by Tom Skwara

Example wireless sensor network virtualizing physical measurements.

Virtualizing physical measurements using wireless sensor technology brings new solution choices to mission challenges.  Wireless sensors are smart devices that realize measurements without the using external data acquisition (DAQ) equipment or external power sources. 

The hardware needed to directly digitize low-level signals, resides within each sensor. The analog-to-digital converter (ADC) found in a highly integrated microcontroller is typically accompanied by a multiplexer (MUX) and programmable gain amplifier (PGA). 

Further inclusion of other mixed-signal peripherals such as comparators and digital-to-analog converters (DAC), promote the sensor as a complete measurement processing system.

The self-contained sensor can buffer consecutive high-speed measurements within its internal memory in preparation for sophisticated post-processing analytics. 

A Fourier transform (FFT) for example can be applied to the captured waveform for conversion to a magnitude and phase component of each frequency.  This analysis in the frequency domain reveals which frequencies are present in the waveform and builds a characterization model that describes the waveform’s signature. It is this signature as opposed to raw measurement data that can be communicated wirelessly to destination hardware in the sensor network pathway. 

There is an important benefit to this approach. The power required to periodically transmit high-level descriptive information is magnitudes less than what is required to continuously stream raw measurement data. Since the wireless sensor is battery-powered, energy savings is critical.

Example wireless hardware, including sensors and gateway units.

Ultimately the waveform signature description is received by another device that can conceivably be another sensor. A sensor that receives high-level description data from another sensor could use this auxiliary information to improve its own accuracy by discriminating against the frequencies of the first sensor. 

Alternately, the destination device may be a gateway that reconstructs the waveform from the descriptive signature transmitted from one or more sensors in the network. The waveform in turn can be fed to the inputs of a legacy DAQ system that would otherwise see the sensor’s direct output.

Before considering the details of data acquisition, post-processing and RF transmission, one must review the type of physical measurement that is being made and select the appropriate fundamental sensing method.

Pressure measurement can be accomplished using conventional sputtered film technology where a metal diaphragm deflects under pressure and its surface tension and compression regions affect change in the resistive legs of a Wheatstone bridge.

A constant current or voltage generated by the microcontroller’s DAC provides excitation to the bridge, whereby a small differential voltage is developed that is proportional to the pressure on the diaphragm.  The ADC converts the differential signal with the assistance of the MUX and PGA as previously described.

Temperature measurement can be accomplished using a platinum RTD that changes its resistance proportional to its temperature.  A constant current sourced by the microcontroller’s DAC provides excitation to the RTD, whereby a small voltage is developed that is proportional to the RTD temperature. The ADC converts the signal with the assistance of the MUX and PGA as previously described.

Flow measurement can be accomplished using a turbine flow meter whereby impeller fins disrupt a magnetic field. The field disruption is detected by a proximity switch that triggers a comparator within the microcontroller at a rate proportional to the rotational speed of the impeller. The frequency of the comparator output is measured using the microcontroller’s accurate time base.

Signal conversion to the digital domain within the sensor enables realization of the measurement in engineering units when correlated to internally stored calibration values. 

In addition to using post-processing methods to derive complex meaning of measurements (e.g. pump stroke signatures), the sensor may simply be delegated to reporting measurements in engineering units or percentages of full scale. 

This can occur on a fixed cadence, or in response to a programmed threshold of change. In either case, the receiving hardware is faced with some important decisions to make should the reporting interval be interrupted, or vary due to special circumstances such as a static measurement becoming dynamic

Virtualizing physical measurements using wireless sensor technology relies on deterministic behavior of the sensor network, but must also have agility to react gracefully under fault conditions.  Receiving hardware relies on regular updates from sourcing hardware in order to preserve the integrity of the original physical measurement.  If a sensor’s transmission is blocked or is disrupted momentarily, the downstream receiving hardware must have the ability to properly handle the exception. Additionally, mesh sensor networks may dynamically reform themselves based on link quality and other factors, and so the resulting delay forces receiving hardware to handle the exception. 

An exception handling strategy is a system requirement since the nature of a disconnected system is that where disruptions are bound to occur.  For instance, the system designer may establish an acceptance tolerance on the number of consecutive lost updates. 

Lost updates within the acceptable tolerance band may be interpolated, while an out-of-tolerance reporting condition may initiate a special sequence of events or generate an alarm.

While there are several RF technologies to consider for use in wireless sensor networks, it is ZigBee technology that represents the best combination of features that fit the needs of measurement virtualization solutions. 

ZigBee is a global standard, highly reliable, low cost, low power, low data rate, and highly secure.  The mesh network topology that ZigBee technology enables provides reliability and agility when faced with changing environmental conditions.

Mission-critical systems using wireless sensor technology require a reliable power source within each sensor.  Important criteria for battery selection includes superior energy density, long shelf life, low self-discharge rate and low weight.  Lithium cells offer the best combination of these attributes.  Several types of lithium batteries can be considered, including lithium thionyl chloride, manganese dioxide, and sulfuryl chloride.

Virtualizing physical measurements using wireless sensor technology enables a host of new solution choices.  The elimination of wired connections and external data acquisition equipment is the obvious benefit, but the true power of measurement virtualization lies in the advanced signal processing capabilities within each sensor. 

The ability to communicate descriptive information between sensors to enhance measurements or enable complex system measurements, adds to the possibilities.  Understanding virtualization concepts, selecting the best hardware, and making the right design choices all play a critical role in helping ensure your critical mission is a success.