PhD Opportunities

Interested in the electric revolution, future of transport or Power Electronics? We're looking for enthusiastic and motivated students to join the EEMG.

Current PhD opportunities

EEMG Brochure (PDF)

Sensor-driven electronics

Originating from research on energy harvesting and wireless power transfer, we are developing Sensor-Driven Electronics to eliminate batteries and energy harvesters in microsensors. To date, this has permitted lifetimes of certain battery-powered sensors to be extended by decades. This research aims to develop battery-free sensor systems, zero-power receivers, and the harvesting of intermittent, ambient energy at the micro-scale.

More information

For more information on sensor-driven electronics, please contact Bernard Stark.

Our ultra-long-life sensor technology is now being developed and commercialised through Sensor Driven Ltd.

Using sensors to power systems

We are developing wireless sensor nodes that, for most of the time, use no mains or battery power, using instead only the signal from a sensor to power the system. An example is bridge vibration detectors that are powered by the accelerometer output signals shown in the figure. An interesting challenge is to use the same signal for several purposes: e.g. for monitoring a strain threshold, for powering this monitoring, and for precise measurement of a strain event if the strain does pass a threshold. The sensor-driven wake-up blocks that we have developed, listen to the signal, use a small proportion of the power in the signal, leave the data content of the signal intact for subsequent sensing, and leave most of the energy in the signal for energy harvesting from this same signal. Sensor-driven technology will therefore extend the life-time and functionality of equipment and environmental monitoring. It will permit monitoring functions to be embedded into structures. It will provide better data from better locations for longer.

Battery-powered versus sensor-driven sensor nodes

Battery-powered versus sensor-driven sensor nodes Threshold detection is used in many industries to monitor a range of physical parameters. A sensor is periodically polled to check if the sensor signal exceeds a safety threshold. Between polling events, these systems are unaware of the sensor signal. Therefore, the shorter that relevant events are, the more frequent the polling needs to be. It follows that the detection of short events have a relatively high-power consumption. Sensor-driven detectors, by contrast, are continuously aware, as they use the sensor signal as their wake-up call and power source. This makes continuous monitoring between events possible without use of a power source. It allows the battery power to be conserved for useful tasks, such as sampling more frequently during events, processing, or communication. We have addressed several problems here: The first is that sensor nodes do not typically have enough energy budget to sense continuously, but must instead sleep much of the time in order to save power. Sleep times are usually set to achieve sufficient time between sensor-maintenance cycles. A sensor-driven approach has made this monitoring continuous. The second problem is that even the sleep power can eat up a high percentage of the total energy budget, and thereby curtail life-time and functionality. The third problem is that some wireless sensor nodes spend much of their energy budget on receive modes or sensing modes, without actually receiving anything useful. For example, this applies to intruder alarms, crack detection, detection of bearing failure, voice detection, or even the infrared receiver on your TV at home during standby. In all of these cases, for most of the time, power-hungry sensing provides nothing useful. In some cases, the our sensor-driven approach has eliminated the sleep monitoring power to zero, and in others to below 10 nanowatts, see reference below.

References

S. E. Adami, G. Yang, C. Zhang, P. Proynov and B. H. Stark, A 10 nW, 10 mV signal detector using a 2 pA standby voltage reference, for always-on sensors and receivers," 2018 IEEE Applied Power Electronics Conference and Exposition (APEC), San Antonio, TX, USA, 2018, pp. 1065-1070. Won an IEEE Best Paper Award.

Sensor-driven wake-up

The simplest form of sensor-driven monitoring can be implemented using the UB20M, one of a range of custom voltage detectors designed at Bristol, originally for energy harvesting (datasheet at www.sensordriven.com, other voltage ratings and thresholds available). A voltage-generating sensor, such as a piezoelectric displacement sensor, provides both an input signal and power to the UB20M detector. As the input signal rises high enough, the output of the detector operates a switch to connect various electronic sub-systems to a battery. This chip uses only a few pW (see its datasheet), which leaves the sensor’s output signal intact for data or power extraction.

Zero-power and near-zero power sensing

This simple circuit shown above provides zero-power monitoring for any voltage-generating sensor, such as piezoelectric accelerometers, electret microphones, infrared diodes, antennas, etc., if the signal amplitudes during a valid even reach around 0.5 V. To monitor passive sensors (e.g. for strain, temperature, and humidity), and those with output voltages well below 0.5 V, enhanced voltage detectors are available, see for example the reference listed above. These consume 1-10 nW, which is generally referred to as near-zero, as it is less than the self-discharge of a modern coin cell, and therefore the impact on battery life is usually negligible.

Nanopower computation and power management 

We are now working on ways of bringing other power drains such as data-capture, computation, and transmission, to within the nW-power budget of a sensor, thereby completely eliminating batteries from sensor nodes. A prototype of this is shown in the figure, where a power management circuit with a quiescent current of a few tens of nW is actively matching its input impedance to a high impedance (80 MΩ) energy harvester with short (10 ms) intermittent output pulses.

Zero-power sensing

Video introduction to zero-power sensing.

Zero-power infrared testing

Video demonstration using a TV

Radio-frequency wake-up

Video demonstration of listening for radio waves without using power.

Research background

This work has been carried out under these projects: 
SPHERE IRC (EPSRC grant)
SAVVIE (EPSRC grant)

This research is a collaboration between the Electrical Energy Management, the μElectronics, and the Digital Health Engineering Research Groups at the University of Bristol.

External University partners include the University of Newcastle (Alex Yakovlev's group), and the University of Southampton (Steve Beeby's group).

Edit this page