FIELD OF THE INVENTION

This invention relates to methods and apparatus for determining the position of transmitters, particularly but not exclusively interference transmitters, in the uplink of a satellite communications system, particularly but not exclusively a geostationary satellite communication system. BACKGROUND

Satellite communications systems comprise at least one satellite, at least one terminal (which may be a portable terminal) and at least one ground station. An example is the present applicant's Broadband Global Area Network (BGAN) system, carrying two-way voice, circuit- and packet- switched data. Each geostationary satellite acts as a bi-directional relay between a terminal and a terrestrial Satellite Access Stations (SAS), interconnecting a terminal uplink channel which receives signals from a terminal with a SAS downlink channel which transmits said signals to the SAS in the from-terminal (hereafter "return") direction, and an earth station uplink channel with a terminal downlink channel in the to-terminal (hereafter "forward") direction.

Satellite user terminals often use relatively low transmit powers, and the terminal uplink channel is therefore susceptible to interference from any source of co-channel radio frequency emissions elsewhere on the Earth's surface within the relatively broad coverage area of the satellite. This problem is not unique to the applicant's satellite communications systems.

Such interference can be mitigated with knowledge of the interference signal, and if the positions of the interferer and the terminal are known, the interference can be mitigated more effectively. Some prior interference cancellation techniques are described in EP1006678, EP1006679, EP1035664, WO0035125, WO0049735, and WO0048333.

Our earlier patent EP0843918 discloses a method of determining in which of a plurality of satellite spot beams a user terminal is located. However, it does not locate the terminal within that beam. Nowadays, determining the positions of user terminals is straightforward since such terminals almost invariably comprise satellite positioning system receivers (e.g. GPS, Glonass, Compass or Galileo receivers) and are therefore able to self-report their precise positions. However, interference signals are, by definition, not in communication with the satellite system. Ranging and Doppler techniques for interference location rely on relative motion of two (or more) satellites which provide coverage in the relevant region at the relevant frequency. It may be the case that suitable satellites either do not exist or the user does not have access to them. Further, systems which employ this technique must monitor (and possibly configure) multiple satellites and are therefore inherently more expensive, error prone and time consuming to operate.

The present invention therefore aims to provide improved methods and apparatus for determining the position within an uplink spotbeam of transmitters (such as interference sources) using a single satellite of a communications system (typically geostationary, though use in non- geostationary systems is not excluded).

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings in which:

Figure 1 is a schematic diagram of a satellite communications system incorporating an interference geolocation system according to an embodiment of the present invention;

Figure 2 is a diagram showing an exemplary beam pattern produced by a satellite of Figure l ;

Figure 3 is a schematic diagram of a satellite payload for use in the embodiment;

Figure 4 is a schematic diagram illustrating spot beam overlap and gain profiles in the spot beam pattern of Figure 3;

Figure 5 is a schematic block diagram of an interference location station according to the embodiment of Figure 1 ;

Figure 6 is a flow diagram illustrating the process of a first embodiment of the invention;

Figure 7 is a flow diagram illustrating the additional steps of the process of a second embodiment of the invention usable with the first;

Figure 8 is a flow diagram illustrating the additional steps of the process of a third embodiment of the invention usable with the first or second embodiments; and

Figure 9 is a flow diagram illustrating the additional steps of the process of a fourth embodiment of the invention usable with the first, second or third embodiments. DESCRIPTION OF THE PREFERRED EMBODIMENTS

The system in which the invention is used includes three geostationary Inmarsat-4™ satellites 2, one of which is shown in Figure 1. "Geostationary" here includes satellites which are in geosynchronous 24 hour orbits but exhibit some movement above and below the equator. Each satellite 2 generates a large number of spot beams 8 spanning the entire coverage area of its global beam 6 (although some are omitted for clarity from Figure 1). The spot beams 8a-e are used predominantly for communications traffic, while the global beam 6 is used predominantly for call set-up signalling. Also provided are a set of 19 regional beams 7 (although only one is shown for clarity in Figure 1). These also collectively cover the same area as the spot beams 8 and the global beam 6. They allow terminals to operate with smaller antennas than would be required when using global beams.

The spot beams 8 are arranged in an approximately hexagonal beam pattern covering the portion of the surface of the Earth which is within the field of view of the satellite 2. The footprint of each beam is represented as having a conventional cellular hexagonal shape, for clarity, but in reality will have an approximately circular section distorted by the projection of the beam onto the surface of the earth, as shown in Figure 2, the transmission and reception beam patterns being substantially reciprocal. A frequency re-use pattern is applied so that the same frequency channel is re-used between spot beams 8 having at least a minimum separation distance. As will be discussed, the beam footprints overlap significantly.

A preferred embodiment of the invention is implemented in the present applicant's BGAN satellite communications system. In this system, the user terminals 5 can be as small as a laptop computer and some are therefore portable. The SAS is connected to other SASs via a high speed data network (e.g. of dedicated leased lines), and also to Public Switched Telephone Networks (PSTNs), mobile networks (PLMSs), Integrated Services Digital Networks (ISDNs), and the Internet. The system is therefore able to provide voice, Standard/Streaming IP or ISDN traffic in either direction between the terminals and terrestrial networks.

For each satellite 2, a plurality of SASs 4a-b act as satellite base stations and gateways to terrestrial networks. Each SAS 4 communicates at C-band over a bidirectional feeder link 10 via the satellite 2. Each further comprises conventional modulators and demodulators, and a control system comprising a processor arranged to allocate channels and to perform other signalling to and from the terminals. The spot beams 8 carry user traffic and signalling for reception by user terminals 5. A single user terminal 5 is shown in Figure 1, but the system is able to provide satellite communications services to a number of such terminals in each spot beam. Each terminal is conventional and comprises an antenna, a radio frequency transceiver, a GPS receiver, and a signalling processor and is connectable to input and output devices such as mobile telephones, and personal computers.

Figure 3 shows a satellite 2 payload. In the to-mobile direction it comprises a C-band receive antenna 14 which receives feeder link channel transmissions from the LES 4, which are amplified by a low noise amplifier (LNA) 16 and digitally converted by an analog-to-digital converter (ADC) 18. A digital beam former and channeliser (shown as two parts 20, 120 for the to-mobile and from-mobile directions respectively) performs the mapping between feeder link channels and frequency channels in spot beams 8, and outputs a set of analog converted signals which are amplified by a bank of high power amplifiers (HPAs) 22 and fed to a corresponding set of transmit antenna elements in an L-band transmit antenna 24. The sum of signals radiated by the antenna elements generates the downlink spot beam pattern. In the return direction, the feed elements of an L-band receive antenna 124 (in this embodiment, 120 elements) are connected via a bank of amplifiers 122 (and ADCs, not shown for clarity) to the digital beam former and channeliser 20b which combines the feed inputs into beams, performs the mapping between frequency channels in spot beams 8 and feeder link channels, and outputs a digital signal which is converted to analog by an analog-to-digital convertor (DAC) 118, amplified by a low noise amplifier (LNA) 116 and fed to a C-band transmit antenna 114 which transmits downlink channel transmissions to the LES 4.

The frequency channels are each 100 kHz in bandwidth, in the uplink and downlink feeder and mobile links. Each frequency channel may be subdivided in frequency and time to define individual user channels, with differing modulation schemes.

Each channel received in each spot beam 8 at L-Band is mapped by the satellite 2 to a corresponding channel at C band in the feeder link according to a variable channel mapping determined by the configuration on board the satellite 2. Each satellite is therefore essentially a "bent-pipe" transponder which amplifies and frequency-translates between the L band and the C band and vice versa. Each mapping is referred to hereafter as a "filter", and is defined by an uplink centre frequency, a downlink centre frequency, a beam and a direction (forward or return). In the following embodiment, only the return filters are utilised. The filter configuration (i.e. the set of mappings) is controlled by commands received from a SAS 4 by a TT&C antenna 26, demodulated and decoded by a TT&C interface 28 and provided as input to the digital beam former and channeliser 20, as described in greater detail in our earlier US Patent 7792485.

As shown in Figure 4, there is a substantial degree of overlap between adjacent L-band

(mobile) uplink beams 8a, 8b. At all positions on the Earth (except at the Poles), a single terrestrial transmitter will be received not just in one spot beam but in each of a first ring of nearest-neighbours to that beam, and perhaps in a further ring outside the first.

The gain of each spot beam 8 varies significantly across the surface of the earth, with the peak gain (boresite) falling within the coverage area of the beam, with the result that the gains of the beam 8 at each of a grid of points within the coverage area of the beam on the surface of the Earth vary quite significantly.

The grid of points on the Earth's surface corresponds to a grid of beam directions defined by pairs of azimuth and elevation point values within the downlink antenna pattern.

Each satellite 2 is typically not absolutely geostationary but moves over time. The satellite position relative to the Earth is accurately known at all times, enabling the beams to be coarsely electronically steered to maintain the same centre points on the surface of the Earth, but as this steering is imperfect (and as the reflector antenna may slightly change shape over a day), the gains at a given set of points on the surface of the Earth vary likewise over time.

Thus, for a given fixed grid of points on the surface of the Earth, or fixed grid of azimuth/elevation points in the beam pattern, such as those indicated in Fig 4, the gain at that point of each of a plurality of beams 8a, 8b ... can be calculated from knowledge of the beam shapes and the instantaneous satellite position (i.e. its orbital ephemera).

Referring to Figure 5, an interference locator device 100 is illustrated. For convenience it is co-located with the SAS 4, but it could alternatively be remote as referred to in our earlier US Patent 7792485. It comprises a spectrum analyser 30, a controller 34 and a database 38 which stores the satellite filter configuration and controls the satellite 2 to modify the satellite filter configuration through a payload control system 36 which communicates the modifications to the satellite 2 via the SAS 4.

Referring to Figure 6, the process of geolocation performed in the preferred embodiment will now be described.

In a step 202, repeated periodically (of the order of every second) the controller 34 determines, via the analysis (step 201) by the spectrum analyser 30, whether an interference source is present in the global or any of the regional uplink beams of the return signal and, if so, determines its frequency.

If so, then in a step 204, the controller 34 temporarily sets up filters encompassing the frequency of the interference source (in the return direction) for each of the regional beams 7, and determines within which of the regional beams the interference is most strongly received (step 205). In a step 206, the controller 34 sets up filters (in the return direction) for each of the set of uplink spot beams 8 within the relevant regional beam 7, encompassing the frequency of the interference source. In each case, the feeder downlink C-band frequencies are made spectrally adjacent so as to minimise any frequency-dependent effects.

In a step 208, the spectrum analyser 30 measures the signal strength of the interference signal within each of the spot beams of the set, to give an interference power vector R of received signal powers in each of these spot beams.

Next, in step 210, the controller 34 calculates, for each of a set N of azimuth/elevation positions P in the beam pattern (corresponding to positions within the relevant regional beam 7 when projected on the Earth' s surface), the gain of each of the set of uplink spot beams 8, from stored data representing the gain patterns of the beams in azimuth and elevation. The result is therefore a grid of gain vectors G, one for each point Pi, P ₂ , ... Pi ... PN, each vector consisting of a set of gain values, one for each of the beams of the set. At least three spot beams are required. However in practice, the vector may consist of values for the nearest 7 beams (the interference centre beam, plus a ring of 6 nearest neighbours in the beam array).

Next, in step 212, for each point Pi of the total of P points, the controller 34 compares the vector R of received signal powers with the gain vector G. This may be, for example, by calculating the "power offset" as the difference at that point Pi in directivity (in dB) of each beam and received power (in dBW) in each beam. If the difference between each of the respective components of R and one gain vector Gi for a particular point Pi is the same, then (as the differences in db/dbW correspond to ratios of signal power), the received signal strengths in each beam are in the same mutual ratios to the calculated gains for the beams at that particular point. Thus, the variance of R - Gi is a measure of how close the interferer position is to the point Pi.

Next, in step 214, the controller 34 selects that point Pi _nt having the gain vector Gi _nt most similar to interference power vector R (e.g. that with the lowest variance in power offset).

Finally, in step 216, the controller 34 translates the point Pi _nt (which is defined in azimuth/elevation within the beam pattern) into a position on the Earth taking into account the current position and orientation of the satellite 2. This is then used as the position of the interference source.

Having completed the measurement, the controller 34 releases the allocated frequency channel mappings.

It will be seen that the initial step of locating the regional beam in which the interference signal is detected with the highest power gives a coarse indication of position. The detection of the spot beam in which the interference signal is detected with the highest power would give a finer indication. However, using multiple spot beam power measurements, and more preferably matching the set of received powers with the beam gains at each of a plurality of points, allows positioning within this spot beam to give a much finer positioning technique.

Yet further accuracy, however, can be gained in preferred embodiments, in several ways.

CALIBRATION EMBODIMENT

Firstly, referring to Figure 7, in a step 220, the controller 34 detects, from the reported positions of the terminals 5 (each of which contains a GPS receiver and periodically signals its position to the network via the satellite), the terminal 5a closest to the point Pm _t . In step 226 (as in step 208), the controller 34 then sets up filters (in the return direction) for each of the set of uplink spot beams 8 within the relevant regional beam 7, encompassing the transmit frequency of the selected terminal 5a. In steps 228-236 (as in steps 208-216), the controller 34 measures the vector R of received signal powers, calculates the grid of gain values G for the terminal 5a, compares it with the vector R of received signal powers in different beams for the terminal 5a, and determines the closest grid point P _T on the basis of the comparison - in other words, determines the position of the terminal 5a in exactly the same manner as for the interference source.

In step 238, the controller 34 calculates the offset vector between the terrestrial position of the point P _T and the self-reported position of the terminal 5a determined by the terminal GPS. If the distance between the two is greater than the distance between adjacent grid points, then in step 240, the controller adjusts the terrestrial position of the point Pi (and preferably all other grid points) by the offset vector, so as to calibrate the points P _.

INTERPOLATION EMBODIMENT

Secondly, referring to Figure 8, in a further embodiment, the controller 34 considers not only the "closest" point Pi _nt but the surrounding 8 points (for a square grid). Thus, following step 216, in a step 248, the controller 34 selects the 8 points surrounding Pi _nt; in step 250, calculates a grid of intervening azimuth/elevation points interpolated (either bi-linearly or bi-cubically) from these known 9 points; and calculates additional gain vectors for each such interpolated point. There may, for example, be 16 additional points equidistant between the original 9, to total 25. This further localises the position of the interference source. The remaining process of Figure 6 is then performed from step 216.

CORRELATION EMBODIMENT

In some cases, the signal received in some of the spot beams may be very weak, or the background noise power may be high. In this case, referring to Figure 9, in a preferred embodiment, a correlation based technique is used to reduce noise within the relative power measurements.

Since the interference signal in each beam originates from a single source, and travels the same path, the images within the different uplink beams are strongly correlated. Thus, in step 250, the beam in which the interference signal is most strongly detected is identified. In step 252, the signals from the other beams are then cross -correlated with the strongest, to extract the interference signal component from other signals and noise, even at low signal-to-noise ratios.

The peaks of the cross-correlation functions thus calculated are determined in step 254, and then the power measurements are calculated therefrom in step 256.

Given:

S(t) - is the signal in the time domain,

Si(t) - refers to the strongest image of the signal from one of the beams;

S ₂ (t), S ₃ (t), ... - refers to the second strongest image, the third strongest image, etc.;

peak(Sn,Sm) - is the peak of the cross correlation function applied to Sn(t) and Sm(t); ¾ is the power measurement (in W) of the nth strongest image of the signal; then:

ft (2) INTERFERENCE MITIGATION

Having determined the position of the interferor according to the present invention, various measures may be taken to utilise the position to reduce the effect of the interference. Although these do not form part of the disclosed technique of transmitter geolocation, a brief description of some examples will be given for completeness of understanding.

For example, firstly, the interference transmitter location sometimes corresponds to a known friendly facility e.g. an airfield, an oil rig or a ship. Thus, the controller can signal to take action to turn off the interference.

Secondly, in allocating frequency channels, user terminals 5 close to the position of the interferor can be allocated channels on different frequencies, with the frequency of the interference source being re-used elsewhere in the uplink beam pattern far enough away so as to reduce interference.

Thirdly, the if the interference cannot otherwise be mitigated, the beam patterns produced by the uplink antenna can be re-formed so as to locate a null at the interference position.

OTHER EMBODIMENTS

In some cases it may be desirable to reduce the domain (time and frequency) over which the interference power measurement is taken. This is more desirable the closer in frequency the interferer (the wanted signal) is to the traffic being carried (the unwanted signal). This can be achieved either:

• Manually by an operator analysing a time/frequency plot of the signal and selecting the domain or;

• Automatically, by the system analysing bursts and excluding those whose modulation type, symbol rate etc. fit a given profile and are therefore likely to be traffic.

Whilst operation using an Inmarsat-4 L-band satellite has been described, many alternatives are possible. It is proposed to add Inmarsat 5 satellite operating in the Ka band, with terminals 5 being able to use either. If regional beams are unavailable (e.g. on an Inmarsat 5 satellite), the preliminary search discussed above can be carried out using all narrow beams, or a subset of narrow beams spread across the satellite coverage. As explained above, the problems addressed by the invention are not confined to Inmarsat™ satellites, geostationary satellites or repeater satellites. Whilst an interference signal source has been described, the same technique could be used to position other transmitters where, for some reason, self-reporting of position is not suitable.

Whilst use of a self-reported mobile terminal position for calibration is described, a fixed transmitter of fixed position, known a priori, could instead be used.

Whilst calculating gain values at an array of points is described, it will be apparent that the points could be defined on the Earth's surface rather than as azimuth/elevation patterns (although the latter is preferred where a satellite is in motion as it reduces the number of calculations required). Knowledge of the beam gain patterns could also be used in other fashions than comparison with an array of precalulated points.

Whilst use of the variance between a pair of logarithmic vectors is described, many mathematical equivalents which have the effect of comparing the received set of signal strengths with the calculated or predicted gain patterns can be envisaged.

Whilst interpolation by adding extra points in a square array is described above, triangular or hexagonal arrays could be used instead. The numbers of beams mentioned are purely exemplary.

Many alternative embodiments may be envisaged, which nevertheless fall within the scope of the invention as defined by the claims. It is understood that protection is sought hereby for any and all novel subject matter and combinations thereof disclosed herein. Unless the contrary is stated, the invention is that features of each embodiment may be combined with features of any other.