1 of 28

�CS60055: Ubiquitous Computing 

Energy Harvesting

INDIAN INSTITUTE OF TECHNOLOGY

KHARAGPUR

Sandip Chakraborty

sandipc@cse.iitkgp.ac.in

Department of Computer Science and Engineering

2 of 28

Low Power Autonomous Devices

  • Connected low-powered devices are the backbone behind majority of the ubiquitous systems
    • Sensing and actuation (Conventional IoT devices)
    • Supports on-board ML for inferences (TinyML)
    • Edge devices with GPU support (Nvidia JetSon)

  • Powered by on-device power supply
    • Currently realized by various types and sizes of batteries depending on the application requirements

  • Power sources become one of the major performance bottleneck
    • batteries store a finite amount of energy – needs recharging or replacement

Indian Institute of Technology Kharagpur

3 of 28

Low Power Autonomous Devices

  • Various applications demand constant power sources which are unlikely to be available
    • Example: Biodiversity monitoring (animals in a forest, fishes in the sea)
  • Given the limited lifespan of batteries, we need alternative power sources
  • Power requirements for typical IoT devices:

Indian Institute of Technology Kharagpur

4 of 28

Power Autonomy Challenges

  • Use non-conventional energy sources (along with harvesting)
    • Example: Solar energy
  • Energy harvesting as a potential energy source
    • Thermal energy harvesting
    • Solar energy harvesting
    • Vibration (piezoelectric) energy harvesting
    • Kinetic (mechanical) energy harvesting
    • Inductive (electromagnetic induction) energy harvesting
    • Ambient Light Energy Harvesting (Indoor Light Harvesting)
    • Electrostatic Energy Harvesting
    • Magnetic Field Energy Harvesting
    • RF (Radio Frequency) Energy Harvesting
    • ...

Indian Institute of Technology Kharagpur

5 of 28

Radio Frequency Energy Harvesting

  • Captures ambient radio frequency energy and converts it into usable electrical power

  • Use the available RF devices for energy harvesting
    • GSM Antennas
    • Radios
    • WiFi devices
    • TV Broacast

  • Typically used in a scenario where many RF sources are available in the area
    • Advantage: "Free" energy source
    • Challenges: Low power density and variable/unpredictable power

Indian Institute of Technology Kharagpur

6 of 28

How RF Energy Harvesting Works

  • Antenna: Responsible for capturing RF signals from the surrounding environment
    • Typically captures a specific predefined frequency (say, 5GHz WiFi)
    • Multiband antennas can capture signals from multiple frequency bands

  • Matching Network: Connects the antenna to the rectifier and maximizes power transfer
    • Adjusts impedance to ensure that the maximum possible RF power is delivered from the antenna to the rectifier

  • Rectifier (AC-to-DC Conversion): Converts the captured RF signal into DC power
    • RF energy is AC, while most electronic devices require DC

Indian Institute of Technology Kharagpur

7 of 28

How RF Energy Harvesting Works

  • Energy Storage/Management Circuit: The harvested energy is typically stored in a small capacitor, rechargeable battery, or supercapacitor to provide stable power to a device
    • Regulates the power output and ensures stable voltage for the device.

  • Loads (Device): The device or circuit that uses the harvested energy

Indian Institute of Technology Kharagpur

8 of 28

Energy Harvesting Network

Indian Institute of Technology Kharagpur

9 of 28

Principle of RF Energy Harvesting

  • The power received by an RF energy harvesting antenna depends on several factors
    • Distance from the RF source
    • frequency of the signal
  • Friis Transmission Equation provides the power received (Pr) by an antenna from an RF source:

    • Pt is the transmit power of the RF source in watts
    • Gt is the gain of the transmitting antenna
    • Gr is the gain of the receiver antenna
    • λ is the wavelength, and d is the distance between the transmitting and the receiving antennas

Indian Institute of Technology Kharagpur

10 of 28

Principle of RF Energy Harvesting

  • The antenna efficiency (η) affects the power delivered to the rectifier
  • The Effective Power Captured by the antenna (Pant) is:

    • In practice, an impedance matching network is used to maximize power transfer from the antenna to the rectifier
    • If the antenna and rectifier are perfectly matched, the maximum power is transferred; otherwise, power is reflected, reducing efficiency.

  • The rectifier converts the AC signal received from the antenna into DC power. The DC Power Output after rectification is computed as,

ηr is the rectifier efficiency

Indian Institute of Technology Kharagpur

11 of 28

Principle of RF Energy Harvesting

  • To calculate the power available to the load or device, we consider both the rectifier’s output and any energy stored in capacitors or supercapacitors

  • Let C be the capacitance of the storage capacitor, and V be the output voltage across it. If the capacitor is charged to a voltage V over time t, the stored energy (E) is:

  • If the capacitor discharges to power a load, the average power delivered (Pout) can be computed as,

Indian Institute of Technology Kharagpur

12 of 28

Battery-free Wireless Sensing

  • Use the wireless signal for joint sensing and energy harvesting
    • The received signal can be used for analyzing the environment as well as the DC component of the signal can be used for energy harvesting

Indian Institute of Technology Kharagpur

13 of 28

Battery-free Wireless Sensing

  • Use the wireless signal for joint sensing and energy harvesting
    • The received signal can be used for analyzing the environment as well as the DC component of the signal can be used for energy harvesting

ACM MobiHoc 2024

Indian Institute of Technology Kharagpur

14 of 28

Challenges with WiFi Sensing

  • Not all WiFi cards support CSI extraction
    • Most of the CSI-extractable NIC cards are no longer adopted by the latest Wi-Fi devices
    • Needs additional hardware on the WiFi devices for CSI extraction

  • WiFi NICs consume a good amount of power

Indian Institute of Technology Kharagpur

15 of 28

Challenges with WiFi Sensing

  • Not all WiFi cards support CSI extraction
    • Most of the CSI-extractable NIC cards are no longer adopted by the latest Wi-Fi devices
    • Needs additional hardware on the WiFi devices for CSI extraction

  • WiFi NICs consume a good amount of power

Use an energy harvester to power the NIC and sensing can be performed by extracting CSI information from the same RF signal

Indian Institute of Technology Kharagpur

16 of 28

Core Idea

  • Human activity significantly affects the amount of energy that the harvester captures from the ambient RF signals
    • Depending on the position and the pose of the human, the signal reflecting off the human body may superimpose constructively or deconstructively with multipath signals in the propagation environment at the harvester
    • Leverage the RF energy harvester as a novel battery-free activity sensor

  • Measure the time-varying harvested power (voltage) as the sensing signal
    • Apply a deep neural network to recognize various human activities

Indian Institute of Technology Kharagpur

17 of 28

Fresnel Zone Diffraction Model

  • Divides the area around the direct LoS path between a transmitter and a receiver into a series of ellipsoidal zones called Fresnel zones
    • Each zone represents an area where electromagnetic waves travel with different path lengths, resulting in constructive or destructive interference based on the phase differences at the receiver

  • Describes how obstacles in the �Fresnel zones affect signal �transmission
    • Helps predict the signal strength �and attenuation caused by diffraction, �which occurs when an electromagnetic �wave encounters an obstacle along its �propagation path

Indian Institute of Technology Kharagpur

18 of 28

Fresnel Zone Diffraction Model

  • First Fresnel Zone (FFZ): Contains most of the transmitted energy
    • Keeping the first Fresnel zone clear of obstructions minimizes diffraction losses and allows for a stronger received signal

  • Higher-order Fresnel Zones: These alternate between causing constructive (odd-numbered zones) and destructive interference (even-numbered zones)
    • Signals passing through these zones can interfere with each other at the receiver, depending on the path differences

Indian Institute of Technology Kharagpur

19 of 28

How Fresnel Zones are Defined

Zone

Path Difference Range

Relative Phase

1st Fresnel zone

Between 0 and λ/2

0° → 180° (mostly in phase)

2nd Fresnel zone

Between λ/2 and λ

~180° out of phase

3rd Fresnel zone

Between λ and 1.5λ

~in phase again

4th Fresnel zone

Between 1.5λ and 2λ

~out of phase again

Indian Institute of Technology Kharagpur

20 of 28

Fresnel Zone Diffraction Model

  • The radius of the first Fresnel zone at any point along the path between the transmitter and receiver can be calculated using

Indian Institute of Technology Kharagpur

21 of 28

Fresnel Zone Diffraction Model

  • To estimate signal attenuation due to diffraction, the model uses a parameter called the Fresnel-Kirchhoff diffraction parameter (ν), which quantifies the impact of an obstacle on the signal:

    • h is the height of the obstacle in meters

Indian Institute of Technology Kharagpur

22 of 28

Fresnel Zone Diffraction Model

  • To estimate signal attenuation due to diffraction, the model uses a parameter called the Fresnel-Kirchhoff diffraction parameter (ν), which quantifies the impact of an obstacle on the signal:

    • h is the height of the obstacle in meters
  • If 𝜈 <0.7: Minor obstruction, minimal diffraction loss.
  • If 𝜈≈1: Moderate obstruction, potential for significant diffraction loss.
  • If 𝜈>1: Severe obstruction, high diffraction loss, substantial signal attenuation.

Indian Institute of Technology Kharagpur

23 of 28

Impact of Human Subjects on the FFZ

  • Human subject on the FFZ will create signal diffraction
    • The diffracted signals from the human body will attenuate with the Friis gain at the RF harvester

Indian Institute of Technology Kharagpur

24 of 28

Impact of Human Subjects on the FFZ

  • The energy generated by the RF energy harvesting device is the combination of three factors
    • Friis Gain (Free-space propagation)
    • Diffraction Gain (Constructive attenuation from the FFZ)
    • Path loss from the other objects (Constructive/destructive attenuation from other objects in the surroundings)

Indian Institute of Technology Kharagpur

25 of 28

Harvested Voltage Signals for Various Activity Recognition Tasks

Indian Institute of Technology Kharagpur

26 of 28

Overview of REHSense

Indian Institute of Technology Kharagpur

27 of 28

Battery Free Sensing System

Indian Institute of Technology Kharagpur

28 of 28

Happy Learning!

Some resources related to this topic

Introduction

Related Work

Background

Observation

Methodology

Evaluation

Conclusion