Piezo Harvesting Powers Underwater Camera and Acoustic Backscatter Link

What you’ll learn:

How a battery-free, untethered underwater camera gets its operating power via wave-energy harvesting.
How the camera captures images with minimal power use.
How the camera communicates captured image data back to the control station with net-zero power.

Powering autonomous, untethered vehicles is always a challenge, and doing so for underwater situations adds to this challenge. A multi-departmental team of MIT researchers has taken a tangible step to overcome this problem by developing a battery-free underwater camera, powered by energy harvesting, which also implements a wireless data link.

Their approach exploits that “other” wireless energy carrier, namely acoustic energy, by harvesting it via on-board piezoelectric transducers while using the same energy and transducers for the bidirectional data link (Fig. 1). They use ambient wave energy from natural water action, ship’s wakes, and other disturbances.

1. (a) A remote acoustic projector (top right) transmits sound on the downlink. The acoustic energy is harvested by a piezoelectric transducer and converted to electrical energy that powers up the battery-less backscatter sensor node. The energy accumulates in a supercapacitor that powers up an FPGA unit, a monochromatic CMOS sensor that captures an image, and three LEDs that enable RGB active illumination. The captured image is communicated via acoustic backscatter modulation on the uplink, and a remote hydrophone measures the reflection patterns to reconstruct the transmitted image. (b) The battery-less sensor is shown in an experimental trial where it’s used to image an underwater object with active illumination that enables capturing color images. (c) The plot shows the voltage in the supercapacitor, which is harvested from acoustic energy and varies over time as a function of the power consumption of different processing stages. (d) The spectrogram shows the frequency response of the signal received by the hydrophone over time, demonstrating its ability to capture reflection patterns due to backscatter modulation and decode them into binary to recover the transmitted image.

The device takes color photos even in dark underwater environments. The camera powers up from harvested acoustic energy, captures color images using ultra-low-power active illumination and a monochrome image sensor, and communicates wirelessly at net-zero power via acoustic backscatter.

When the stored energy reaches a minimum required threshold, it autonomously activates a power-management unit to regulate the voltage and supply it to an on-board processing and memory unit (Fig. 2). The processing unit and oscillator trigger an ultra-low-power monochromatic CMOS camera and on-board active illumination to capture the image of an underwater object. The entire imaging process is powered by the harvested energy stored in the supercapacitor.

2. Schematic of the hardware design: The harvester node at the bottom is connected to a multi-stage rectifier followed by a supercapacitor, which stores the harvested energy. The supercapacitor voltage is fed to a 2.8-V LDO and to a 1.4-V dc-dc step-down converter. The output of the dc-dc converter is used to power the FPGA core, and the output of the LDO power the FPGA banks. The FPGA also is connected to two external clocks (32 kHz and 4 MHz) as well as the camera via several GPIO pins (pixel clock, line valid, data, power, master clock). The FPGA controls the operation of the MOSFETs connected to the communication transducer on the top left. This transducer is responsible for sending camera data via backscatter communication.

The project required advances in a technique for ultra-low-power underwater communication, as harvested power from the acoustic source typically ranges from a few tens to hundreds of microwatts. Meanwhile, state-of-the-art low-power underwater communication modems require 50 to 100 mW to communicate over tens of meters.

Piezoacoustic Backscatter

To overcome this gap, they leveraged piezoacoustic backscatter to communicate the captured image on the uplink, extending a recently developed net-zero power communication technology to enable telemetry of imaging data.

Underwater piezoacoustic backscatter communicates messages by modulating the reflection coefficient of its piezoelectric transducer. The battery-free node encodes pixels into communication packets by switching between different transducer loads (here, inductors). The camera either reflects a wave back to the receiver or changes its mirror into an absorber so that it doesn’t reflect back.

A remote hydrophone next to the transmitter measures the received acoustic signal to sense changes in the reflection patterns due to backscatter and if a signal is reflected back from the camera. If it receives a signal, that’s a “1” bit, and if there’s no signal, that’s a “0” bit. The switching is done by simply controlling two transistors and requires just 24 nW of power.

The reflection patterns are decoded and used to reconstruct the image captured by the remote battery-free cameras. Robust end-to-end communication is achieved by implementing a full networking and communication stack that incorporates underwater channel estimation, packetization, and error detection. Working range is about 40 meters, and they’re working to extend that distance.

The battery-free imaging system incorporates three monochrome LEDs: red, green, and blue. An always-on LED system would obviously consume far too much precious energy. Thus, the ultra-low-power processing and memory unit alternates between activating (flashing) each of these LEDs and capturing monochromatic images with each active illumination cycle (Fig. 3).

3. (a) To recover color images with a monochrome sensor, the camera alternates between activating three LEDs—red, green, and blue. The top figures show the illuminated scene, while the bottom figures reveal the corresponding captured monochromatic images, which are transmitted to a remote receiver. (b) The figure shows the color image output synthesized by the receiver using multi-illumination pixels that are constructed by combining the monochromatic image output for each of the three active illumination LEDs. (c) A side view of the camera prototype demonstrates a larger dome that houses the CMOS image sensor and a smaller dome containing the RGB LEDs for active illumination. The structure is connected to a piezoelectric transducer. (d) The circuit schematic demonstrates how the imaging method operates at net-zero power by harvesting acoustic energy and communicating via backscatter modulation. (e) The plots show the power consumption over time. The power consumption peaks during active imaging and drops when the captured images are being backscattered.

The monochromatic images are acoustically backscattered to the remote receiver. After decoding each of the images, the receiver synthesizes the received packets into multi-illumination pixels by applying them to the RGB channels of a digital pixel array to reconstruct color images.

Energy Harvesting

The active illumination using the LEDs is the most power-consuming aspect of the battery-free imaging system. For acoustic communication rates of 1 kb/s, empirical measurements demonstrate an average power consumption of around 276 μW for active imaging phase. In a demonstration of passive monochromatic imaging where active illumination isn’t used, the camera drops to about 112 μW. For either configuration, the entire energy budget is harvested from underwater acoustics.

Among the many tradeoffs they evaluated “on paper” was use of a single versus two transducers, with one for harvesting and one for backscatter communications. In principle, it’s possible to use a single transducer for both roles, since both transducers are identical. However, doing so would result in less harvested energy because less energy may be harvested in the open-circuit state than in the inductively matched state. Thus, they chose to use two transducers so that both processes can occur simultaneously without either of them reducing the other’s efficiency.

The entire device is packaged in a domed waterproof housing (Fig. 4). They also built a “base station” hydrophone transceiver that’s used to ping, interrogate, and receive data from the batteryless camera—a somewhat easier project.

4. Exploded view of the camera dome and LED dome: (a) The LED dome contains the red (R), green (G), and blue (B) LEDs, and a layer of polyurethane gasket is added to the dome base to make it waterproof. (b) The camera PCB contains the Himax image sensor, supercapacitor for harvesting energy, power-management electronics, and an FPGA for processing and memory. It also contains programming pins to program the FPGA and change camera parameters. The PCB is enclosed in a transparent dome, and the entire structure is tightly screwed to make it waterproof.

In field tests, they captured color images of plastic bottles floating in a New Hampshire pond, and they took such high-quality photos of an African starfish that tiny tubercles along its arms were clearly visible (Fig. 5). The device also was effective at repeatedly imaging an underwater plant in a dark environment over the course of a week to monitor its growth.

5. (a) The figure shows a photo of a prototype deployed in Keyser Pond for monitoring pollution from plastic bottles on the lakebed. (b) The RGB image output obtained from the imaging method while monitoring pollution in Keyser Pond. (c) RGB image output for Protoreaster linckii, demonstrating qualitative success in recovering its color and numerous tubercles along the starfish’s five arms. d) The imaging method was used to monitor the growth of an Aponogeton ulvaceus over a week. The figures show the captured images on different days of the week.

As we’ve seen in other MIT research projects, the associated analysis and documentation are thorough and comprehensive, The basic nine-page paper “Battery-free wireless imaging of underwater environments” is published in in Nature Communications, plus there’s a posted 38-page Supplementary Information file that goes into further details covering all aspects of the concept, design, modeling, operating cycles, evaluation, power budgets, and analysis of each operating phase, unit fabrication, BOM, circuit description, costs, test arrangements, and evaluation.

Going beyond the written collateral page, there’s a brief video (1:30) showing the color image-capture scheme in action, as well as this one with more coverage of field trials: