Scientific, and Medical band, and claim that it achieved a spectral efficiency of 2.0 bps/Hz using 4-QAM modulation, showing effective bandwidth utilization.
According to the Pusan team, this lays the groundwork for a reliable and efficient backscatter system for multiple applications, including consumer electronics, healthcare monitoring, smart infrastructure for urban management, and environmental sensing.
Gartner Distinguished VP Analyst Bill Ray, Chief of Research for Emerging Technology and Trends, expressed skepticism.
"There's nothing wrong with backscatter communications – it builds on the RFID concept. Passive RFID tags reflect a radio signal, and modulate that reflection to incorporate a serial number. Backscatter takes that a step further by allowing random data to be encoded in the reflection," Ray told us.
But "the technology is already highly efficient, so improving it is of limited value," he claimed, adding: "The challenge is competing with highly efficient Bluetooth (or LoRa, or even 802.15.4) that run from low-cost, rechargeable, batteries."
The technology could find a niche in IoT applications where the devices are too small to easily fit a battery, or in environments where retrieving the device for this purpose is not advisable. ®