Goals of a Terrestrial Sensor Node
The goals of this monitoring system can be summarized in five objectives:
1) To monitor micro-climates on the islands,
2) To monitor soil quality,
3) To serve as a point of reference for the triangulation of the location of the electron tags placed upon certain species,
4) To do a portion of the seismic activity monitoring, and
5) To serve as the information backbone of the entire network of stationary sensing nodes and tags.
While this is a sizable task, our sensor node must remain relatively small, easily portable by a team of scientists on foot, and not too intrusive into the environment. All of the tasks we are assigning to it are currently available and implementable, if a little on the experimental side, but by placing the sensors we are automating the tasks that would otherwise require a more obtrusive human presence.
Design
The sensors will be based around a combinational synergy of commercially available sensors, platforms, power sources, and communication interfaces that provide real world functionality with the flexibility to be expanded and enhanced as sensors technology and need increase. Thus we find in the design an inherent elegance that ensures its survival for generations and iterations to come.
To address the task of monitoring micro-climate in the Galapagos we will measure many of the quantities that you would expect to see on the evening news broadcast: rainfall, humidity, temperature, pressure, windage, and pH (acidity). Most manufacturers suggest that this part of the system be placed on a pole 10 centimeters in diameter and approximately 3 meters long, buried between half and a full meter in the ground. This pole will be constructed of the material as described by the Materials portion of the website. Many companies produce commercial units that are able to monitor the quantities mentioned, and are even suited for rugged environments. The power source will be a solar cell atop the pole supplying a battery system with near continuous power during daylight hours. The remaining portions of the weather sensor portion of the sensor will be placed near the top of the pole saving approximately a 25 cm wide swath to allow for the placement of radio transmitters, which will use the Zigbee and ARGOS (island mother node only) protocols.
Soil quality is a very subjective quantity that can be judged only from a baseline sample of the area as to whether or not the soil is conducive to the growth of a certain type of flora. The sensors which we are including in this sensor network are all judged to have a vast impact on the surrounding environment and greatly effect how nearby flora grow. We wish to monitor pH, heavy metal content, water content, percent organic matter, amounts of potassium, nitrogen, and phosphorous. All of these quantities can be easily monitored from a tethered, buried sensors placed below the base of the pole. If this is impossible due to excessively shallow bedrock or dirt that is packed to firmly the sensor should be placed next to the pole at the greatest possible depth achieved using hand tools.
The game tracking tags described in the electronic tags section on the website describes a semi-autonomous portion of the sensor network which requires the terrestrial sensing nodes to operate. A brief description of what occurs is one of the electronic tags emits a short, high power burst of information every three minutes. Only two pieces of data are transmitted in this burst: a tag identification number that correlates to that specific tag and, therefore, a specific animal, and a time stamp set to the sensor system's clock. Then the terrestrial nodes that receive it, depending upon the location and the animal on which the tag was placed, numbering between one and three store that information until their node's data transfer time. The information will be relayed in the same process as can be found in the Communications and Data Processing portion of the website. This data is then accumulated and processed by a means described in Algorithm Design and Data Handling that attempts to triangulate, or best approximate given the number of sensor nodes that receive theinformaton, the position of the tag. As with all of the other data this information is stored on the offsite server for further analysis by scientists and researchers. At each of the nodes, to accommodate for this task, a receiving antennae approximately one meter in length by one half meter wide will be placed on the pole and operate at 2.4 gigahertz to receive the data.
Seismic monitoring is a two tiered system described in the Seismic monitoring section of the website. Half of it occurs directly on the slopes of the volcanoes, the other portion of it occurs at the mother terrestrial sensor node on each island. That which occurs at the site of the mother node is done with a seismometer to collect relatively general information about how the island is moving and the vibrations in the crust. This portion of the system then interfaces with the communications described in Sensor Node Transmission Handling. This monitoring has a frame of reference broad enough that only one sensor like it will be necessary per island while the infrasonic data is so specific 10 to 15 will be needed to provide the picture of the volcano. The placement of this sensor will be buried in the ground approximately 2 meters beneath the soil surface and placed the minimum distance it can from the sensing pole as possible to minimize from the product wastage and footprint of the sensor.
The communications infrastructure that we chose to implement for the vast majority of the sensors on the Galapagos is based entirely around the communications setup of these nodes, their ability to "talk" to each other and to broadcast data out to the ARGOS system. We selected this option because these nodes provide a fixed location from which to reference much positional data that could not be gained in other ways. Further, we were worried about the power source of these sensor nodes. While these sensors are using high capacity battery and a solar panel to recharge this battery, if we selected to have them each transmit the signal to the other side of the world our power consumption would be enormous. Thus, in requiring them to only transmit for a relatively short period of time for distances of no more than 5 km for every node, excluding the mother node which must communicate with ARGOS, we ensure we are conserving power for that inevitable day when there is no sunlight to collect and we must rely wholly on the battery. Beyond simple transmission of data, however, onsite data storage is key. Between transmission windows and when receiving data from other sensors there must be some way to store the data. We chose a weatherproof "datalogger" which primarily consists of a weatherproof case attached to the pole and a hard disk to store data. We estimate that it should hold at least four days worth of data for up to five nodes. We believe that this is on the order of two gigabytes of data per day for the entire network and as such a storage device of approximately 10 gigabytes will be needed. Rather than use of a conventional spin-up disk drive, we opted for a solid-state flash-type storage solution, as it would use less energy and be less succeptable to damage from being hit or jostled. For further information about the transmission guidelines please refer to Sensor Node Transmission Handling for additional information about the communications system.
Beyond the sensor portion of the sensor they will also need a power source. We have in mind a redundant system combining a 55 amp-hour rechargeable lithium-ion battery and a solar panel to recharge the battery. The solar panel will be a 20 watt panel attached to the top of the pole anchored in the ground. The solar panel will be approximately 50 cm by 30 cm. The battery will be sealed so it cannot leak contaminants into the environment, and will be kept inside the waterproof canister that is also storing the datalogger. We estimate that at most the sensor monitors and the communications system may consume 5 amp-hours a day. If the solar panel goes bad sometime during the course of monitoring it can send a trouble-shooting message to the server so we may dispatch a repair crew before valuable data is lost.
The inter-node and satellite communications will look much similar to the communications antennae of the Zigbee system. The radio frequency transmitter will be a high gain antennae that is unidirectional so that it may send and receive data at the extremes of the 5 km range that we established. It will more than likely be a mono-antennae extending from the side of the shaft anchoring the sensor node to the ground. The mother node on each island will also have one more communication device that other nodes on its island will not: a satellite transponder capable of transmitting the data we gather to the ARGOS satellite system to be stored on our offsite server. These come in various shapes and sizes, but for our application high connectivity, unidirectional precision is needed so we will use a satellite dish about 30 cm across, similar in size and scope to a satellite dish used for satellite television programming in many homes. It will be mounted to the side of the pole opposite the radio transmitter and pointed due south at the ARGOS satellite in polar orbit.
Locations on the Islands
Placement of the nodes is key if we are to collect the information that we wish to at the accuracy that is required by the scope of the project. As such we looked at different maps displaying the Galapagos. In our search we came across a series of map that display the micro-biomes present on the Galapagos. This is particularly valuable to us because each biome has slightly different environmental qualities from one another, and therefor different data, and we believe that we will find it invaluable to know what is occurring in each specific part of the island at any given time.
With thought to the micro-biomes communications standards of a low power radio frequency and Zigbee incorporated into the tags, and the specificity that we wished to receive with our sensors we decided that a maximum sensor separation of approximately 5 km was the best possible distance. Further, due to the biomes we decided that we wished to place at least one sensor within each biome on each island. While this occasionally provides for some violation of the 5 km rule it is more imperative that we collect complete, accurate data than follow a standard that we implemented and omit valuable data. While we did not enumerate explicit locations for node placement, we did so consciously. While we have access to topographical maps we cannot fully justify placing a sensor in a location without either a visual or more information than we could find. If these two guidelines are followed, however, the system will be performing at its peak.
Installation and Maintenance
As mentioned earlier the sensor must be small and light enough to be carried in by a team of scientists on foot for placement. A team of approximately 4 scientists must be able to carry at least two sensor nodes and hand tools, a shovel and post hole digger, for placement into the ground. The team will be dropped off on the island on the shore and will then hike in on a specified trail avoiding known areas of high animal traffic so as to avoid interrupting their habitats. They will install the system during daylight and be off the island by nightfall if possible. If that is infeasible for some of the more inland locations they will need to at least spend the night. In this event they will employ Leave No Trace principles (Leave No Trace Organization, 2004) to aid in minimizing their impact on the environment. While there will be an established location for a sensor that they are going to plant as they go out, if they find a more favorable location in the same micro-biome they have the authority to modify the plan.
Maintenance on these sensors will be evaluated on a continual basis by examination of the data that they are transmitting. If it is corrupted or a message is transmitted enumerating a problem somewhere with the sensor a team much like the installation team can be dispatched quickly to address the problem. Further, the scientists placing tags on the animals will do a physical evaluation of the condition of wear and tear on the sensor and evaluate whether there is a need for invasive repairs or replacements. While we do not predict widespread animal or human destruction of the sensors, the possibility is always there. ORGALA will deal with the policy in the Ecuadorian government to deter humans from vandalism and punish harshly those that do. If animal destruction occurs, the sensor will be repaired and relocated from that site to another site within the same micro-biome if possible. If it is not then the sensor will be placed at the discretion of the scientist team. In either case the node should not be placed within a 1 km radius of the previous site. The damaged pieces are to be removed and returned to the storage facility to await analysis for why and how they were damaged so that in future iterations of the system we can attempt to account for the inadequacy.
Sources:
Leave No Trace Organization. (2004, July). Leave No Trace Principles [WWW Document]. URL http://www.lnt.org/programs/lnt7/index.html (visited 2004, November 18).
