GB2565106A

GB2565106A - Apparatus and method of locating a radio emitter

Info

Publication number: GB2565106A
Application number: GB1712408.2A
Authority: GB
Inventors: James Gear Benjamin
Original assignee: UK Secretary of State for Defence
Current assignee: UK Secretary of State for Defence
Priority date: 2017-08-02
Filing date: 2017-08-02
Publication date: 2019-02-06
Anticipated expiration: 2037-08-02
Also published as: GB2565106B; GB201712408D0

Abstract

A radio emitter locator comprises a phased array antenna 2 and a computer apparatus. The apparatus is arranged to receive first waveforms comprising a first group of signals from respective directions and to receive second waveforms comprising a second group of signals from respective directions. The first and second waveforms are processed to determine the angles of arrival of each signal. The angles of arrival of the first group of signals are compared with the angles of arrival of the second group of signals to determine the locations of a group of radio reflecting objects 4. The time at which each of the radio reflecting objects reflected its respective signal is then calculated. For each reflecting object a range of possible emitter locations is determined (see light-cones in fig 6, not shown). By combining generated probability distributions associated with each reflecting object, the most probable location of the radio emitter 1 is identified.

Description

The present invention relates to an apparatus and method of locating a wireless radio emitter based on reflections of the radio energy from the emitter, and is applicable to situations where there is no direct line of sight to the radio emitter. The invention is of use in various situations such as locating a skier's emergency transmitter in a mountainous terrain, or locating a person by their mobile phone under rubble after an earthquake.

Location of a radio emitter in the situation of No Line Of Sight (NLOS) is a difficult challenge. One way to estimate the location is to identify a number of directions which the relevant radio signal appears to be coming from, and to perform a weighted average of these directions. This provides an adequate estimate when the reflective materials are homogenous, however the natural environment, and especially the urban environment is highly non-homogenous, making this (and any similar approaches) highly inaccurate.

With such methods it is possible to determine the location of the emitter in the case that:

1. the various reflections are characterised by highly discrete reflecting point objects of known location,

2. the time of arrival of the signal from each of the reflecting point objects is precisely known, and

3. the density of obstructing objects is low such that any reflection can be assumed to be a direct, single reflection from the emitter only via the reflector and on to the receiver.

Unfortunately, those NLOS situations where the emitter is obstructed from view, tend to also be situations where the various reflections are not a simple number of discrete reflecting point objects, and also the density of obstructing objects means it is likely that a proportion of the identifiable reflections are unlikely to be single (direct) reflections.

Additionally, it is generally extremely difficult to accurately determine the location of the most clearly discrete reflections, because even with a reasonably sophisticated phased array antenna radio receiver it is difficult to very accurately determine the relative angle of the reflection source, and it is even more difficult to accurately determine the distance of each reflecting source.

Accordingly it is an object of the present invention to provide an apparatus or method of locating a radio emitter which can determine the location of a radio emitter in such challenging situations.

According to a first aspect of the present invention there is provided an apparatus as set out in claim

1. According to a second aspect of the present invention there is provided a method as set out in claim 7.

This has the advantage that when the apparatus is carried by a user or on a vehicle, and used while moving, the location of the emitter can typically be estimated with reasonable reliability and accuracy in a challenging NLOS environment such as an urban environment.

In essence, a radio receiver is used to identify locations of possible reflections from a radio source. By repeating this operation (after/while moving the receiver) the locations of the various reflecting objects can be identified, and based on the apparent time difference of arrival of the radio signal at each reflection source (rather than at the receiver), a probability distribution can be established comprising light cones extending from each identified reflection source location in space-time. The light cones are not geometrically planar but rather each surface has varying density distributed across a thickness such that it is diffuse. Each light cone represents a probability distribution and by multiplying these probability distributions together a total probability distribution for the location of the emitter is obtained, and the maximum value of this indicates the location of the radio emitter. Identifying the locations of reflection sources multiple times while moving the receiver, coupled with accounting for measurement errors to generate diffuse light cones, gives rise to a more reliable assessment of the emitter location (as compared to taking a single reading and/or treating each light cone as a substantially geometric (thin) cone and looking only for intersections of cones).

According to preferred embodiments of the invention, in determining each of the light cone distributions, account is preferably taken of a probabilistic distribution in an error in angle of arrival from the respective reflection source to the phased array antenna radio receiver. Indeed account preferably is taken of a probabilistic distribution in an error time in time of arrival from the respective reflection source to the phased array antenna radio receiver. Indeed account is preferably taken of a probabilistic distribution in an error in time of flight from the potential radio emitter to the respective reflection source. These have the advantage of providing a balance between maximising the accuracy with which the radio emitter can be located, and maximising the chance that the radio emitter is located. Preferably all three probabilistic distribution errors are taken into account.

According to a preferred embodiment of the invention the probabilistic distribution in the error in time of flight from the potential radio emitter to the respective reflection source is of substantially cropped exponential form. This provides an improvement in accuracy of locating the radio emitter. This takes into account the probability that the flight path includes second order reflections and the likely delays caused by such second order reflections, and enables the approach to locate emitters more accurately especially when the location is based on a larger number of primary reflections.

According to a preferred embodiment of the invention, in combining the plurality of light cone probability distributions to generate an emitter probability distribution, a mathematical operation is performed that is, or substantially is, multiplying the plurality of light cone probability distributions with one another. This has the advantage that detections based on a larger number of 'light cones' intersecting inaccurately is given more weight than a smaller number of 'light cones' intersecting accurately, to an appropriate degree, with the result that fewer false positive or missed detections should occur.

Further embodiments are set out in the claims.

A preferred embodiment of the invention will now be described by way of example only, with reference to the figures in which:

Figure 1 is an illustration of a bird-eye-view of an urban environment, showing a radio emitter and a receiver with no line of sight between them;

Figure 2 is an illustration of the urban environment showing a path of the receiver and three sets of signals received from reflections of the emitter, received at three different locations along the path;

Figure 3 is an illustration of how in principle a precise knowledge of the locations of the reflecting objects and a precise knowledge of the time of flight of the signal from the emitter to those reflecting objects would permit identification of the location of the emitter;

Figure 4 is an illustration of the urban environment showing the path of the receiver and various sets of signals received from various reflections of the emitter, received at nine different locations along the path;

Figure 5 is a diagram illustrating how by plotting the locations of three reflecting objects and extending light cones from them in the negative direction of the time axis these in principle intersect at a location which would in principle correspond to the location of the emitter;

Figure 6 is an extension of the diagram of figure 5 showing further light cones extending from the locations of the reflecting objects in the forward direction of the time axis, and illustrating how in principle these would intersect at a location corresponding to the receiver;

Figure 7 shows three probabilistic distributions, the first being a probabilistic distribution of the angle of arrival at the receiver, the second being a probabilistic distribution of the time of arrival of a signal at the receiver, and the third being a probabilistic distribution of the time of flight of a signal from the emitter to the reflector;

Figure 8 is an illustration showing how the signal may take a great number of possible paths from the emitter to each of three reflectors, showing how the most common paths are either direct or nearly direct and that highly indirect paths are possible but less common;

Figure 9 is an illustration of a side view of a probability distribution of the possible locations of an emitter, showing three light cones associated with three reflectors, which have a diffuse nature due to accounting for one or more errors in measurement;

Figure 10 is an illustration of a side view of a probability distribution showing a fourth individual probability distribution resulting from an incorrectly identified reflector location;

Figure 11 is an illustration of a combined density distribution where the intersection of the three individual probability distributions gives risk to a strong peak at the location of the radio emitter;

Figure 12 is a diagram showing the intersection of six light cones shown in cross section as circles, illustrating how failure to sufficiently accurately account for errors in the position and shape of the light cones and by adding instead of multiplying individual probability distributions it becomes more likely that an incorrect emitter location will be identified;

Figure 13 is a graph showing a modelled path of a receiver, a number of sources of reflection, and shows a probability distribution arising from the product of the various individual probability distributions associated with reflector;

Figure 14 is a second graph where in contrast to that of figure 10 the path of the receiver is longer, and as a result the overall probability distribution of the location of the emitter exhibits a narrower peak and its maximum is closer to the actual position of the emitter;

Figure 15 is a third graph where in contrast to that of figures 10 and 11 the path of the receiver is short, and as a result the peak of the probability distribution does not accurately reflect the location of the emitter; and

Figure 16 illustrates how a receiver moving along a path receives signals via reflecting objects (upper section), and this is measured and translated into light cone probability distributions (middle section), which in turn translates into identification of the emitter location (lower section).

With reference to figure 1, in some environments such as cluttered urban areas, it is common that a radio emitter needs to be located, but a user with a radio source locator may often find that there is no line of sight to that emitter and so the emitter cannot be located with any accuracy. Figure 1 shows such a situation where the emitter 1, emits radio transmissions which are largely or wholly obstructed or deflected by buildings. This is a simplified picture as it is not just building walls that obstruct or deflect radio transmissions but street furniture, vehicles, fencing and electrical wiring and white goods in buildings. A user has a phased array antenna radio receiver 2 which is located in a different area of the urban environment and lacks a direct line of sight path to the transmitter in order to locate it accurately. Moving within the urban environment may eventually lead to fleeting or sustained line of sight contact, but this is unreliable.

Figure 2 shows how by moving the phased array antenna radio receiver it is possible to triangulate the locations of various particularly reflective parts of buildings or other objects. In practice it has been found that complex non-homogenous environments tend to give rise to a moderate number of particularly reflective locations which reflect the radio energy towards the receiver strongly enough to stand out from other minor reflections which can then be treated as noise. In this example three building corners 4 happen to reflect radio energy towards the receiver particularly strongly, and when the user moves the phased array antenna radio receiver, these locations remain reasonably consistent. With further movement of the receiver some or all of these reflecting locations may cease to be particularly strong, and other reflecting locations may become dominant instead.

As the phased array antenna radio receiver is capable of determining the angle of arrival of these reflections, then by either maintaining the orientation of the phased array antenna radio receiver while moving (preferably in a straight line) or by monitoring and compensating for rotation in the orientation of the phased array antenna radio receiver it is possible to gather data on how the angle of arrival from each reflection varies with movement of the receiver. Figure 2 shows three locations on a path of movement of the receiver, along with incidence rays 5 (only one ray is labelled for clarity) from the various reflecting locations. By means of trigonometric calculations the locations of the reflecting locations can be determined to a degree of accuracy.

Having determined the locations of three reflecting locations their distance from the receiver is known. As the time of arrival of the signals from the reflecting locations is different it should be possible to identify a location from where a single emitter could give rise to those signals reflecting from those locations at those times. In particular in figure 3 on the left hand side it is shown that circles 6 can be drawn around the three reflecting locations, with the radii of the circles being different according to the difference in the time that the signal was reflected by each of the three reflecting locations. In this example on the left hand side there is no location where the three circles intersect. However as shown on the right hand side, by enlarging all three circles by the same amount it is possible to find a location where the circles do overlap 7, and provided that:

1. the position of the reflecting locations does not vary with the motion of the receiver,

2. the position of the emitter does not change while the receiver is moved,

3. the correct sources of radio signals have been identified as reflections,

4. accurate data on the locations of those reflectors is obtained, and

5. all reflections are solely direct (single) reflections of the emitter (rather than secondary reflections or more complex reflections), then the location of the emitter should be identifiable.

Furthermore, figure 4 shows how by moving the receiver along a longer path 8, a greater amount of data can be collected, regarding the location of the reflectors, more reflectors 4 may be identified, and the position of the reflectors can be determined more accurately too.

Figure 5 shows how this approach may be implemented using light cones. A light cone 9 is the locus of points through space (horizontal axes) and time (vertical axis) which the radio signals pass through, or - in this case - from which the radio waves may have come from in order to reach a reflector 4. Each light cone 9 is wider in the past (negative time direction) and converges on one of the reflectors 4.

For completeness, dashed arrows show how the radio signals pass from the reflectors 4 to the receiver 2. Where each of the two light cones intersect 10 this is shown as a thick line. Near the bottom of the diagram it can be seen that there is a region where the three cones intersect. This is also the location where each of the three intersects 10 themselves intersect (see enlarged view 11). Thus it is not necessary to iteratively try out different emitter times to see whether this gives rise to a match (as set out with respect to figure 3), but rather it is possible to evaluate the location of overlap in a 3 or 4 dimensional space-time (rather than just a 2 or 3 dimensional spatial region).

For completeness in figure 6 it is shown how by completing the light cones in the positive time direction 9' above each reflector 4, it should always be the case that the light cones similarly intersect at the location of the receiver (small circle).

However, while in principle this method should work, the inventor has found that in practice it generates acceptable results in only in hypothetical scenarios where the errors of measurement are small. In practice, with a typical portable phased array antenna radio receiver being moved a short distance through a dense urban environment, the measurement errors are so great that the light cones do not intersect.

Figure 7 shows illustrates three of the most significant errors, namely:

error in angle of arrival to the receiver which generally follows a normal distribution, error in the time of arrival of the signal from each reflector which equates to an error in the height of the respective light cone, and error in the time of flight between the emitter and the identified reflector which statistically is expected to follow an exponential distribution (a cropped exponential distribution) because while a direct path from the emitter to the reflector is possible, small deflections are very common and large or multiple deflections are rarer. This is illustrated in figure 8 where various signal paths 5' from the emitter 1 to the reflector 4 are shown.

By combining one or more of these errors when generating the light cones, this gives rise to probability distributions for each reflector, and it becomes easier to find an intersection of the individual probability distributions. This is illustrated in figure 9 where the three light cones of figure 5 are shown (as a pixelated needlepoint representation for printing and image reproduction purposes) as respective individual probability distributions 12.

Having evolved the light cones into probability distributions it becomes easier to find a match. However, this does not take into account that some apparent reflections may be incorrectly identified as such (false positives), and other apparent reflections might result from two different overlapping reflections leading to a gross error in an identified reflection position.

Figure 10 illustrates a situation where a wrongly identified reflector gives rise to an additional light cone which cannot be matched with the other light cones, however by using probability distributions it is nonetheless possible to identify the location of the maximum probability density, and this is not affected by falsely identified reflectors and is much less affected by errors in the position of the light cones. This is because, especially when a larger number of reflectors are identified (not illustrated for simplicity) the errors average out and the most probable location of the emitter can be identified.

To minimise the effect of one or more very inaccurately positioned probability distributions, the density of the probability distributions are multiplied together, rather than being added together. This means that the location where most or all of the light cones most closely intersect will stand out more strongly from other values. This effect is illustrated in figure 11 where due to the multiplicative effect the location of the emitter 1 corresponds to the strongest peak in overall probability density while a different intersection (the three light cones intersecting on the right hand side of figure 11) should give rise to a weaker peak even if those three light cones are better aligned.

As shown in figure 12, due to one of the light cones being measured particularly inaccurately (or indeed a radio source being wrongly identified as a reflection), three light cones (shown in section as circles) which intersect neatly (in this case the intersection is nearly perfect) could give rise to an incorrectly identified location 13, whereas the five light cones that intersect less precisely (at the correct location) should give rise to a stronger peak in total probability density. By correctly accounting for errors in the position, shape and distribution of the light cones, and by multiplying (rather than adding) the probability distributions the imperfect intersection of five light cones gives rise to a higher value than the combination of the three perfectly intersecting light cones. As a result the correct location is identified.

Referring to figures 13 to 15, probability distributions are shown that have been determined based on modelling of a scenario where an emitter 1 emits a radio signal which is reflected towards the receiver 2 via eight reflecting objects/locations/features 4. Unlike with figures 9 to 11 which are illustrative, figures 13 to 15 are the result of calculations from detailed modelling.

The receiver 2 is moved a distance along the path shown and based on the radio signals received a calculation is made of the locations of the reflectors 4 to estimate the location of the emitter 1. The method used involves accounting for errors in the angle of arrival at the receiver, the time of arrival of the signal at the receiver, and the time of flight between the emitter and the reflectors. The simulation shows that when the errors are accounted for and data is collected along a reasonably long path, then using a typical portable phase array antenna in a typical challenging environment, it should be possible to identify the location of the radio emitter despite the lack of a line of sight between the emitter and the receiver.

In figure 14 it is evident that by following a longer path, greater precision is achieved in locating the emitter.

By contrast in figure 15 the receiver follows only a very short path and as a result the accuracy with which the emitter is located is greatly reduced.

It should be noted that the strength of each individual probability distribution can advantageously be varied in accordance with a level of confidence that the apparent reflection is indeed a genuine reflection and not a noise artefact, however an alternative is to evaluate radio emissions from a range of directions and identify as a binary operation whether a particular source is a reflection or not. In either case this is done by comparing the signals to identify whether they correlate with either a reference or with each other.

It should also be noted that a 'light cone' is a mathematical concept relating to the spreading or converging of electromagnetic radiation in space-time and the invention does not require providing of a graphical representation of this. Instead the values of the individual and multiplied probability distributions can simply be evaluated to find a peak in the multiplied value. If however a representation were provided with two space dimensions and one time dimension then this presents a cone, while in three space dimensions and one time dimension it presents a hypercone (a hypercone is a sphere expanding or contracting over time).

Figure 16 provides a summary illustration whereby the top section illustrates how a receiver moving (top arrow) through a non-homogenous environment that lacks line of sight to an emitter, receives radar signals from various emitter locations. The middle section illustrates how the identification of the locations of reflective objects coupled with identification of the time that the signals were reflected at the various objects, gives rise to light cones identifying possible signal paths, which taking into account at least one measurement error gives rise to light cone probability distributions. The lower section illustrates how multiplying (or applying a similar operation) to the light cone probability distributions gives rise to an emitter location probability distribution with a peak corresponding to, or near to, the actual location of the emitter, thus identifying the location of the emitter.

Claims

Claims:

1. A radio emitter locator comprising:

A phased array antenna, and

A computer apparatus arranged to:

Receive waveforms using the phased array antenna radio receiver, each of the waveforms comprising a first group of signals from respective directions,

Process the waveforms to determine therefrom angles of arrival corresponding to each of the first group of signals,

Identify a most probable location of a radio emitter,

Characterised in that:

The computer apparatus is further arranged to:

Receive second waveforms using the phased array antenna radio receiver, each of the second waveforms comprising a second group of signals from respective directions,

Process the second waveforms to determine therefrom angles of arrival corresponding to each of the second group of signals, compare the angles of arrival of the first group of signals with the angles of arrival of the second group of signals to determine the locations of a respective group of radio reflecting objects, and determine, for at least one group of signals, based on the time of arrival of each of those signals from each radio reflecting object, the time that each of those radio reflecting objects reflected its respective signal, and;

The computer apparatus is further arranged to identify the most probable location of the radio emitter, by:

determining for each of the group of radio reflecting objects a light-cone of possible radio emitter locations, in each case based on the determined location of that radio reflecting object and the determined time that it reflected the respective signal;

generating for each light cone, a light cone probability distribution, based on at least one error in measurement thereof, and;

combining the plurality of light cone probability distributions to generate an emitter probability distribution, and;

identifying as the most probable location the maximum value of the emitter probability distribution.

2. A radio emitter locator according to claim 1 wherein:

in determining each of the light cone distributions, account is taken of a probabilistic distribution in an error in angle of arrival from the respective reflection source to the phased array antenna radio receiver.

3. A radio emitter locator according to claim 1 wherein:

in determining each of the light cone distributions, account is taken of a probabilistic distribution in an error time in time of arrival from the respective reflection source to the phased array antenna radio receiver.

4. A radio emitter locator according to claim 1 wherein:

in determining each of the light cone distributions, account is taken of a probabilistic distribution in an error in time of flight from the potential radio emitter to the respective reflection source.

5. A radio emitter locator according to claim 4 wherein;

The probabilistic distribution in the error in time of flight from the potential radio emitter to the respective reflection source is of substantially cropped exponential form.

6. A radio emitter locator according to any preceding claim wherein;

In combining the plurality of light cone probability distributions to generate an emitter probability distribution, a mathematical operation is performed that is, or substantially is, multiplying the plurality of light cone probability distributions with one another.

7. A method of locating a radio emitter comprising the steps of:

Providing a phased array antenna radio receiver, and

Controlling a computer apparatus to:

Identify a most probable location of a radio emitter,

Characterised in that:

The computer apparatus is further controlled to:

Receive second waveforms using the phased array antenna radio receiver, each of the waveforms comprising a second group of signals from respective directions,

The computer apparatus is further controlled to identify the most probable location of the radio emitter, by:

determining for each of the group of radio reflecting objects a light-cone of possible radio emitter locations, in each case based on the determined location of that radio reflecting object and the determined time that that respective radio reflecting object reflected the respective signal;