Shadow Matching: A New GNSS Positioning Technique for Urban Canyons
Summary (2 min read)
1. I N T R O D U C T I O N. Poor performance of Global Positioning System
- Section 2 presents modelling results to support this, showing how the availability in However, an urban canyon affects the geometry as well as the number of the available GNSS signals.
- Signals with lines of sight going across the street are much more likely to be blocked by buildings than signals with lines of sight going along the street.
- The shadow-matching positioning method reverses this by using the signal availability determined by the receiver to work out the user location.
- These include handling database errors, the effect of along-street position errors on shadow matching, reliable determination of whether a satellite is directly visible and efficient preparation, dissemination and storage of the building database.
2. A V A I L A B I L I T Y M O D E L L I N G.
- A series of simulations were run to determine the impact of GNSS constellation size and urban canyon aspect ratio on GNSS position solution availability.
- In two scenarios, the street was aligned north-south and in the remaining four, it was aligned east-west.
- For each aspect ratio, the satellite constellation was stepped through a 24 hour period in 60 s increments and the percentage availability of four or more direct LOS signals calculated.
- Within the east-west plane, a greater proportion of the satellites have lower elevations, whereas within the north-south plane, the distribution is more even, with a peak around 77x.
- Also, a good position solution may often be obtained from three direct LOS signals and a terrain height database.
3. C O N V E N T I O N A L P O S I T I O N I N
- The performance of multi-constellation GNSS positioning in urban canyons using conventional methods may be predicted by multiplying the average dilutions of precision (DOP) by an estimated user equivalent range error (UERE).
- Also, the path delay can be up to twice as large where the direct LOS signal is received through a gap between buildings.
- Values are omitted for solution availabilities below 10% to avoid misleading outlier effects.
- 5 because the ionosphere, troposphere and multipath error standard deviations will be larger than assumed in the UERE estimate.
- As can be seen from the figures, the cross-street positioning accuracy is slightly poorer than its along-street counterpart for the north-south street alignment and substantially poorer for the east-west street alignment.
4. A S I M P L E S H A D O W -M A T C H I N G A L G O R I T H M.
- The principle of shadow matching is simple.
- Note that satellites that are either directly visible throughout the street or obstructed throughout the street do not contribute positional information.
- Figure 8 depicts a simple shadow-matching algorithm.
- To determine the performance of this shadow matching algorithm, it was tested using the 100-satellite constellation over each of the six scenarios used for determining urban canyon position solution availability and conventional positioning performance in the preceding sections.
- Position containment is the difference between the maximum and minimum cross-street position as determined by the shadow matching process.
Boundaries of all-satellite localisation region
- Figure 9 (left) shows the root mean square (RMS) cross-street position error using shadow matching as a function of street aspect ratio.
- Performance in the car scenarios was poor for the lowest aspect ratio of 0 .
- 1 because the GNSS shadows did not extend to the middle of the street.
- The results presented here represent a lower bound on the expected accuracy of a practical shadow-matching system.
- Additional position errors will occur due to errors in the 3D city model, the conventional GNSS position solution used to select the correct region of the model and the ability of the GNSS user equipment to determine whether a received signal is direct line of sight.
5. P R A C T I C A L S H A D O W M A T C H I N G. A practical implementation of shadow matching requires:
- 3D city models are now available from a range of suppliers for cities all around the world.
- As a strongly reflected signal will have left-handed circular polarisation (LHCP), a receive antenna with a good polarisation discrimination will help reduce its C/N 0 .
- Cross-street errors in the positions of buildings in the database directly translate into errors in the boundary between the receivable and non-receivable regions.
- It has been shown, using mathematical modelling, that although combining GPS with other GNSS significantly increases signal availability in urban canyons, the positioning accuracy in the cross-street direction will generally be poor because the unobstructed satellite signals travel along the street, rather than across it.
- Issues affecting a practical implementation of shadow matching, such as database dissemination, satellite visibility determination and handling of database errors, have been explored and mitigating action proposed.
Did you find this useful? Give us your feedback
Citations
262 citations
175 citations
Cites methods from "Shadow Matching: A New GNSS Positio..."
..., 2012) and the generation of an intelligent urban positioning solution by combining augmented conventional positioning with the shadow-matching technique (Groves, 2011; Wang et al., 2012)....
[...]
...…using a 3D city model, when the user position is only approximately known, as discussed in (Groves et al., 2012) and the generation of an intelligent urban positioning solution by combining augmented conventional positioning with the shadow-matching technique (Groves, 2011; Wang et al., 2012)....
[...]
150 citations
Cites background or methods from "Shadow Matching: A New GNSS Positio..."
...Another solution is GNSS shadow matching, which can potentially improve the across-street positioning accuracy by comparing the observed GNSS signal availability with that predicted using a 3-D city model (Groves, 2011; Wang et al., 2011; Groves et al., 2012)....
[...]
...D city model (Groves, 2011; Wang et al., 2011; Groves et al., 2012)....
[...]
127 citations
116 citations
Cites background or methods from "Shadow Matching: A New GNSS Positio..."
...Consequently, the signal geometry, and hence the positioning accuracy, is much better along the direction of the street than across the street (Groves, 2011)....
[...]
...3 Figure 2: The shadow-matching concept: using direct signal reception to localise position (Groves, 2011)....
[...]
...Shadow matching is a new positioning technique using GNSS, assisted by knowledge derived from 3D city models, that has the potential to provide metres-level cross-street accuracy in urban canyons (Groves, 2011; Tiberius and Verbree, 2004)....
[...]
References
4,475 citations
1,351 citations
"Shadow Matching: A New GNSS Positio..." refers background or methods in this paper
...Along-street and crossstreet DOP were then calculated using the conventional method [9, 10]....
[...]
...These compare quantities calculated from different combinations of measurements to determine whether they are consistent [9]....
[...]
...Finally, dead reckoning [9] could be used to bridge a navigation solution between intersections, though the accuracy will depend on sensor quality and environment....
[...]
...The UERE estimation was based on the following assumptions [9] :...
[...]
407 citations
77 citations
Related Papers (5)
Frequently Asked Questions (9)
Q2. What is the way to bridge a navigation solution between intersections?
dead reckoning [9] could be used to bridge a navigation solution between intersections, though the accuracy will depend on sensor quality and environment.
Q3. What is the practical implementation of shadow matching?
A practical implementation of shadow matching requires:’ Efficient preparation, dissemination and storage of the city models ; ’ Reliable determination of whether a satellite is directly visible ; and ’ Quantification and mitigation of errors.
Q4. What is the way to mitigate along-street errors?
One way of mitigating these along-street errors is to repeat the shadow matching calculation for a number of different along-street positions, applying a consistency check in each case.
Q5. What is the scaling factor for the shadow matching algorithm?
where only one path is blocked, an erroneous shadow match will occur, whereby the shadow matching algorithm assumes a blocked signal when the true signal is available or an available signal when the true signal is blocked.
Q6. How many street widths from the edge was the receiver located in the simulations?
In half the scenarios, the receiver was located 0.1 street widths from the edge, representing a pedestrian; in the remaining scenarios, it was located 0.35 street widths from the edge, representing a car.
Q7. What is the way to determine the C/N0 of a satellite?
For applications where the antenna is kept (approximately) horizontal, the angle of incidence at the antenna will be known, so antenna gain calibration can be used to determine the expected C/N0.
Q8. How was the performance of multi-constellation GNSS in urban canyons predicted?
The performance of multi-constellation GNSS positioning in urban canyons using conventional methods may be predicted by multiplying the average dilutions of precision (DOP) by an estimated user equivalent range error (UERE).
Q9. Why are the values omitted for building height to street width?
They are also omitted for building height to street width aspect ratios below 0.5 because the ionosphere, troposphere and multipath error standard deviations will be larger than assumed in the UERE estimate.