The most important criterion must be the purpose to which the user wants to put the system. Such questions as:  "Is he returning to a place he has been?" and "Is he navigation to abstract coordinates provided?" will quickly answer whether absolute or relative is appropriate.

Each of the applications you posit have very different environments in which to operate. 

• Airborne navigation has few nearby landmarks useful for confirmation. Takeoffs and landings are short-lived exceptions.
• Waterborne navigation will often have some useful landmarks except in open sea. Near to and at port landmarks are frequent.
• Landborne navigation usually has a plethora of useful landmarks with the possible exception of desert travel. Swamp and jungle may provide too many similar marks to be useful.

The three types of accuracy estimation to which you refer are commonly defined as follows:

• Predictable:  sometimes called geodetic is a measure of how close to coordinates of a certain datum the navigation system is able to perform consistently. 
• Repeatable:  is a measure of how close to a place visited previously, with the same equipment, the navigation system is able to perform consistently.
• Relative:  is a measure of how closely a measured position agrees with another similar system at the same time. This is very like repeatable accuracy, but across the whole type of system instead of just one system alone.

There is another, often more, possibly most important, accuracy estimation that should be considered for navigation while moving.

• Dynamic:  is a measure of the aforementioned criteria while the system is moving. A properly tuned Kalman Filter will always make a good navigation system better and a poor one acceptable. It is often the case that a system may be optimized for static operation, which detracts greatly from dynamic performance. A system tuned for an aerobatic platform will serve a cargo vessel poorly.

A fishing charter operator who has visited his "hot spots" and wants to return to them with clients will want a repeatable system and benefit if it is also predictable.

A pilot ordered to rendezvous with a midair fuel tanker will require predictable and relative accuracy and hope the same for the tanker. Both need excellent dynamic accuracy.

A crop-duster will want predictable and dynamic accuracy usually augmented with differential corrections.

An automobile driver's system usually has an additional "cheat sheet" in the form of a background map. The system can afford to have less predictable accuracy by assuming the vehicle is on a road and display position by forcing the illustration to show the vehicle on the road nearest instead of where it is actually calculated.

On the other hand, a farm implement has no road map to guide its report and it is moving. His requirements are very similar to the crop-duster but without the degree of risk.

A hiker will benefit most from a predictable system with good static performance because he may stop to take a position and orient.

I hope this is helpful. I firmly believe there is no one answer with the exception of "give me all three improved and it will be better."

Best wishes,

JAC

Maybe repeatable accuracy is most important for Marine, airborne and Land.

The simple answer: the best possible accuracy. There will always be asked for accuracies which are better than possible today (e.g. using GNSS).

The most important criterion must be the purpose to which the user wants to put the system. Such questions as: "Is he returning to a place he has been?" and "Is he navigation to abstract coordinates provided?" will quickly answer whether absolute or relative is appropriate.

Each of the applications you posit have very different environments in which to operate.

Airborne navigation has few nearby landmarks useful for confirmation. Takeoffs and landings are short-lived exceptions.

Waterborne navigation will often have some useful landmarks except in open sea. Near to and at port landmarks are frequent.

Landborne navigation usually has a plethora of useful landmarks with the possible exception of desert travel. Swamp and jungle may provide too many similar marks to be useful.

The three types of accuracy estimation to which you refer are commonly defined as follows:

Predictable: sometimes called geodetic is a measure of how close to coordinates of a certain datum the navigation system is able to perform consistently.

Repeatable: is a measure of how close to a place visited previously, with the same equipment, the navigation system is able to perform consistently.

Relative: is a measure of how closely a measured position agrees with another similar system at the same time. This is very like repeatable accuracy, but across the whole type of system instead of just one system alone.

There is another, often more, possibly most important, accuracy estimation that should be considered for navigation while moving.

Dynamic: is a measure of the aforementioned criteria while the system is moving. A properly tuned Kalman Filter will always make a good navigation system better and a poor one acceptable. It is often the case that a system may be optimized for static operation, which detracts greatly from dynamic performance. A system tuned for an aerobatic platform will serve a cargo vessel poorly.

A fishing charter operator who has visited his "hot spots" and wants to return to them with clients will want a repeatable system and benefit if it is also predictable.

A pilot ordered to rendezvous with a midair fuel tanker will require predictable and relative accuracy and hope the same for the tanker. Both need excellent dynamic accuracy.

A crop-duster will want predictable and dynamic accuracy usually augmented with differential corrections.

An automobile driver's system usually has an additional "cheat sheet" in the form of a background map. The system can afford to have less predictable accuracy by assuming the vehicle is on a road and display position by forcing the illustration to show the vehicle on the road nearest instead of where it is actually calculated.

On the other hand, a farm implement has no road map to guide its report and it is moving. His requirements are very similar to the crop-duster but without the degree of risk.

A hiker will benefit most from a predictable system with good static performance because he may stop to take a position and orient.

I hope this is helpful. I firmly believe there is no one answer with the exception of "give me all three improved and it will be better."

Best wishes,

JAC

Form my experienced, vehicle appplication with sub-meter accuracy is needs for safety monitor system.

Article Driver behavior detection based on PPP-GNSS technology

To Anthony,

Thank you very much for the comprehensive description and a clear explanation. 

Especially type called (by You) "dynamic" - including the Kalman filter - is interesting becouse of testing problems.  In dynamics we don't know:  is it accuracy of  system or method of filtration or filter settings.

I believe that it is better to talk about the type of measurement as a "dynamic" (in opposition to static) than about "dynamic accuracy". We can measure system accuracy during  movement  (dynamic trials) but all devices (like INS, magnetic sensors and also ... Kalman Filter) should be turn off its.

Athony,national radio navigation plans (USA, Sweden, EU, GLA ...available in net) define the minimum requirements for the accuracy associated with the type of navigation tasks. Exactly as you wrote. Therefore, there are few other questions related to system accuracy:

1. How (methodology!) These values have been determined?

2. Is there a relationship between accuracy and availability of position?

3. Is the relations between minimum accuracy and the speed ? (of course... but how to white it as a equation?) 

Cezary

To Carl

I'm afraid the answer is not so simple. If selected receiver with the best possible accuracy then in your car you would have geodetic GNSS receiver working with the danish active geodetic network (like EUPOS) with accuracy aprox. 3 cm and the price of 20,000 Euro. Then what about the car price? Another aspect of this case is the lack of practical use of such a high accuracy. Yet you probably have a simple code GPS receiver with an accuracy of 9m (p = 0.95) horizontally and 15 m (p = 0.95) vertically (with map system)  and most importantly it performs all navigational  functions for a car.

cezary

To Raksit

Thank you for Your comment. Does all safety monitor systems needs sub-meter accuracy? What about the differences between environmental conditions  as: air, land and sea?

PS. Publication is very interesting !!

cezary

To Qingzhou Mao

I think You are right provided that we use only the coordinates displayed by the receiver for navigation (no system of map). In contrast, if we display the position (execute a navigation) on the map in a specific reference system (like WGS-84) I think predictale accuracy is a more representative measure. What is Your opinion?

cezary

The main problem in car is urban canyon, to make the detection correctly in urban is really hard, sometime accracy level has to measure for shut the detection off in some places. For highway, there is less problem.

Raksit Thitipatanapong 

Thanks for answer. Did You analysed position availability of GPS (in %) in the urban canyon in compassion to the highways (without map aiding) ? I realized similar measurements in 1996 (!). Article: Trials of DGPS in the Area of Northern Poland (available in RG). Then (in 1996)  were not dedicated GPS car receivers (GPS accuracy aprox. 100 m - before SA off). They also should not use the reflected signals and had a very low sensitivity (I used maritime DGPS receiver). How it looks today in Your studies?

I have not made such an analysis. The advances of receiver and computer technology advances at such a pace as to render such tests a mere snapshot in time of capability.

Many of the newer techniques involve using augmentation from sources other than GPS (GNSS), from WAAS to signal-matching to positioning with cell phone antennas and Wi-FI sources and, of course various levels of inertial navigation.

Taken as a whole, the amalgam tends to reinforce more general categories, but complicates comparisons between products with "hard numbers."

The ever growing expectations of near perfect performance and raised the importance of setting one's system in the correct datum to a similar level as the quality of the hardware.

JAC

Hi

if I understand your question correctly what you should be reading to is GPS SPS (Standard Performance Service) SIS (Signal in Space). In there you will find information about SPS  performance characteristics which are : availability, health, accuracy (it is described in two ways, one is the 95% SPS SIS user range error (URE) at a specific age of data (AOD), and the other is the 95% SPS SIS URE over all AODs), continuity (is the probability that the SPS SIS will continue to be healthy without unscheduled interruption over a specific time interval, and the UTC (USNO) accuracy (95% error in the parameters (ref. 20.3.3.5.2.4 of IS-GPS-200 contained in that SPS SIS which relate GPS time to UTC (USNO).

Another standard that you should be looking to is MOPS (Minimum Operational Performance Standards) 229D for GPS/WAAS airborne equipment. This standards only provides performance indicators for single frequency equipment commonly used nowadays (although a bit old fashioned technique) in aviation.

Hope this helps

Best regards

Stavros

Stavros

Thanks for Your ansver. We  discuse about USER minimum performance (defined by e.g.  national radionavigation plans). Not about the system minimum performace.

cezary

Nowadays, single frequency U-blox series 7 or 8 is really good, for vehicle safetly is good enough for velocity mesurement. However, position accuracy slightly difference within same path.

This depends o the application and whether it s or not for a SoL service:

- as far as regards terrestrial navigation, accuracies ranging from 1 m to 10 meters are possible, depending on the involved application, on the kind of vehicles (private cars, ambulances, trucks,..) and on the service requirements (navigation, SoL applcations, ..);

- as far as regards maritime navigation, it depends if only navigation is required or also precise harbor approaching s needed; in general Stand-Alone GPS is not sufficient and a WAAS/EGNOS solution (1-2 meters), DGPS solution (3-5 meters) or RTK (Real Time Kinematic) solution can be required;

- for aerial navigation more strict requirements are needed. EGNOS has been certified in 2011 for integration into GNSS positioning solutions for aerial applications; a planimetric accuracy of  1-2 and a vertical accuracy of 3-5 m ,  can be achieved.

To start with, Augmentation help improves the limitations of GPS, GLONASS, BEIDOU, GALILEO (GNSS) in providing Integrity, Availability, Accuracy and Continuity as performance parameters. These performance parameters are essential in the  area of applications you listed above with aviation (Air) topping the order followed by maritime especially at the stage of entering a port.

For instance, minimum system performance for Aviation (CAT I) are as follows:

ACCURACY: 1.0metre (Horizontal 95%) and 1.5 metre (vertical 95%)

AVAILABILITY: Greater or equal to 0.999

INTEGRITY RISK: Less Than or Equal to 4.0 E-7/Approach

ALERT TIME: Less than or equal to 5.2 seconds

CONTINUITY: Greater than or equal to 1-5.5E-5/Approach.

Some regional Space Based Augmentation System (SBAS); across the world (Operational and Planned) are listed below:

a. USA: Wide Area Augmentation System (WAAS), expanded to Canada as CWAAS and planned expansion to South America.

b. Europe: European Geostationary Navigation Overlay System (EGNOS) with expected expansion over Africa. 

c. China: Chinese Satellite Navigation Augmentation System (SNAS).

d. Japan: MTSAT Satellite Augmentation System (MSAS). MTSAT means Multi-functional Transport SATellite.

e. India: GPS-Aided Geo-Augmented Navigation (GAGAN).

f. Russian Federation: System Differential Correction and Monitoring (SDCM).

g. AFRICA: Nigerian Satellite Augmentation System (NSAS): First with Nigeria and expansion to African countries including oceans to the extent of NIGCOMSAT-1R coverage.

N.B: Some modern generational GNSS (i.e Third (newer) generation of American GPS) strives to meet some of the performance requirements without augmentation.

To Raksit

Thank you for your reply. I agree that the series ublox 7 or 8 is very good. In the manual I read that the accuracy is 2 m (CEP). Can you confirm it?

cezar

To Rafaela

Thank you for a very substantive answer. I share your view. JAC gave a similar response. Do you think that You presented  accuracy are only related to the predictable accuracy?

cezar

To Lasisi

Thanks for detailed presentation requirements for CAT-1. First time I'm hearing about the Nigerian Satellite Augmentation System. Can you recommend to us  some materials ?

cezar

For multi-gnss u-blox 8, it is around 1.5 meter. For true point, its might not good as carrier phase RTK. However, velocity vector It is good enough.

Thanks Raksit

cezar

To Cezary.

 Cezary, the Nigerian Satellite Augmentation System (NSAS) is relatively new.

Nigeria’s first communication satellite (NIGCOMSAT-1), a quad-band high powered satellite with navigational capability and capacity launched on 13th May, 2007 was Africa’s first contribution to the Global Navigation Satellite System. It was however de-orbited on the 10th of November, 2008 due to an irreparable single point of failure on-board the satellite. All broadcast, telecommunication services being offered by the satellite including strategic navigational plans and objectives were disrupted. The NIGCOMSAT-1R launched on the 20th of December, 20011 is the insurance replacement for the NIGCOMSAT-1 satellite with improved L-band payload. The NigComSat-1R Navigation (L-band) payload is a Navigation Overlay Service (NOS) similar to the European Geostationary Navigation Overlay Service (EGNOS) system. The system will augment the Global Navigation Satellite System (GNSS) over Africa as Nigerian Satellite Augmentation System (NSAS). The navigation payload of NigComSat-1R has been designed to support and operate in both the L1 and L5 frequencies. Furthermore, the piggyback payload uses a 10MHz ultra stable crystal oscillator to meet the performance requirements of frequency conversion stability and accuracy. The system functionality is identical and similar to European geostationary Navigation Overlay Service (EGNOS), where a number of ground reference stations monitor the GPS satellites’ signals and provide their observations to one or more Master Control stations (MCS). An augmentation message is then generated by the MCS and two (2) signals, C1 and C5, are transmitted via uplink stations within the uplink coverage areas on the C-band. The navigation payload down converts the C-band signals to L-band, L1 and L5, and retransmits these signals globally to users as  L1 and L5 exploiting a helical antenna with a backfire cavity (reflector) for global coverage from 42.5E orbital slot. (Footprint is almost half of the earth's surface with a bore sight angle of 8.7 degrees).  Nigerian Communications Satellite Ltd operates and manages the hybrid Communications Satellite (NIGCOMSAT-1R). http://www.nigcomsat.net. 

Where are the GNSS reference stations (ground segment) for this system ?

cezar 

The Master Control Stations (MCS) is located at the Abuja Headquarters of NIGCOMSAT Ltd, Federal Capital City of Nigeria. Six GNSS reference stations in six geo-political regions of Nigeria will be completed before the end of 1st Quarter of 2015  as pilot project. Another MCS is planned for North Africa after validation of the pilot project as proof-of-concept.

Dear Lasisi

Thanks for Your explanation. I am very impressed. Congratullations to Nigeria!

cezar

Dear Colleagues,

With what measure to express predictalbe, repeatable and relative accuracy?

rms, CEP, SEP or other ..?

cezar

Cezery,

All the the expressions you cite are useful since they are, in essence the same animal in different clothes. With each there is a most-probable or selected (mean) result and because the result of a measurement contains a degree of doubt, the expressions are means of expressing the degree of conformity or doubt to be assigned to the result.

The exact same data will have the exact same mean result. The quality or consistency of that data (precision) will be expressed as CEP, RMS, 2-sigma, 2DRMS. If the data set is in close agreement (precise) all of these will express lower numbers and the converse if the data is imprecise.

Since that is the case, any of them may be used in expressions of predictability, repeatability, and relative accuracies. The caveat is that in order to express these expected results it must be understood they are dependent upon certain constraints such as the systems must be alike, &c.

JAC

Dear JAC

Thanks for Your explanation. I think that this topic was finished. Thank You for Your Excellent Activity.

best regards

Cezary