Demo: Bi-static Scatter Underwater Acoustic Communications

Scatter radio, i.e., communication by means of reflection rather than radiation, although not a new idea, has recently been used for underwater sensor network applications [2].

Because communication can be achieved with a single transistor front-end, scatter underwater acoustic communications can minimize both energy requirements and monetary cost of each sensor. This allows for widespread collection and analysis of data from Internet of Underwater Things (IoUT) devices, on a large scale, and enable a plethora of applications. These range from marine conservation, coastal resilience, aquaculture, offshore energy, defense, among others [1].

Extended field coverage is a key challenge i.e., the communication range between the sensor tags and the receiver has to be maximized. Achieved ranges are inherently limited due to the following: 1) passive communications, since tags rectify a continuous wave (CW) signal, transmitted by the emitter, and the achieved range is limited by the so called "power-up link"; 2) high bit rates, thus small bit duration lead to reduced energy-per-bit and signal-to-noise ratio; 3) monostatic architecture i.e., the carrier emitter and receiver are co-located in the same box, which means that these systems suffer from round-trip path loss.

Our demo shows a passive sensor/tag in a bistatic topology (i.e., carrier emitter and receiver placed in different locations) to communicate sensor data by means of acoustic reflections (see Fig. 1). The emitter illuminates the tag with a carrier continuous wave at f_c = 35.9 kHz. To modulate sensor information, the tag terminates a piezoceramic transducer between two loads. Ideally, when the transistor is switched on, the transducer is short-circuited and the incident wave is scattered back with a negative reflection coefficient (phase change π). Respectively, when the transistor is switched off, the transducer is open-circuited and the incident wave is scattered intact (i.e., no phase change). Modulation on the tag is achieved by switching the transistor at two rates i.e., f₁ − f_c and f₀ − f_c, thus changing between two reflection coefficients. To ensure orthogonality for noncoherent frequency shift keying (FSK), the spacing between the two subcarrier frequencies is |f₁ − f₀| ≥ 1/2T, where f₀ = 40 kHz, f₁ = 44.5 kHz and T = 10 ms is the i-th bit duration. At the receiver we correlate against frequencies f₀ and f₁ for signal demodulation (see Fig. 2).

The efficiency of energy transfer was measured by the percentage of current that could be collected at the sensor tag. This changed with three main variables that required careful tuning: 1) the potting material used for the piezoceramic transducer which affected the selection of the frequency for the carrier wave at the emitter, 2) the subcarrier frequencies and 3) the design of the electrical matching circuitry. The most pertinent variable was the potting material. We found out that hard plastic materials perform better than soft silicon and rubber-based plastics. We observed that application of a thin layer of UV-resin resulted in better frequency response compared to thicker tabletop epoxy or polyurethane mold-based materials. This was even the case when the thickness of the epoxy/polyurethane was the optimal (as indicated in the literature) one fourth of the signal wavelength. As soon as UV-resin demonstrated good power transfer results, we measured the new resonance frequency of the piezo in water and used it to match the piezo's electrical impedance. We discovered that the nominal resonance frequency of the piezo measured in air is not nearly as effective as the second harmonic that appears at twice its resonance frequency.

Specifically, we observed that the tag worked better around the 2nd harmonic when tested in a denser environment where it experiences a higher loading from the water. Subcarrier frequencies f₀ and f₁ were selected based on the impedance measurements and the minimum frequency separation criterion for non-coherent FSK. We avoided selecting f₀ and f₁ to be multiples of each other since both would get illuminated at the same time by the incident carrier wave. The final challenge was impedance matching at the transmitter side. High power inductors are not as readily available (as their 0.5 W versions), thus we had a limitation on the amount of maximum power that we were capable of transmitting. This resulted in limited communication range for our system.

Our demo will be set up in a small 10 gallon fish tank, see Fig. 1. The bistatic underwater acoustic backscatter communication system consists of a carrier emitter, a sensor tag and a software-defined underwater acoustic receiver. The battery-free tag prototype consists of a piezoelectric transducer, electric impedance matching, voltage multiplier circuitry, and a low-power microcontroller that can interface to a sensor unit. The demonstration will begin with the carrier wave being amplified and transmitted through the transmitter piezo. The power of the carrier wave will be controlled by the tunable voltage supply supporting the transmitter. The sensor tag will be connected to a breadboard with the energy harvesting, and impedance matching circuitry. As the sensor tag begins to power up, voltage will be measured and displayed in real time. When the sensor tag has collected sufficient voltage, it will audibly reflect the carrier wave and send its (pre-set) data vector to the receiver. Voltage will visibly drop at the same instant the audible reflection will take place. The receiver will be connected to an oscilloscope, which will record the transmission of data as well as display the FFT spectrum at the receiver in real-time. The received data will then be loaded to MATLAB for decoding and display.