Science & Technology

Embedded sensor technology.

Wireless sensor networks (WSN) typically comprise an array of small microprocessing devices, distributed over a wide area and running on low-power sources which wirelessly transmit data from in situ sensors to a base station. The hardware platform is generally a microprocessor/radio board (MPR), often referred to as a mote. A complete node is a mote in conjunction with a data acquisition board (DAQ) and attached sensors that can be embedded in environmental media (e.g. water, soil or air). Our wetland/lake application required unobtrusive, rugged, waterproof motes that allowed two-way communication, a variable data collection frequency (minutes to hours) and the capability to function unattended for several months.

Crystal Bog WSN.

The wireless network at Crystal Bog was composed of nodes for monitoring water level, conductivity and water temperature of the open water (CB pond), and the porewater at several peatland sites. In addition, nodes were deployed for a rain gauge and an evaporation pan (added in 2009). The base station was mounted on a GLEON buoy deployed on CB pond. Data from the sensors nodes was transmitted from the base station to the host PC at Trout Lake Station via GLEON maintained Freewave radios (TCP/IP).

CrossBow Motes. The embedded sensor network in Crystal Bog was built in-house using the MICA2 mote and the MDA300 data acquisition board from CrossBow Technology, Inc . CrossBow Motes have the ability to form self-healing and multi-hopping mesh networks - when motes are distributed across a broad area they can form their own communication mesh that self-adjusts to interference and optimizes lines of communication between individual nodes.

CrossBow Motes. The embedded sensor network in Crystal Bog was built in-house using the MICA2 mote and the MDA300 data acquisition board from CrossBow Technology, Inc . CrossBow Motes have the ability to form self-healing and multi-hopping mesh networks - when motes are distributed across a broad area they can form their own communication mesh that self-adjusts to interference and optimizes lines of communication between individual nodes.

Wireless network software,- The mesh network of sensor nodes (motes with attached sensors) and gateway requires three distinct software tiers: 1) XMesh is the software that runs on the cloud of sensor nodes. This software provides the networking algorithms required for reliable communication among all the nodes within the mesh cloud. 2) XServe provides services to route data to and from the mesh network with higher level services to parse, transform and process data as it flows between the mesh and outside applications. 3) database management and data analysis software.

Node Components. Our CrossBow-based wireless node has four main components: the MICA2 mote, the MDA300 DAQ board and the mainboard PCB, all hosed in a waterproof enclosure plus the connected sensors.

The MICA2 mote is the hardware used to transmit the acquired data via a wireless connection to the base station and to run the sensor data acquisition firmware. The components on the MICA2 include the CC1000 radio chip, which is used to transmit and receive data packets, the Atmega128L, the microcontroller which simultaneously runs the data acquisition and the network communication firmware (TinyOS), flash memory for the firmware and data buffers, and the antenna. Motes in the Crystal Bog Mote Project broadcast over the 903MHz channel.

The MDA300 DAQ board is used mainly as an interface for the MICA2 to the rest of the system. Both the MICA2 and the MDA300 contain 51-pin connectors which facilitate this connection. The MDA300 contains the I/O pins that are used to collect data from the sensors and also to control them. Also contained within the MDA300 are the analog-to-digital converters allowing the microcontroller on the MICA2 to process the voltage signals coming from the sensors.

The mainboard PCB contains the circuitry to interface with the sensors, and it allows for all the components within the node to connect to one another. The mainboard was designed by Mike Morrow and Steve Yazicioglu at the Electrical and Computer Engineering Department, UW-Madison.

The sensors currently deployed are WL400 water level sensors and WQ301 conductivity sensors from Global Waters and thermistor probe assemblies from U.S. Sensor (p/n USP7881). Rainfall is monitored with NovaLynx 260-2500 tipping rain gauge . The evaporation pan was fabricated by putting flotation on a plastic tub and mounting a level sensor and thermistor on the bottom and deploying the device on CB pond filled with pondwater

Mainboard:

The mainboard has four main functions: 1) controlling power to the sensors, 2) scaling and converting the sensor output, 3) filtering the sensor output, and 4) physically connecting this hardware to the MDA300 DAQ board. The sensors are powered by 12-AA dry cells with the voltage maintained at 12V by a voltage regulator. The sensor power circuit is enabled by a 3.3 V excitation signal from the MDA300. The sensors are powered for 3 seconds warm-up period before the output signal is sampled by the MDA300. On the mainboard, the signal output from the sensors is converted to a voltage if necessary and scaled down to the dynamic range of the single-ended analog channels on the MDA300 (0 to 2.5 V). Finally, the signal then is filtered and passed to the MDA300.

Battery packs JP2, JP3, and JP4 supply the 18V power source which is used to power the sensors. The ENABLE is connected to the E3.3 pin on the MDA300. This pin supplies a 3.3V signal when the sensors need to be turned on. When ENABLE goes high, current is allowed to pass through Q2 drain to source. This creates a positive voltage at Q1. 18V is then applied to the input of U1. This 18V gets scaled down to 1/10th its original value through R8 and R10 to be connected to the MDA300 through VBATT_SENS so that it can be remotely monitored. U1 creates a stable 12V at its output, which is used to power the sensors. It is also scaled down to 1/10th its original value by R12 and R13 to be monitored through VREG_SENS.
Trout Bog INW WSN
For the embedded sensor network in Trout Bog, we used a commercially available product: the INW WaveData® Wireless data collection system obtained from Instrumentation Northwest, Inc. of Kirkland, Washington. The INW components used in this project included the PT2X™ and CT2X™ data-logging sensors and the WaveData® radio-telemetry units.

The PT2X™ is a smart sensor with internal data-logging capabilities of 520,000 records. The recorded variables are pressure, temperature and time. The sensor’s internal datalogger has programmable, multi-phase, logging sequence capabilities. The onboard chipset is programmed to record and output barometrically compensated pressure (gauged or absolute), temperature, and time over either Modbus® or SDI-12 communication protocols. The PT2X™ is powered by two AA batteries, which act as a backup power source when the sensor is connected to auxiliary power. In our application, PT2X sensors were deployed in shallow piezometers to monitor the TB wetland water table.

INW’s CT2X™ sensor is built on the same technology and platform as the PT2X™, but with added conductivity measuring and recording capabilities, as well as temperature, time and pressure. In our application, CT2X sensors were deployed in the bog pond, evaporation pan and in two wetland piezometers to monitor the concentration of bulk ionic solutes as well as water level and temperature.

INW’s Aqua4Plus™ software was downloaded from the company’s website free of charge. The software was used to program and calibrate the sensors, upload and download data files, and establish communication via the WaveData® radios.

As deployed in TB, seven INW nodes communicated by RF directly to a single INW host radio that was mounted on the GLEON buoy moored in the center of the bog pond. The longest communication path in this star-shaped network was <200m. The short-range host radio was connected to a longer range Freewave Ethernet radio via a Moxa NPort 5210 which was configured to act as a TCP serial server with an IP address on a VPN. Using the radio link between the Trout Lake lab and the GLEON buoy, we were able to communicate with field-deployed sensors from any remote location with Internet access. The TB network collected data continuously at 30 minute intervals from May to November 2009 without requiring a battery change.