Arthur Lange
The explosion in interest in precision agriculture technology has been accompanied by a blossoming uses for a number of enabling technologies, the two most important of which are the Global Positioning System (GPS) and Geographic Information Systems (GIS). While GIS technology offers tremendous capabilities for more informed Agriculture Management decision making, rendering competent decisions still depends on having reliable data. This paper deals with two issues related to obtaining reliable data. First, the importance of accurately identified locations to which all field mapping and subsequent treatments can be linked. Second, how Global Positioning System (GPS) technology can be used to build your data base and to collect data efficiently. What is reliable? What is accurate enough data? It depends on the functions and the objectives for which the GIS data base was developed. Some precision agriculture applications can be performed with less accurate data which cost much less to acquire. However, other applications, like spray control with GPS may require higher accuracy in order to prevent overlapping applications of chemicals. Determination of accurate coordinates for weed or insect infestations is essential for proper assessment requiring grower intervention. Once the grower determines where a problem requiring treatment exists, it is important to only treat the problem area, and leave the rest of the field alone, both to save money and not harm "good insects, " both important aims of an IPM program.
This paper's goal is to give an overview of GPS and how it may be used in precision agriculture applications including yield mapping, field scouting, pest management and custom chemical application.
The Navigation Satellite Timing And Range Global Positioning System, or NAVSTAR GPS, is a satellite based radio-navigation system that is capable of providing extremely accurate worldwide, 24 hour, 3-dimensional location data (latitude, longitude, and elevation). The system was designed and is maintained by the US Department of Defense (DoD) as an accurate, all weather, navigation system. Though designed as a military system, it is freely available with certain restrictions to civilians for positioning. The system has reached the full operational capability with a complete set of at least 24 satellites orbiting the earth in a carefully designed pattern.
The NAVSTAR GPS has three basic segments: space, control, and user. The space segment consists of the orbiting satellites making up the constellation. This constellation is comprised of 24 satellites, each orbiting at an altitude of approximately 11,000 nautical miles, in one of six orbital planes inclined 55 degrees relative to the earth's equator. Each satellite broadcasts a unique coded signal, known as Pseudo Random Noise (PRN) code, that enables GPS receivers to identify the satellites from which the signals came, and makes positioning possible.
The control segment, under DoD's direction, oversees the building, launching, orbital positioning, monitoring, and provides two classes of GPS service. Monitoring and ground control stations, located around the globe near the equator, constantly monitor the performance of each satellite and the constellation as a whole. A master control station updates the information component of the GPS signal with satellite ephemeris data and other messages to the users. This information is then decoded by the receiver and used in the positioning process.
The user segment is comprised of all of the users making observations with GPS receivers. The civilian GPS user community has increased dramatically in recent years, due to the emergence of low cost portable GPS receivers and the ever expending areas of applications in which GPS was found to be very useful. Some of these applications are: surveying, mapping, precision agriculture, navigation and vehicle tracking. The number of civilian users now greatly out number military users.
There are two classes of GPS service; the Precise Positioning Service (PPS) which is available only to users authorized by the military, and the Standard Positioning Service (SPS) which is available for civilian use.
Though GPS can provide worldwide, 3D positions, 24 hours a day, in any type of weather, the system does have some limitations. First, there must be a (relatively) clear "line of sight" between the receiver's antenna and several orbiting satellites. Anything shielding the antenna from a satellite can potentially weaken the satellite's signal to such a degree that it becomes too difficult to make reliable positioning. As a rule of thumb, an obstruction that can block sunlight can effectively block GPS signals.
The receiver must receive signals from at least four satellites in order to be able to make reliable position measurements. In addition, these satellites must be in a favorable geometrical arrangement. The four satellites used by the receiver for positioning must be fairly spread apart. In areas with a relatively open view of the sky, this will almost always be the case because of the ways these satellites were placed in orbit.
The position of a point is determined by measuring distances (pseudo-ranges) from the receiver to at least 4 satellites. The GPS receiver "knows" where each of the satellites is at the instant in which the distance was measured. These distances will intersect only at one point, the position of the GPS receiver (antenna). How does the receiver "know" the position of the satellites? Well, this information comes from the broadcast ephemeris that are received when the GPS receiver is turned on. The GPS receiver performs the necessary mathematical calculations, then displays and/or stores the position, along with any other descriptive information entered by the operator from the keyboard.
The way in which a GPS receiver determines distances (called pseudo-ranges) to the satellites depends on the type of GPS receiver. Basically, there are two broad classes: carrier phase based and code based.
The carrier phase receivers, used extensively in geodetic control and precise survey applications, are capable of sub-centimeter (cm) accuracy. These receivers calculate distances (called pseudo-ranges) to visible satellites by determining the number (N) of whole wavelengths and measuring the partial (phase) signal wavelength there are between the satellites and the receiver's antenna. Once the number of wavelengths is known, a pseudo-range may be calculated by multiplying 'N' by the wavelength of the carrier signal (L1 and/or L2, 19 cm and 24.4 cm respectively) plus the partial wavelength. It is then a straight forward (albeit complex) task to compute a baseline distance and azimuth between any pair of receivers operating simultaneously. With one receiver placed on a point with precisely known latitude, longitude, and elevation, and with the calculated baseline (distance between 2 points), the coordinate for the unknown point may be determined.
The relative cost of the carrier phase receivers is high, but technological advances have made the dual frequency (using both, L1 and L2) carrier phase receivers of today much more efficient than the single frequency (using L1 only) receivers that were state-of-the-art only a few years ago. With some of the newest dual frequency receivers very precise measurements (+ 1 cm) can be made in real-time. These receivers will be used in machine control applications requiring a high degree of accuracy.
Though less accurate than their carrier phase cousins, code-based receivers have gained widespread appeal for applications such precision farming. This popularity stems mainly from their relatively low cost, portability and ease of use.
Instead of using the number of signal wavelengths to establish pseudo-ranges, code-based receivers use the speed of light and the time interval that it takes for the signal to travel from the satellite to the receiver, to compute the distance to the satellites. The time interval is determined by comparing the time in which a specific part of the coded signal left the satellite with the time it arrived at the antenna. The time interval is translated to a range by multiplying the interval by the speed of light constant (c=186,000 miles/second). Ranges from at least four satellites are needed in order for a receiver to produce a position fix. Position fixes are made by the receiver roughly every second, and the more advanced receivers enable the user to store the position fixes in a file that can be downloaded to a computer for post processing.
Under normal circumstances, autonomous standard position fixes (SPS) made by code-based receivers would be accurate to within 25 meters. The DoD however began imposing its selective availability (SA) policy in July of 1992, which limits position fix accuracy to within 100 meters. The purpose of SA is to deny potential hostile forces accurate positioning capabilities. Military P-code (Y-code) receivers are not affected by SA, but as mentioned earlier are not available for the general public. In order to overcome the limited positioning accuracy, differential GPS (DGPS) techniques have been developed. DGPS enables the user to improve SPS and also to remove the effects of SA and some other sources of error. These differential correction techniques can produce positions generally accurate to within a few meters. A recent (April 1996) Presidential Directive indicated that SA may be discontinued by the DoD within the decade. Industry observers point out that this policy will probably not take place in less than four years.
Even if SA is discontinued, many precision agricultural applications will still require differential GPS (see below) to attain the accuracy required for their application.
There are also now code based receivers capable of sub-meter differential accuracy. Most sub-meter receivers require longer data collection times (about ten minutes), and perform best under very favorable satellite geometry, and with an unobstructed view of the sky. Some newer receivers can provide sub-meter in seconds. These receivers cost more at the outset, but provide a good return on the added investment by providing substantially higher productivity.
It is very important that users of code based receivers understand the position accuracy limitations of the receiver. Due to SA, each coordinate viewed on a non-differential GPS receiver's display is only accurate to within 100 meters. This accuracy can be improved on by taking an average of 200 or so repeated position observations of the same point. The resulting accuracy would still be below what many users would consider acceptable quality. In order to produce acceptable results, GPS data collected in the field must be differentially corrected either in real-time, or by post-processing the data.
Differential GPS (DGPS) can be employed to eliminate the error introduced by SA and other systematic errors. Differential GPS requires the existence of a base station, which is simply a GPS receiver collecting measurements at a known latitude, longitude, and elevation. The base station's antenna location must be located precisely, using carrier phase GPS or other traditional surveying techniques. The base station may store measurements (for post processed DGPS), broadcast corrections over a radio frequency (for real-time DGPS), or both.
The assumption made with the base station concept is that errors affecting the measurements of a particular GPS receiver will equally affect other GPS receivers within a radius of 200-300 miles. If the differences between the base station's known location and the base station's locations as calculated by GPS can be determined, those differences can be applied to data collected simultaneously by receivers in the field. These differences can be applied in real-time (especially applicable for accurate navigation) if the GPS receiver is linked to a radio receiver designed to receive the broadcast corrections. In some GIS mapping applications, these differences are applied in a post-processing step after the collected field data has been downloaded to a computer running a GPS processing software package. GPS processing software is typically integrated with GPS hardware and thus is provided by the receiver manufacturer. As a rule, post-processed DGPS is considered slightly more accurate than real-time DGPS.
There are many permanent GPS base stations currently up and running in the United States that provide over electronic bulletin boards to the users of code based receivers the data necessary for differentially correcting positions.
In addition, installation and testing has started in many parts of the US of the Coast Guard's DGPS beacon system. These stations are part of a coastal network of stations the US Coast Guard is planning that will provide real-time corrections over a radio frequency.
There are a number of commercial suppliers of real-time differential correction data. Two vendors, ACCQPOINT and DCI, use a sub-carrier on a number of commercial broadcast FM stations. Other vendors, John E. Chance and Racal, distribute their data from geo-stationary satellites.
Several key issues need to be explored when considering GPS as a tool for capturing coordinate data for a precision agriculture application. First and foremost, what are the position accuracy requirements? If the data will be used for site specific analysis that require position accuracy to be within three feet, code based differential GPS receivers will be used. If better accuracy is required, as for a spray controller, then carrier phase GPS techniques have to be employed.
Every GIS database must be referenced to a base map or base data layer. Ideally, the database should be referenced to a large scale, very accurate base map. If instead the base map is smaller scale (quad scale or smaller) there could be problems when attempting to view the true spatial relationships between features digitized from a small scale map and features whose coordinates were captured with GPS. This can be a real problem if an grower decides to use a particular GIS data layer that was originally generated using small scale base maps as a base to which all new data generated is referenced. The best way to avoid such incompatibility one should consider developing an accurate base data layer, based on geodetic control and photogrammetric mapping.
As mentioned earlier GPS has its limitations. The most important one is that it requires clear view of the satellites. Buildings, trees, overpasses and other obstructions that block the line of sight between the satellite and the observer (GPS antenna), make it impossible to work with GPS. Urban areas are especially affected by these types of difficulties. Bouncing of the signal off nearby objects may present another problem, that of distinguishing between the signal coming directly from the satellite and the "echo" signal that reaches the receiver indirectly. In areas that posses these type of characteristics, traditional surveying techniques must be used instead or to complement GPS positioning.
GPS is a positioning system that can also be used as a real-world digitizer for mapping point and line features such as roads or wetland boundaries. However, for large volume data collection which includes measuring many points, mapping, contouring, etc., one should consider photogrammetry, as a more efficient data collection tool.
Field portable GPS receivers are available for rapidly mapping insect infestations and this data can be accurately communicated to the field manager who may employ a custom spray operator to apply the correct chemicals only where they are needed. In addition, the spray operator will be able to provide a permanent record back to the field manager with GPS data of where and when the treatment took place. Yield monitors will be connected to GPS receivers to map yield. The resultant yield maps will help identify areas of the field requiring different treatments.
Arthur Lange, Precision Agriculture Product Manager, Trimble Navigation, Ltd., 485 Potrero Avenue, Sunnyvale CA 94086, phone: 1-408-481-2994, fax: 1-408-481-6074, email: art_lange@trimble.com