GPS: GPS Q&A: Industry experts answer reader's GPS questions

Q.How is the new millennium date going to affect GPS? I've heard rumors that the equipment I currently use will only be good for a few more years!
C.G. Santa Fe, N.M.

A. Matthew Florio, Trimble Navigation: GPS time is divided into GPS weeks, which commenced on Jan. 6, 1980. The GPS ICD 200 (Interface Control Document) states that the GPS control segment allocate 10 bits when broadcasting the GPS week. This means that a maximum number of 1023 can be broadcasted. Week 1024 commences on Aug. 22, 1999.
      With regard to receivers, there exists two possible scenarios. One, some receiver types will rollover, restarting the GPS week at 0. Two, some receiver types will rollover, restarting the GPS week at 0. Two, some receiver types will report the true week (i.e. 1024, 1025, etc.), computed from Jan. 6, 1980. Whether your receiver reports scenario one or two depends upon the receiver design. There may be compatibility issues using receivers of different designs in the same job.
      Most GPS manufacturers are aware of the issues regarding the rollover and are taking appropriate steps to resolve them.

Leonard R. Kruczynski, Ashtech, Inc.:
If you decided that two characters were sufficient to identify a year and that those two characters should be the last two digits of the calendar year, then there is a problem in the year 2000. Is 1 January 00 before or after 31 December 99? This 1000-year ambiguity has recently received much publicity as we approach the new millennium. The GPS navigation message has a similar challenge whose denouement will occur a few months before New Year's Day of 2000.
      Midnight, Saturday night, 21 August 1999, will mark the end of GPS week number 1023 and the beginning of GPS week number 1024. The GPS broadcast data uses week number and time of week to time tag the data and the date of validity of data. Week number is used in almanac information, in UTC offset data and in the ephemeris and clock data. The longest field for week number is in the ephemeris and clock data and that field is 10 bits long. Ten bits can represent numbers from 0 to 1023. At midnight on 21 August 1999, the week number will become 0, not 1024.
      The navigation data will continue to be accurate and a receiver should be able to compute valid and accurate position, velocity and time of week. If the software is designed to output a calendar date, then there may be a problem, There is no information in the navigation data that explicitly differentiates between the week of 22 August 1999 and the week of 6 January 1980. The GPS Standard Positioning System Signal Specification says that "Users must account for the previous 1024 weeks in conversions from GPS time to a calendar date."
      Without external assistance or protective algorithms, it is quite likely that GPS equipment will provide good position, velocity and time of week but erroneous calendar dates. There are ways to account for the week rollover and to provide correct calendar date without the need for a modification to the GPS navigation message itself. Customers can expect that manufacturers will address this issue and, if necessary, will provide updated firmware for the equipment. GPS receivers should have no problems on the transition from 1999 to 2000.

Jon Vavrus, Magellan Systems Corp.:
Magellan does not anticipate problems with our GPS equipment in the year 2000 and beyond because several years ago we added code to all our two-channel and 10-channel receivers that will handle the rollover.
      The questions about the new millennium's effect on GPS equipment arise from the Global Positioning System's time standard, which is based on numbered weeks. All Magellan GPS software has a ROM almanac associated with one of those numbered weeks. Almanacs are typically one to six weeks old at the time of release, meaning a GPS 3000 receiver associated on Oct. 22, 1995 and released on Nov. 17, 1995 would have a ROM almanac corresponding to week 824 in GPS time.
      The software assumes that the current date is occurring after the week in the ROM almanac. At the turn of the millennium, when the satellites will broadcast a numbered week that is less than the numbered week in the receiver's ROM almanac, the software code will increase the number by 1024 to account for the week rollover. The adjustment will enable the receiver to operate for 19 years after the software production date. In other words, that GPS 3000 receiver released on Nov. 17, 1995 will operate until GPS week 1847, or May 31, 2015. After 19 years, units cannot be initialized but will continue to function as long as they do not lose their memory (for instance, if the receiver lost battery power for a prolonged period of time).

Francis Yuen, NovAtel:
There are two major issues surrounding the way GPS receivers will handle the time information during the transition from this millennium to the next - how will the receiver handle the UTC time offset that will occur at the end of the year, and how will the receiver manage the "GPS week roll over" that will occur on the week of Aug. 22, 1999.
      UTC Time: The GPS is based on an universally synchronized time frame called GPS time, GPS time first started on Jan. 6, 1980 at 0:00 hours, and at that time it was perfectly aligned with UTC time. However, GPS time is not corrected for leap seconds whereas UTC time is. Therefore, every time a leap second correction is applied to UTC, an additional offset is introduced between the UTC and GPS time scales - this offset is currently 11 seconds. Fortunately, GPS satellites broadcast this offset and it is a relatively trivial matter for the GPS receiver to interpret and apply the offset information to its calculations. Thus the transition from 1999 to 2000 and beyond) will be easily dealt with.
      GPS Week Roll-Over: The roll-over problem is a little more complicated. GPS week number is represented by a 10 bit binary number, hence, it can be represented by any integer number between 0 and 1023. We are currently in week 874, and a quick calculation reveals that by Aug. 15, 1999, it will be week 1023. So the question is how are we supposed to represent the week after Aug. 22, 1999? According to the official GPS specifications (ICD-GPS-200, paragraph 20.3.3.3.1.1), "...the 10 bits shall represent the number of the current GPS week...," it explicitly specifies the size of the week counter (i.e. 10 bits). In other words, as per the ICD document, the receiver should reset the GPS week number back to 0 (i.e., reset the 10 bit GPS time counter) and not advance it to 1024. Since there are many discussions and arguments over the interpretation of this so called GPS week roll-over problem, different manufacturers will no doubt interpret the ICD document differently and, as a result, users may find incompatibility between different makes of receivers.
      The question will continuously be raised, and left unresolved, until the U.S. Department of Defense issues a definitive statement which governs this GPS week roll-over issue. In the absence of such a directive, when purchasing GPS equipment, users should ensure that the equipment is capable of handling either or both of the above scenarios.

About the Participants:
Matthew Florio is a senior mapping trainer at Trimble Navigation in Sunnyvale, Calif. He may be reached at 408-481-8000 (phone) or 408-481-8699 (fax). Leonard R. Kruczynski is director of strategic relationships at Ashtech Inc. in Sunnyvale, Calif. He may be reached at 408-524-1505 (phone), 408-524-1500 (fax), or e-mail: [email protected] Jon Vavrus is chief technical officer at Magellan Systems Corp. in San Dimas, Calif. He may be reached at 909-394-5000 (phone) or 909-394-7050 (fax). Francis Yuen is a customer service manager at NovAtel in Calgary, Alberta, Canada. He may be reached at 403-295-4650 (phone), 403-295-4901 (fax), or e-mail: [email protected]

Back