GB2521092A

GB2521092A - Radio systems

Info

Publication number: GB2521092A
Application number: GB8615564.5A
Authority: GB
Inventors: Simon A B Avarne
Original assignee: Racal Research Ltd; Racal Ses Ltd
Current assignee: Racal Research Ltd; Racal Ses Ltd
Priority date: 1986-06-25
Filing date: 1986-06-25
Publication date: 2015-06-17
Anticipated expiration: 2006-06-25
Also published as: GB2521092B; GB8615564D0

Abstract

A frequency hopping MFSK radio system which transmits one symbol on each channel is arranged to transmit a header control block encoded using a powerful error correcting code on all available channels before message transmission commences. This header control block contains a sequence of flags identifying a subset of the available channels which are least occupied. The information from which the flags are set is contained in a probability table which is updated from at least one of the following: the output of a receiver decoder for decoding an error correcting code applied to the data to be transmitted relating to the presence and channel position of errors detected in received data; channel spectrum information derived from an MFSK demodulator at the receiver; and the setting of an automatic gain control device at the receiver.

Description

II

RADIO SYSTEMS

The present invention relates to radio systems, and, more particularly, to radio systems using multiple frequency shift keying (MFSK) and frequency hopping.

A frequency hopping MFSK radio system is described in GB-B-2 129 655. In this radio system one of a number of possible fixed tones is transmitted relative to a base frequency in dependence on a symbol of data to be transmitted. The base frequency is hopped pseudo randomly amongst a number of available cha.nnels with a hop period which is equal to the symbol period.

Such a radio system has very good resistance to recognition, tracking, direction finding or jamming so that, overall, there is a very low probability of intercept (LPI) for such a radio system.

A fundamental technical problem with such a frequency hopping MFSK radio system arises out of the relatively high level of channel occupancy that is normally encountered when such a radio system is operated in the HF band. A channel is said to be occupied if the level of noise or interference is such that a transmitted tone cannot be reliably decoded by the receiver. In order to be effective, a a radio system must be capable of tolerating o-ccupancy levels of 40 -50%.

To some extent this problem can be solved by encoding the data to be transmitted using a forward error correcting code. Because an MFSK radio system transmits symbols, a Reed Solomon (RB) code is particularly suitable for this type of radio system.

Such codes are well known. Reference may be made to Reed 1.5. and Solomon G. "Polynomial Codes over Certain Finite Fields" J.Soc. md. Appl. Math. Vol 8.

pp 300-304 June 1960 for a discussion of their definition and properties. For maximum efficiency, two such codes are concatenated, that is, the data is first encoded with an outer code, and the resulting encoded data is encoded again with an inner code.

The output from the encoder which applies the inner code is then transmitted.

Despite the use of forward error correction codes, the maximum information rate as computed by Shannon's Fundamental Theorem is still reduced to unacceptably low levels when the channel occupancy is 50%. In practice, using the best known error correcting codes, an information rate of about 5Bauds could only be achieved at a 50% occupancy level. Since this

I

implies that less than one 7 bit character per second can be transmitted, such a rate is unacceptably low the for majority of applications.

The present invention is directed to increasing the overall information rate of a frequencyhopping MFSK system.

The present Invention accordingly provides a frequency hopping MFSK radio system comprising a transmitter unit and a receiver unit; the transmitter unit comprising encoding means for encoding data to be transmitted using at least one forward error correcting code and outputting encoded symbols to be transmitted, an MFSK modulator for converting each of the encoded symbols into a tone, and a frequency hopping transmitter for transmitting the tones and comprising a frequency synthesiser, and means for controlling the synthesiser to produce for each successive symbol period a frequency representing a channel frequency plus the tone representing the symbol to be transmitted during that symbol period, each symbol period being equal to a hop period for which the channel frequency is constant; the receiver unit comprising a frequency hopping receiver for outputting received tones, an I4FSK demodulator for & outputting symbols from the received tones, and decoding means for decoding the symbols and outputting the transmitted data and information relating to the presence and channel location of errors detected in the transmission, the receiver unit further comprising control means comprising means for storing information for each of the channels available to the system, which information represents a probability that that channel is occupied, and means for updating said probability information at least from said information output from the decoder, and means for periodically transmitting back to the transmitter on all channels, data relating to said stored probabilities, the transmitter unit comprising means for selecting a subset of the available channels for use during subsequent transmission in response to the received data from the receiver unit.

In this manner, the channels which are most likely to be unoccupied are used for transmission of data. All channels are used for transmission of the data relating to the channel occupancy. This allows infbrthation to be accumulated on channels which were previously considered blocked and the!efore, unused.

In order that the channel occupancy information can t be received accurately, it is sent at a very low information rate with a powerful forward error correcting code applied to it. Although this portion of the data is transmitted at a very low information rate, it enables a higher information rate to be used subsequently because the channel occupancy level in those channels which are used by the system will ideally be zero and generally be much lower than the channel occupancy level when all available channels are considered.

In a preferred embodiment, the demodulator at the receiver unit outputs information relating to the entire spectrum received on the channel in addition to the demodulated symbol, said updating means updating said probability information from this spectrum information in addition to said information output from the decoder. Preferably the said channel spectrum information relates to the position of the peak received signal in the frequency domain relative to the nearest valid tone frequency. Where the channel is relatively unoccupied, this peak will lie on a valid tone frequency. However, in the case of an occupied channel, the peak may occur anywhere within the spectrum with a substantially flat probability distribution function. Therefore, t S depending on the position of the peak, the probability that the channel is occupied can be assessed.

In one alternative embodiment, the receiver comprises an automatic gain control device, and the gain setting of this device is fed to said updating means.

Since a channel with a strong interfering signal will require less gain than an unoccupied channel, the setting of the gain can also be used to provide an assessment of the probability that the channel is occupied.

More specifically, the present invention provides a frequency bopping MFSK radio system in which data is transmitted only on a selected subset of the available channels, the radio system comprising meafls for selecting said subset in dependence on the contents of a probability table which stores information for each channel relating to the probability that that channel is occupied, the probability table being updated from information derived from at least one of the following: the output of a receiver decoder for decoding an error correcting code applied to the data to be & transmitted relating to the presence and channel position of errors detected in received data; channel spectrum information derived from an MFSK demodulator at the receiver; and the setting of an automatic gain control device at the receiver.

A frequency hopping MFSK radio system embodying the invention will now be described, by way of example only, with reference to the accompanying diagrammatic drawings, in which: Figure 1 is a block diagram of a radio station including a transmitter unit and a receiver unit; and Figure 2 is a block diagram showing a control unit of the radio station shown in Figure 1 in more detail.

A frequency hopping MFSK radio system use.s radio stations of the type illustrated in Figure 1. Each radio station comprises a transmitter unit 2 and a receiver unit 4 which operate under the control of a control unit 6. & S

Input message data in digital form is fed on line 8 to a data packer 10 of the transmitter unit. The data packer assembles the message data bits into blocks of an appropriate length to be handled by the encoder 12. The data packer 10 also receives an input of digital control data on line 14 from the control unit 6. Data received on the input 14 is sent at the beginning of each transmission as a header block and also at predetermined fixed intervals during message transmission. The contents of the header block will be described in more detail below. The output of the data packer 10 is fed to the encoder 12 which encodes the data in accordance with a forward error correcting code strategy determined by the control unit 6. The most powerful error correcting code strategy is always used for transmitting the header block but the strategy used for transmitting the message data is variable. For example, the encoder 12 may first encode the input data using an outer code which in this example is always the truncated RB (240,210) code. The encoder 12-then encodes the output of the outer code with an inner code. The inner code is selected from the following group of Reed Solomon codes: RB (15, 2), RB (15, 4), RB (15, 6), RB (15, 8), RB (15, 10) and -RS (15, 12). The particular inner code to be used is determined by the input applied to the encoder 12 on line 16 from the control unit 6. Since -RS codes have been used, the output of the encoder 12 is in the form of a sequence of symbols, which are fed to an MFSK modulator 18. The modulator 18 outputs a predetermined one of' sixteen available tones in dependence on the symbol to be transmitted. It will be appreciated that modulators using more or less tones can be employed. A sixteen tone system can transmit a four bit symbol on each tone. If the number of tones is increased, then the data rate will increase. However, this is of little use as increasing the number of tones increases the bandwidth of each channel and so increases the occupancy level. -The tones output from the modulator 18 are fed to transmitter 20. Transmitter 20 is a conventional frequency hopping radio transmitter, which transmits one symbol during each hop period. Therefore, each new tone is transmitted on a different carrier frequency. The carrier or channel frequencies are pseudo randomly changed in accordance with a predetermined sequence. Typically 256 channels are available to the system, of which only a subset made up of the least occupied 128 are used in this example. The channels to make up this subset of' 128 channels are selected by the control unit 6 in response to the header block received from another radio station in a manner to be described in more detail below. Information concerning the subset of channels to be used is fed to the transmitter 20 on a control line 22 from the control unit 6.

Synchronisation between the transmitter 20 and the receiver of the other radio station is achieved by any suitable means known in the art. For example, synchronisation can be carried out by the method described in our co-pending Patent Application No. 86.08328.

The receiver unit 4 of the radio station comprises a frequency hopping receiver 24 which receives the incoming signal and passes it to an MFSK demodulator 26. In this example the receiver 24 has an automatic gain control device 25, which sets the output power level from the receiver to a constant value regardless of the strength of the input signal.

Information relating to the setting of the device 25 is fed on a line 28 to the control unit 6. This information reflects the strength of any interfering signal present on the channel. Where there is a t strong interfering signal, less gain will *be required than if the *only received signal is the transmission from the other radio station. The receiver 24 also receives information from the control unit 6 on line 29 relating to the subset of channels which will be used by the other radio station for transmitting the message data.

The MFSK demodulator 26 may be of any conventional type. For example, it may include a series of parallel correlators, each of which receives the signal output from the receiver 24. One correlator is matched with each of the possible tones transmitted. The demodulated tone is taken as the tone corresponding to the correlator which produces the highest output. However, if' the channel spectrum information available from all of the correlators is utilised in the demodulator, information regarding the likely occupancy of the channel can be retrieved.

For example, if the ratio of the output from the correlator of the demodulated tone and the next highest correlator output is taken, this can be fed as information related to the occupancy of the channel on line 30 to the control unit 6. Where the channel is occupied, this ratio would be relatively close to one, whereas if' the channel is unoccupied, S.. S the ratio would be higher representing the signal to noise ratio. Demodulation can also be carried out by measuring the strength of the signal at a number of intervals within the frequency spectrum in order to determine where the peak power within the frequency spectrum is received. The number of sampling points is at least several (for example 8) times the number of tones. The difference in the frequency domain between the position of the peak power and the nearest valid tone frequency can then be output on line 32 to the control unit 6. This difference also relates to the probability that the channel is occupied. If the channel is occupied by a strong interferring signal, then the peak level may occur anywhere within the channel spectrum with a substantially flat probability distribution function.

However, if the channel is unoccupied and noise free, the peak will be located on a valid tone frequency.

However constructed, the MFSK demodulator 26 converts the received signal tones into the symbols which they represent and these are fed to a decoder 28. This decoder is a pair of RS decoders. The first decoder decodes the inner code and then the second decoder decodes the outer óode applied by the encoder 12.

When decoding the inner code, the Reed Solomon decoder produces information about the presence of errors and ensures and their location. This information is used to update a history file maintained in a random access memory associated with the decoder. The history file contains 1 byte of memory for each of the 256 channels. If after error correction has been carried out by the decoder of the outer code an error was found in a symbol received on a particular channel the entry in the history file for that channel is left shifted and a 1 inserted.

If the symbol received on a particular channel is not in error then the history file entry is left shifted and a 0 is inserted. In this way the history file contains a 1 byte wide binary number that represents the probability that each channel is occupied. The larger the probability the more likely it is that the channel is occupied. The contents of the history file may also be used by the decoder itself in the decoding process and also passed to the following outer decoder, which utilises such information to provide a more accurate output. Information about the channel on which each symbol is received is fed directly from the receiver 24 to the first decoder 28 on line 311. The contents of the history file are also passed to the control unit 6 on line 32. The decoder also receives an input on line 36 from the

A

control unit 6. This conveys information about the variable inner code which has been applied to the transmission from the other radio station. In this way the decoder 28 selects the appropriate decoding strategy to produce a correct output. The output from the decoder 28 is fed to a data unpacker 38.

The data unpacker 38 diverts control data such as the contents of the header block along line AID to the control unit 6 and passes the data relating to the message transmitted from the other radio station to output line 42.

Figure 2 shows a block diagram of the control unit 6 of the radio station in more detail. The control unit 6 is preferably implemented as one or more microprocessors with associated memory operating under the control of a stored programme. The control unit 6 has two primary functions. Firstly, to process the information input on lines 28, 30 and 32 from the automatic gain control device 25 of the receiver, the MFSK demodulator 26, and the decoder 28 respectively, in order to maintain a probability table 50 stored in random access memory (RAM). In the present example, this table is 256 long by 1 byte wide. Each byte represents one channel and stores information relating to the probability that that channel is occupied. A channel and code selector 52 associated with the probability table derives the contents of a header block from the contents of the table and outputs this information on line l't to the transmitter unit so that the header block can be transmitted to another radio station when a link is being set up between the two radio stations. The second function of the control unit 6 is to process the information in the header block received from the other radio station and input on line 40 in order to provide the necessary control information to the encoder, decodier, receiver and transmitter so that the appropriate channels and codes are used.

The contents of the probability table at a first radio station relate to the noise and interference conditions prevailing over the communications link when the first radio station containing the control unit is the receiver and another radio station is the transmitter. Therefore, a header block derived from the probability information is transmitted back to the other radio station which uses it to select the channels on which subsequent message data is transmitted to the first radio station. The first radio station control unit receives a header block from the other radio station which relates to the S. S noise -and interference conditions relevant to the first radio station as transmitter. Therefore, the contents. of this received header block are used to provide the output signals on lines 16 and 22 to determine te code applied by encoder 12 and the channels used by transmitter 20. It will therefore be appreciated that the code and channels used by the transmitter unit and receiver unit of one radio station may be different if the noise and interference conditions are perceived as different in the two alternative directions. In an alternative embodiment, the control unit at one of the radio stations may be the master unit and determine the code and subset of channels to be used for communications in both directions. In this case, the * master control unit will always transmit a header block to the other radio station. The transmission of a header block from the other radio station back to the master radio station is not essential, though it may be provided in order to provide further updating information for the probability table in the master control unit.

The inputs on lines 28, 30 and 32 are each fed to respective processors. The input on line 28 relating to the setting of the AGC device 25 is fed to an AGC 17 $ processor 54. In its simplest form the AGC processor 54 will be a lookup table which converts each possible setting of the AGC device into a probability which is fed to a probability combiner 56.

The input signal on line 30 from the MFSK demodulator 26 is fed to a soft decision processor 58. This processor may again be a lockup table which converts the ratio signal or the signal representing the difference between the peak power of the spectrum and the nearest valid tone frequency, depending on the type of demodulator used, into a probability which is fed to the probability combiner 56. The output on line 32 from the decoder 28 is fed to an error correlation processing 60. The error correlation processor 60 may convert the contents of the history file RAM, or may simply be the history file RAM from which the probability information for each channel can be fed to the probability combiner 56. When an amended probability is available for any channel, then it is combined with the previous contents for that channel stored in the probability table 50 in the probability combiner 56 and the result stored as a new value for that channel in the probability table 50. It will be appreciated that the input relevant to a channel is first available from the AGC device in the receiver and shortly thereafter from the modulator. The input for a channel from the decoder is only available after some delay introduced by the decoding process. If all the methods of calculating the probability that a channel is occupied were independent, the probabilities from each source can be combined in accordance with the standard formula for combining two probabilities p and q. This gives a total probability of: pq pq + (1-p)(1-q).

The code arid channel selector 52 produces the information for the header block to be transmitted to the other. station. This header block includes a.256 bit long sequence of flags. Each flag represents one chanel. The flag for a channel is set to 0 or I in dependence upon whether the stored probability is above or below a predetermined threshold. The threshold is set so that 128 channels have their flags set to 0 and are not to be used and the other 128 channels have their flags set to 1 and are to be used. This ensures that the -128 least occupied channels will be utilised. The code and channel selector also determines a coding strategy to be used for the tran3mission to its radio station in dependence upon the level at which the threshold must * -be set. It will be appreciated that if the threshold level of probability in order to achieve 128 free channels is relatively high, then the channels actually used will be subject to some interference and therefore a higher degree of error protection is necessary. However, if the threshold probability needs only to be set relatively low, a lesser degree of coding protection can be used and this will ensure that the maximum information rate is available. The code and channel selector therefore converts the threshold information into data bits which identify which of the available inner codes is to be used, These data bits are added to the header block. The number of bits needed to send this information depends on the number of coding strategies available to the system. In the present example 3 bits are necessary to determine which of the six possible inter codes are used. The output of the code and channel selector 52 is fed on line 1I to the data packer 10 for transmission of the header block. The sequence of flags is also fed on line 29 to the receiver so that the receiver is aware of' which channels will be used in the following message transmission. The data bits from the header block representing the inner code to be used are fed on line 36 to the decoder. The header block is transmitted after synchronisation and just -before each message. The header control block is sent on all 256 channels and encoded using a powerful error correcting code to ensure a high probability that it will be received correctly.

The received, decoded and corrected header control block information is fed to the control unit 6 on line 110 to a header processor 62 where the 256 flag sequence is separated from information relating to the code to be used. This sequence of flags may be used to modify the probability table 50 and is also output on line 22 to the transmitter 20 to determine.

which channels are to be used for transmission. The coding strategy information is fed on line 16 to the encoder 12.

The frequency hopping transmitter and receiver 20, 24 each contain a key generator which determines the next channel to be used in the following hop period.

The key generators at the transmitter and receiver of communicating stations are synchronised in a known manner. However, because only half of the available channels are to be used, it is necessary for a substitute channel to be selected when the key generator outputs a channel which is flagged as occupied. This is conveniently done by using a suitable hashing function which may be repeated until a suitable unoccupied channel is selected. This hashing function is brought into operation if the channel selected by the key generator is flagged as not to be used or is the same as the previous channel. A suitable hashing function is: Next channel (old channel + 89) mod 256 Since channels flagged as occupied are revisited during transmission of the header block, it is possible for them to be recognised as unoccupied once they cease to be occupied.

It will be appreciated that the described system allows increased information rates to be used for message transmission because only a subset of the available channels which are least occupied are used for the message transmission. The selection of the channels to be used is provided by maintaining a probability table which is updated in response to all received transmissions. S *

Claims

CLAIMS1. A frequency hopping MFSK radio system comprising a transmitter unit and a receiver unit; the transmitter unit comprising encoding means for encoding data to be transmitted using at least one forward error correcting code and outputting encoded symbols to be transmitted, an MFSIC modulator for converting each of the encoded symbols into a tone, and a frequency hopping transmitter for transmitting the tones and comprising a frequency synthesiser, and means for controlling the synthesiser to produce for each successive symbol period a frequency representing a channel frequency plus the tone representing the symbol to be transmitted during that symbol period, each symbol period being equal to a hop period for which the channel frequency Is constant; the receiver unit comprising a frequency hopping receiver for outputting received tones, an MFSK demodulator for outputting symbols from the received tones, and decoding means for decoding the symbols and outputting the transmitted data and information relating to the presence and channel location of errors detected in the transmission, the receiver unit further comprising control means -w -comprising means I or storing information for each of the channels available to the system, which information represents a probability that that channel is occupied, and means for updating said probability information at least from said information output from the decoder, and means for periodically transmitting back to the transmitter on all channels, data relating to said stored probabilities, the transmitter unit comprising means for selecting a subset of the available channels for use during subsequent transmission in response to the received data from the receiver unit.
2. A radio system according to claim 1, wherein the demodulator at the receiver unit outputs information relating to the entire spectrum received on the channel in addition to the demodulated symbol, said updating means updating said probability information from this spectrum information in addition to said information output from the decoder.
3. A radio system as claimed in claim 2, wherein said channel spectrum information relates to the position of the peak received signal in the frequency domain of the channel relative to the nearest valid tone frequency. 4 3
4. A radio system as claimed in any one of the preceding claims, wherein the receiver comprises an automatic gain control device, and the gain setting of this device is fed to said updating means.
5. A frequency hopping MFSK radio system in which data is transmitted only on a selected subset of the available channels, the radio system comprising means for selecting said subset in dependence on the contents of a probability table which stores information for each channel relating to the probability that that channel is occupied, the probability table being updated from information derived from at least one of the following: the output of a receiver decoder for decoding an error correcting code applied to the data to be transmitted relating to the presence and channel position of errors detected in received data; channel spectrum information derived from an MFSK demodulator at the receiver; and the setting of an automatic gain control device at the receiver. _t -
6. A frequency hopping MFSK radio system substantially as herein described with reference to the accompanying drawings.Amendments to the claims have been filed as followsCLAIMS1. A frequency hopping MFSK radio system having a predetermined plurality of available channels and in which data is transmitted between a plurality of transmitter/receiver units but only on a selected subset of the available channels, the system comprising selecting means for selecting the subset in dependence on the contents of a probability table which stores information for each channel relating to the probability that that channel is occupied, the selecting means including updating means for updating the probability table and comprising means for applying an error correcting code to the data transmitted by the transmitter of at least one of the units and a receiver decoder in at least one of the units for decoding the error correcting code as received by the receiver with that transmitted data and thereby determining the presence and channel position of errors in the received data, and means responsive to the channel position of those errors for updating the probability table.2. A system according to claim 1, in which the updating means is also responsive to the setting of an automatic gain control device at the receiver of at least one of the units which setting is dependent on Amendments to the claims have been filed as follows the strength of any interfering signal present on the channel on which data is being received.3. A system according to claim 1 or 2, in which the updating means includes means responsive to channel spectrum information derived by demodulating data received at the receiver of at least one of the units.4. A system according to claim 3, in which the said demodulation outputs information relating to the entire spectrum received on each channel in addition to the demodulated data, the said selecting means updating the probability information from this spectrum information.5. A system according to claim 4, in which the channel spectrum information relates to the position of the peak received signal in the frequency domain of the current channel relative to the nearest valid tone frequency.6. *A system according to any preceding claim, in which the transmitter in each unit comprises encoding means for encoding data to be transmitted using the said error correcting code in the form of a forward error correcting code and outputting encoded symbols to be transmitted, an MFSK modulator for converting each ofIAmendments to the claims have been filed as follows the encoded symbols into a tone, and frequency hopping means comprising a frequency synthesiser and means for controlling the synthesiser to produce for each successive symbol period a frequency representing a channel frequency plus the tone representing the symbol to be transmitted during that symbol period, each symbol period being equal to a hop period for which the channel frequency is constant.
7. A system according to claim 6, in which the receiver in each unit comprises a frequency hopping receiver for outputting received tones, an MFSK demodulator for outputting symbols from the received tones, and decoding means for decoding the symbols.
8. A system according to any preceding claim, in which the selecting means includes means at at least one of the units for periodically transmitting to the transmitter at at least one other said unit, and on all the channels, data relating to the stored probabilities, the latter transmitter including means for selecting the said subset of the channels.
9. A frequency hopping MFSK radio system, substantially as herein described with reference to the accompanying drawings.