At the Gateway, data are stored on net-book computer and transmitted via cellular modem to an ... from other nodes to fo
THE UNIVERSITY OF GEORGIA SMART SENSOR ARRAY George Vellidis Crop & Soil Sciences Department University of Georgia
[email protected] The UGA Smart Sensor Array (UGA SSA) consists of smart sensor nodes and a Gateway. A “smart sensor node” is defined as the combination of electronics and sensors installed at each location in the field. A UGA SSA node consists of a circuit board, a radio frequency (RF) transmitter, soil moisture sensors and temperature sensors. Each sensor node accommodates up to 3 Watermark® soil moisture sensors and 2 thermocouples for measuring temperature (Figure 1). In the future, other sensor types (nutrient, pH) can be added to the node. The RF transmitter is a postage stamp-sized intelligent low-cost, low-power radio module capable of acquiring, analyzing, and transmitting sensor data (Figure 2). Data from all the nodes are routed to a centrally located node known as the Gateway at 5 minute intervals. At the Gateway, data are stored on net-book computer and transmitted via cellular modem to an FTP server hourly. The net-book computer is powered by a solar panel (Figure 3). One unique characteristic of the UGA SSA is that it uses wireless mesh networks to communicate between irrigation sensor nodes. As the name implies, mesh networks create a wireless network between the nodes. The RF transmitters act as a repeater to pass along data from other nodes to form a meshed network of nodes. If any of the nodes in the network stop transmitting or receiving or if signal pathways become blocked, the operating software reconfigures signal routes in order to maintain data acquisition from the network (Figure 4). To overcome the attenuating effect of the plant canopy on radio transmissions, the RF transmitter antenna is mounted on spring-loaded, hollow, flexible fiberglass rods approximately 8ft above ground level (Figures 5, 6, and 7). This design allows field equipment such as tractors and sprayers to pass over the sensors – something which no other truly wireless system offers. The effective range of the RF transmitter is about 2500 ft. An important characteristic of our system is its affordable cost – a 20-node system can be installed for a onetime cost of $3500. Installing irrigation sensors throughout an irrigated field is key to understanding and managing the soil moisture variability which exists in all fields (Figure 8). With extensive testing, we have proven that our system is robust, reliable, easy to use, and affordable. We have been developing the UGA SSA for the past several years. We have tested the system in its current configuration under field conditions for two entire cropping seasons including the 2011 season and the system has proved to be extremely robust. Each sensor node is powered by two AA batteries which under field conditions have an effective life which exceeds 140 days. Field Testing the UGA SSA The manager of the field shown in Figure 8 used a commercially available soil moisture monitoring system to schedule irrigation in the field during the 2010 growing season and was pleased with the results. As a result, he elected to use the commercial system to schedule irrigation during the 2011 season as well. Because of poor calibration (the commercial system is
Vellidis / p.2
a capacitance sensor which requires extensive calibration after installation), the system was under-predicting soil water content throughout the 2011 growing season. Because the farm manager followed the commercial system’s recommendations for several weeks (June – July), he over-applied irrigation water in July which promoted fungal diseases in the cotton. To reduce the incidence of disease, he did not irrigate in August even though only one small precipitation event was recorded. The UGA SSA performed well during the entire growing season. There were no failures and all the data were successfully recorded by the netbook computer at the Gateway. As a result of the irrigation scheduling protocol used at the Moss Field, soils as measured by the UGA SSA were wet to moist in June and July (consistently below 30 cb or kPa) and dry to very dry in most of the field during August (exceeding 150 kPa) (Figures 9-13). The strength of the UGA SSA is that its low cost allows for the installation of many sensor nodes which clearly show the variability of soil water tension in the field. Sensors installed at 4, 12, and 20 inches provide excellent information on the effect of irrigation and precipitation events on soil water tension.
Vellidis / p.3
Figure 1. A UGA SSA sensor node consists of the sensors which are installed in the soil (pictured at left) and the electronic components (pictured at right). The three Watermark® sensors are integrated into a shaft which can be easily installed after planting and extracted prior to harvest for agronomic crops. In vegetables production, the sensor can be installed after the bed is prepared and remain in place until the bed is reshaped. The sensors and electronics are installed immediately adjacent to each other to minimize the length of exposed cabling.
Figure 3. Gateway for UGA-SSA system. Enclosure houses the net-book PC. Solar panel above recharges a 12VDC battery.
Figure 2. A sensor node's circuit boards pulled part-way out of its PVC enclosure. The node circuitry is powered by two alkaline AA batteries mounted to the back side of the sensor acquisition board.
Vellidis / p.4
Figure 4. Mesh networks establish communication pathways between nodes (left). GW represents the Gateway node. If a node is disabled, a new pathway is automatically developed (middle). The number of potential data transmission pathways for a dense network of nodes is large (right).
Figure 5. To increase transmittance range and to allow for farm vehicles to pass over the
node, the electronics are kept at the soil surface and the antenna enclosed in an 8 foot, 0.25 inch diameter hollow, flexible, spring-loaded fiberglass rod.
Vellidis / p.5
Figure 6. Sensor nodes from two commercially available systems (left and middle) installed adjacent to UGA SSA node 10 in the Moss Field shown in Figure 8 during the 2010 growing season.
Figure 7. Sensor nodes from two commercially available systems and (left and right) installed adjacent to UGA SSA node 2 (middle) in the Moss Field shown in Figure 8 during the 2011 growing season.
Vellidis / p.6
Figure 8. The Moss Field is approximately 150 acres in size and located near Camilla, Georgia. Three soil moisture sensing systems were evaluated in the field during the 2011 growing season. The numbers in the adjacent figure indicate the location of the 12 UGA SSA sensor nodes. AquaSpy and SmartCrop nodes were installed adjacent to UGA SSA nodes 2 and 10.
Vellidis / p.7
Figure 9. Soil water tension curves for UGA SSA node 2 at the Moss Field – the site where the two commercially available system sensor nodes were also installed. The UGA SSA node contains Watermark® sensors installed at 4, 12, and 20 inches. The top graph presents 01 June – 12 July, 2011. The bottom graph presents 10 July – 24 September, 2011. Irrigation events were scheduled using one of the commercial systems. The farm manager did not irrigate for four weeks in August because over-irrigation in July resulted in boll rot.
Vellidis / p.8
Figure 10. Soil water tension curves for UGA SSA node 6 at the Moss Field The node contains Watermark® sensors installed at 4, 12, and 20 inches. The top graph presents 01 June – 12 July, 2011. The bottom graph presents 10 July – 24 September, 2011. The farm manager did not irrigate for four weeks in August.
Vellidis / p.9
Figure 11. Soil water tension curves for UGA SSA node 3 at the Moss Field – the sandiest area of the field. The node contains Watermark® sensors installed at 4, 12, and 20 inches. The top graph presents 01 June – 12 July, 2011. The lead from the sensor installed at 20 inches was accidentally not connected to the circuit board upon installation. The error was discovered on June 23 at which time the lead was connected. The bottom graph presents 10 July – 24 September, 2011.
Vellidis / p.10
Figure 12. Soil water tension curves for UGA SSA node 8 at the Moss Field which is a wetter part of the field. The node contains Watermark® sensors installed at 4, 12, and 20 inches. The top graph presents 01 June – 12 July, 2011. The bottom graph presents 10 July – 24 September, 2011.
Vellidis / p.11
Figure 13. Soil water tension curves for UGA SSA node 10 at the Moss Field – the site where the two commercially available system sensor nodes were also installed. The UGA SSA node contains Watermark® sensors installed at 4, 12, and 20 inches. The top graph presents 01 June – 12 July, 2011. The bottom graph presents 10 July – 24 September, 2011. Please note that at this location, most irrigation events were not adequate to wet the soil at 20 inches. The farm manager did not irrigate for four weeks in August.