GB2495709A - Assigning orthogonal and non-orthogonal spreading codes in a CDMA transmission system - Google Patents

Abstract

Transmitting signals in a CDMA system comprising: assigning to up to n users an orthogonal spreading sequence derived from a spreading code having n bits and n orthogonal spreading sequences; spreading the signal of each of the n users by applying the respective orthogonal spreading sequence; and transmitting from each user the spread signal; wherein users in excess of n are assigned a non-orthogonal spreading sequence.

Description

COMMUNICATION SYSTEM AND METHOD OF OPERATING THE SAME

The present invention relates to a communication system and to a method of operating the same, in particular to a system and method using random access communications. The communication system and method of the present invention find use in wireless communications, such as mobile telephone and satellite communications.

A random access communication system will typically have a plurality of tO users able to transmit using a common communication capacity. In particular, users communicating at a given time are sharing a common band of frequencies at which data is being transmitted. As a result, signals being transmitted from different users simultaneously can interfere with each other, generally referred to as collision. Such collisions can, in some cases, prevent the successful reception of the transmitted signal. A typical random access communication system employing a satellite is shown in Figure 1. The system comprises a hub station communicating with a plurality of user terminals via a common satellite transponder.

Two general approaches have been applied to random access methods of communication. The ALOHA protocol employs a hub/star configuration, characterised by a communications hub transmitting data to all users on a first frequency and with users transmitting data to the hub on a second frequency. The principle of operation is that a given user transmits data to the hub on the second frequency. Upon receipt of the data, the hub retransmits the received data to all users on the first frequency. The retransmission of the data by the hub allows the user to determine if the transmitted data were received by the hub. Should no acknowledgement be received from the hub by the user or the data received by the user not correspond to the data originally transmitted, the user simply retransmits the earlier signal. Collisions arise when two or more users attempt to transmit data to the hub at the same time. Various improvements to the basic ALOHA system were based on the manner in which the hub retransmits the data to the users. The overall efficiency of the system is determined, in large part, by the manner in which users wait between transmissions. The most significant improvement to the basic ALOHA principle was the introduction of time slots, with users only be able to commence transmission of data at the start of a given time slot. In this way, the collision between users would occur for a complete time slot and collisions involving the partial overlap of transmitted data are removed. This slotted ALOHA protocol doubled the efficiency of the system.

An alternative approach is provided by code division multiple access (CDMA) systems. In COMA systems, multiple users are allowed to transmit simultaneously, in turn allowing interference between user signals. However, the signal transmitted by each user is spread in bandwidth, thereby reducing the effect of the interference.

The signal bandwidth may be spread in a number of ways. The most commonly applied method for spreading signal bandwidth is the use of a binary spreading code or sequence to modulate the signal of the user. In use, the data signal from each user is multiplied by the spreading code, thereby spreading its bandwidth, before the signal is transmitted. To reduce collisions between multiple user signals, each user is supplied with a different spreading code. It is known to use Gold codes and Kasami codes as spreading sequences in COMA systems.

Methods to reduce or avoid collisions between multiple users of a COMA system are proposed in the art.

US 5,537,397 describes a spread ALOHA COMA data communication system and method, in which all users are assigned the same spreading code selected from the class of maximum length shift register sequences for transmissions. Repeated subtractions from the received signal are used to leave a single signal message.

US 5,544,196 discloses an apparatus and method for reducing message collision between mobile stations accessing a base station in a CDMA system at the same time. The method uses one or more randomisation methods to distribute the transmissions from each mobile station. In a first randomisation method, the signal transmitted by a mobile station is spread by a pseudonoise (PN) code. The mobile station time-delays its transmissions by a number of chips of the PN code with which it spreads the transmitted signal. A further randomisation method involves the mobile station randomly selecting a RN code. In a third method, the mobile station inserts a random delay between successive message transmissions or probes, in cases where it does not receive an acknowledgement from the base station after a predetermined timeout period.

US 5,716,236 discloses a system and method for generating signal waveforms in a CDMA system. In the method, FN sequences are constructed that provide orthogonality between the users of the system. In this way, mutual interference between users is reduced, in turn allowing for a higher system capacity and better link performance.

US 6,075,795 addresses the problem of interference between multiple users of a CDMA system by providing for collisions to be detected. The method comprises generating a collision code corresponding to ID information of a station. A collision symbol representing the collision code is transmitted at the beginning of a portion of a random access data communication. The collision symbol is received by a station in the network, which in turn decodes the collision signal carried by the collision symbol. A collision is detected if the collision signal does not satisfy certain predetermined requirements.

US 6,181,683 describes a method for the transmission of data in a CDMA system, in which one of the CDMA channels is used in a time-shared manner for the transmission of data packets from more than one transmitting station to a receiving station.

US 6,580,747 suggests a method and apparatus for generating orthogonal spreading codes in a CDMA radio system. The method includes obtaining control data for the spreading code, including the length, running number and code class of the spreading code. The spreading code is formed such that the running number of the spreading code in the code class is first denoted by a binary code number of the same length as the code class, after which the modified code number is formed by reversing the code number.

A radio communication system including a collision detection function is described in US 2009/02 1 9874.

More recently, US 2010/0238846 describes a wireless communication system based on code spreading-orthogonal frequency division multiple access (CS-OFDMA) and a smart antenna. The system is indicated to avoid inter-symbol interference caused by the conventional transmission of wideband data on a COMA system and to solve the problem of frequency selective fading and inter-cell interferences in known OFDMA techniques.

It has now been found that the collisions between simultaneous users of a CDMA system may be significantly reduced or eliminated if some of the spreading codes used to spread the signals of users are orthogonal to each other. This in turn allows the CDMA system to operate with a better than 100% spectral efficiency. The spreading codes are selected such that the cross-correlation between the spreading sequences is low and the number of spreading sequences matches the system capacity requirements. In particular, it has been found that, to provide maximum system efficiency, the spreading sequences should be from a code that comprises orthogonal sequences as a sub-code. It has further been found that a significant improvement in spectral efficiency of the COMA system may be achieved by assigning such spreading sequences to users in a particular manner.

Accordingly, the present invention provides a method for transmitting signals in a CDMA system using as spreading sequence a code having n bits and n orthogonal sub-codes. In particular, according to a first aspect of the present invention, there is provided a method for transmitting signals in a CDMA system, the method comprising: assigning to up to n users an orthogonal spreading sequence derived from a spreading code having n bits and n orthogonal spreading sequences; spreading the signal of each of the n users by applying the respective orthogonal spreading sequence; and transmitting from each user the spread signal; wherein users in excess of n are assigned a non-orthogonal spreading sequence.

In a further aspect, the present invention provides a method for assigning spreading codes to multiple users in a COMA system, comprising:

S

providing a spreading code having n bits and n orthogonal spreading sequences; assigning each of up to n users a different spreading sequence selected from the n orthogonal spreading sequences; and thereafter assigning each user in excess of n users a non-orthogonal spreading sequence.

In a further aspect, the present invention also provides a CDMA system, the system comprising: spreading code assigning means operable to assign to up to n users an orthogonal spreading sequence derived from a spreading code having n bits and n orthogonal spreading sequences; spreading means operable to spread the signal of each of the n users by applying the respective orthogonal spreading sequence; and a transmitter for transmitting from each user the spread signal; wherein the spreading code assigning means is further operable to assign to each user in excess of n a non-orthogonal spreading sequence.

In a still further aspect, the present invention provides a system for assigning spreading codes to multiple users in a CDMA system, comprising: spreading code assigning means operable to provide a spreading code having n bits and n orthogonal spreading sequences and assigning each of up to n users with one of the n orthogonal spreading sequences; and thereafter operable to assign each user in excess of n users with a non-orthogonal spreading sequence.

The methods and systems of the present invention find use in a wide range of mobile communications, including but not limited to satellite communications and mobile telephone communications.

It has been found that a spectral efficiency of greater than 100% can be achieved in a CDMA system by applying spreading codes designed such that the cross-correlation between sequences is small. In particular, it has been found that the cross-correlation between sequences is avoided when the spreading codes assigned to users are sub-codes of a spreading sequence code that comprises orthogonal sub-codes. A spreading sequence of n bits in length and containing orthogonal sub-codes will comprise n orthogonal sub-codes. In the method and system of the present invention, the first n users of the CDMA system are each assigned a spreading sequence selected from the n orthogonal sub-codes of the spreading code. As the spreading sequences for all of these users are derived from orthogonal sub-codes, there is no cross-correlation between the codes and hence no interference between the signals of the first n users. Thereafter, the n + I and subsequent users of the system are assigned further spreading sequences. As these further sequences are not orthogonal, interference between the users will arise and will affect the transmissions of all users. However, the interference increases, and the efficiency of the system decreases, only as a function of the number of further users assigned the non-orthogonal spreading codes and not as a function of the total number of users of the system. As a result, the capacity of the CDMA system can be significantly increased and/or the spectral efficiency of known systems can be increased. In particular, it has been found that a spectral efficiency of greater than 100% can be achieved using the method and system of the present invention.

This compares most favourably with a typical spectral efficiency of 20% for known ODMA systems using Gold codes or Kasami codes as spreading sequences and a maximum spectral efficiency of 37% achievable with slotted ALOHA systems.

In the methods and systems of the present invention, a spreading code is assigned to each user. In one embodiment, each user is assigned a fixed spreading code that is applied to all signals transmitted by that user at any time. Alternatively, a spreading code may be assigned to a user dynamically, that is on demand, as and when that user wishes to transmit a signal. In this embodiment, a single spreading code may be assigned to different users at different times, according to user demand. In a preferred embodiment, the spreading codes, in particular the n orthogonal spreading codes, are assigned in a dynamic manner to users on demand.

In this way, no interference occurs between users until all n orthogonal spreading codes having been assigned and the number of users exceeds n.

Each user may be assigned a single spreading sequence. As an alternative, a user may be assigned multiple spreading sequences, for example when the system is underloaded and additional information is required to be sent. In particular, a user may be assigned more than one of the n orthogonal spreading sequences. As noted above, for a given spreading code having n bits, there is a maximum of n orthogonal sub-codes. In the case that a user is assigned more than one of the n orthogonal spreading sequences, the capacity of the system is reduced. In particular, the number of users having orthogonal spreading sequences assigned to them and for which interference is avoided is reduced to less than n users. References to the further users as being n + 1 and higher herein are therefore to be reduced accordingly. I0

As described above, the method and system of the present invention rely on assigning spreading sequences to users of the CDMA system derived from a spreading code having orthogonal sub-codes or sequences. In particular, the spreading code having n bits comprises n orthogonal sub-codes applied as spreading sequences in the present invention. Thus, using a spreading code having n orthogonal sub-codes, the first n users of the CDMA system are each assigned a respective spreading sequence selected from the n orthogonal spreading sequences of the spreading code. As the spreading sequences are orthogonal, interference between the signals transmitted by the up to n users is avoided. Thereafter, should there be further users of the CDMA system, each user in excess of n is assigned a non-orthogonal spreading sequence. As the spreading sequences assigned in excess of n will not be orthogonal sequences, interference between the users assigned these further sequences can arise. However, as no interference between the first n users will arise, any reduction in system performance is limited to that arising from the n + 1 and subsequent users.

As noted above, for an appropriate spreading code having n bits there are n orthogonal sub-codes used in the present invention as spreading codes for up to the first n users of the CDMA system. Thereafter, the further users are assigned non-orthogonal spreading codes. The spectral efficiency of the system is determined by the number of further users assigned the non-orthogonal spreading codes. In particular, it has been found that such further users suffer interference from all users of the system. Accordingly, in a preferred embodiment, the system and method of the present invention are operated such that the users assigned a spreading sequence selected from the n orthogonal sub-codes of the spreading code are in the majority. That is, it is preferred that the number of users assigned non-orthogonal spreading sequences is less than the number of users assigned orthogonal spreading sequences. Preferably, the number of users assigned non-orthogonal spreading codes is less than 90% of the number of users assigned orthogonal spreading codes, more preferably lesss than 75%, still more preferably less than 50%.

It is a feature of the method and system of the present invention that the first tO n users are assigned a spreading sequence selected from the n orthogonal sub-codes of the spreading code. Thereafter, further users are each assigned a different, non-orthogonal sub-code. As noted, interference between transmitted signals arises between those users assigned non-orthogonal spreading codes. As a result, the level of interference for such users will increase with an increase in the number of the further users in excess of n. The spectral efficiency may be increased, for example, by allowing those users assigned non-orthogonal spreading codes to transmit more power allowing them to operate at a higher energy per bit to noise power spectral density (EJN0) ratio than those users assigned an orthogonal spreading sequence.

As noted, the interference experienced by users will depend upon the total number of users in excess of the n users assigned orthogonal spreading codes. For example, if 25% of users are operating with non-orthogonal spreading codes then these users will experience on average 4 times the interference compared to users with orthogonal spreading codes. Depending on the received noise level a balance may be struck so that each user experiences the same total average interference plus noise. An adaptive uplink power control system may be used to advantage to achieve this. Typically, users operating with non-orthogonal spreading codes may transmit at higher power levels, for example at 2dB to 10dB, more preferably at 3dB to 6dB, higher power levels compared to users with orthogonal spreading codes.

The present invention may employ any spreading code provided it comprises orthogonal sub-codes for use as spreading sequences.

A linear error correcting code is a linear subspace of the space F2 is the space of all vectors of length n over the binary field. For binary linear codes, every code is commonly described as an [n, k, d] code, wherein n is the length of the code in bits, k is the dimension of the code, meaning that there are a total of 2k codewords, and d is the minimum Hamming distance of the code.

Non-linear error correcting codes aie denoted as (n, M, codes, where Mis the number of codewords in the non-linear code and n and d having the meaning indicated above for linear codes. I0

If codewords of binary error correcting codes are used as spreading sequences, there is a relationship between the Hamming distance between the codewords and their cross correlation value. Defining each codeword as c1(x) for / = 0 to 2k -I and each spreading sequence as s1(x), then the following applies: = and (-i)C.JtJ In this way, the spreading sequences have values of +1-1 mapped from the 1 and 0 coefficients of the codewords.

The cross correlation, X0 between two sequences s(x) and sp(x) is given by the following: k fl Substituting from above gives the following equation: = -1) j = (-1)< C =0 where the addition of c and c is modulo 2. If the Hamming distance between codewords c0(x) and cp(x) is 5, then the following applies:

-

This is the relationship between the cross correlation values that will occur and the Hamming distance of the codewords, when using codewords from an error correcting code as binary spreading sequences.

Preferred codes for use as the spreading code are sequences derived from binary error correcting codes. The codes may be linear or non-linear. In one embodiment, the code is non-linear, in particular a quadratic residue code. In an alternative embodiment, the code is linear, in particular a Reed-Muller code.

Enlarged Reed-Muller codes in which the dimension of the code is increased by adding vectors to the generator matrix are preferred Reed-Muller codes. Such vectors having weight equal to the covering radius of the Reed-Muller code are particularly preferred.

Methods for designing binary error correcting codes, from which spreading codes may be derived, are known in the art, for example as described by MacWilliams, F.J., et al., The theory of Error Correcting Codes', North Holland, 1977, and Boemer, A.B., Binary pulse compression codes', IEEE Trans. Inform. Theory, Vol. lT-13, April 1967, pages 156 to 167.

The spreading sequences may be of any suitable length, preferably at least 64 bits. Preferred sequences have a length between 64 bits and 256 bits, as longer sequences permit more simultaneous users. However, longer sequences also require more bandwidth expansion. I0

As noted, non-linear quadratic residue codes are one preferred class of codes for deriving the spreading sequences for use in the present invention. Every non-linear quadratic residue code may be derived from an odd prime number, p. Prime numbers are classified in terms of the following four possibilities: p=Sm-1; p=Sm÷ 1; p = Sm -3; or p = 8m +3.

For each value of p, a primitive root ci may be found and each codeword c1(x) defined in terms of the even powers of a and an exponent offset i, where / = 0 to p - 2. The weight of each of the p -I non-zero codewords depends upon the number of quadratic residues equal to (p -1)/2. For the cases that p is equal to Sm + 1 or Sm - 3, the number of quadratic residues is even and is equal to 4m or 4(m -I) + 2 For the cases that p is equal to Sm -I or Sm + 3, the number of quadratic residues is odd and is equal to 4m -I or 4m + 1. Each codeword may be expressed as a polynomial and is given by the following formula:

I (I

i=Otop-2.

Where the number of quadratic residues is odd, it has been shown that the Hamming weight between any two distinct codewords is always (p + 1)/2and where the number of quadratic residues is even, the Hamming weight between any two distinct codewords is either (p -1)12 or (p ÷ 3)/2.

For application to spreading sequences for CDMA systems, the odd number quadratic residues are required and p is equal to either Sm -I or 8m + 3. From the relationships set out above, the cross correlation between sequences is as follows: p-2p+l)/2)=-l Orthogonal codewords with zero cross correlation may be produced by extending the code length by one bit to produce spreading sequences of length p + 1 bits by adding an x" term. The length of the orthogonal code is p + I bits and there are 2(p + I) orthogonal sequences, including the inverses of sequences, as follows: -i i=Otop-2 Each orthogonal code may be enlarged by adding additional vectors as additional codewords to the code. These vectors preferably have low correlation with the existing codewords for use as spreading sequences. It is possible to calculate the best performance attainable for the code by calculating the covering radius of the code, as described by Helleseth, T. et al., On the covering radius of binary codes', IEEE Transactions Information Theory, IT-24, September 1978, pages 627 to 628.

Additional vectors for non linear quadratic residue codes may be obtained by adding codewords of the code together. By definition, these vectors have weight (p + 1)12 from above, as the code is orthogonal. As a result of the considerable structure in these codes, the cross correlation between these vectors is strongly constrained, in turn producing good performance as spreading sequences.

Preferred choices for code lengths, the corresponding minimum Hamming distance and the worst case cross correlation values resulting from the additional codewords are given in the following table: Code length Number of 6 Cross Normalised (bits) sequences correlation value 72 284 28 16 0.222 316 32 16 0.2 84 332 36 -20 -0.238 104 412 44 -24 -0.231 108 428 44 20 0.185 128 508 56 -24 -0i875 132 524 56 20 0.151 556 60 20 0.142 152 604 64 24 0.158 168 668 72 24 0.143 A further general class of codes that may be used to generate spreading sequences in the present invention are a sub-set of the BCH codes. As noted above, another preferred class of codes for deriving spreading sequences are the Reed-Muller codes. The first order Reed-Muller codes may denoted by R(1,m) and are binary linear codes of length 2, dimension () --1 and have a minimum Hamming distance 2m1. R(1,m) is equivalent to the [2m, m ÷ 1, 2m1] extended BCH code.

The Reed-Muller codes are orthogonal and in their non-extended form are cyclic codes with parameters [2 -1, m + 1, 2 -1]. In addition, to the Reed-Muller codes, there exist the simplex codes with parameters [2 -1, m, 2] and denoted by J. If R*(1, m) is the non-extended code of R(1, m), which is cyclic, it is the case that: 9(: : *:?* 1. ?fl..

In other words, all of the codewords of the simplex code are contained in the non-extended first order Reed-Muller code of the same length. Indeed, the remaining codewords of the first order Reed-Muller code are basically the inverted codewords of the simplex code, that is: ri 7; 1J(i -1-where 1 is an all l's vector of length 2 -1.

Simplex codes have the unique property that a shifted codeword c(x), added to itself is equal to a shift of itself, the same codeword, that is: I5 U r a P for all j and corresponding Ic.

This arises as, being a cyclic code, every shifted codeword is also a codeword and, having dimension m and length 2 -1, there are 2 -2 distinct shifts of one codeword, which accounts for all codewords, apart from the all zero codeword. A codeword of Jm c1(x) is given by the following polynomial division: r r1 pix fori=Oton-2.

where p(x) is any primitive polynomial of degree m. For example, for n = 127, p(x) = 1 + ÷ x7 is one such primitive polynomial. The weight of c(x) for any i is 2m1 and from the above analysis, the cross correlation between any two codewords is always -1.

Extending each codeword with one additional bit equal to 0 produces a cross correlation of 0 between any two codewords. All codewords may be inverted with changing the minimum Hamming distance between codewords. Accordingly, the total dimension of the extended code is m 4-1 and there are a total of codewords tO in the code. From the above equation, the resulting code after extension and adding the inverted codewords is the first order Reed-Muller code R(1,m).

To increase the number of spreading sequences the Reed-Muller code no,m) may be enlarged by adding additional vectors as codewords, in particular codewords from the extended BCH code, as described above. Helleseth, T. et al., noted above, have shown that the covering radius of R(1,m), denoted by Tm Is given by the following:

-

for even m; and )1 I -- -fl. -. foroddm.

Thus, for the [32, 6, 16] Reed-Muller R(1,5) code, it is known that r5 = 12. For the [128, 8, 64] Reed-Muller R(1,7) code it is known that r7 = 56. In each case, rm equals the lower bound. This means that the minimum Hamming distance of the enlarged code is either known or bounded. For the [128, 8, 64] Reed-Muller R(1 7) code, it is known that an enlarged code having parameters [128, 9, 56] is attainable.

It is possible to have a higher dimension with the same ö, by noting that the [128, 8, 64] Reed-Muller R(1 7) code is equivalent to the [128, 8, 64] extended BCH code, in turn contained as a sub-code within the [128, 15, 56] BCH code. Thus, the [128, 8, 64] Reed-Muller R(1 7) code may be enlarged to [128, 15, 561 by using the BCH code construction. With this code, there are a total of 32768 spreading sequences available and from the above relationships it follows that the worst case cross correlation is 16, with a corresponding normalised value of 0.125.

For those cases where the covering radius or the Reed-Muller R(1 rn) code is equal to the minimum Hamming distance of the [2m, 2m + 1, O] BCH code, there is no loss in performance in using the BCH code construction to produce the enlarged code. The covering radius for R(1, m) as a function of m, the minimum Hamming distance of the corresponding BCH codes and the resulting worst case cross tO correlation values are given in the following table.

Code length Covering o of Number Cross Normalised (bits) radius enlarged sequences correlation value extended value BCH code 32 12 12 2048 8 0.25 64 28 28 1024 8 0.125 128 56 56 32768 16 0.125 256 120 120 8192 16 0.0625 512 240 to 244 240 524288 32 0.0625 At low spreading bit rates, the various users of the CDMA system may be effectively synchronised. However, synchronisation of the users may decrease as higher spreading bit rates are employed. To counteract this effect, the carrier frequencies of users identified as having the greatest timing uncertainty may be offset from each other in multiples of the information bit rate. In this way, orthogonal carrier frequencies (after despreading in the receiver) may be used, similar to that achieved using orthogonal frequency division multiplexing (OFDM). The increase in occupied bandwidth resulting from this technique will depend upon the magnitude of the synchronisation errors, but is typically less than 10%.

In addition, or as an alternative, the spreading codes may be combined with OFDM, so that the spreading sequence bit period is lengthened, thereby reducing the effects of synchronisation errors. Suitable OFDM techniques are known in the art.

Still further, the spreading codes may be multiplied by a covering sequence to minimise increases in the cross correlation due to synchronisation errors.

Embodiments of the present invention will now be described, by way of example only, having reference to the accompanying figures: Figure 1 is a diagram of a typical random access communication system employing a satellite; Figure 2 is a graphical representation of the operation of a system of the general type of Figure 1 employing an assignment of spreading sequences according to one embodiment of the present invention; Figure 3 is a graphical representation of the operation of a system of the general type of Figure 1 employing an assignment of spreading sequences according to a further embodiment of the present invention; and Figure 4 is a graphical representation of the operation of a system of the general type of Figure 1 employing an assignment of spreading sequences according to a still further embodiment of the present invention.

Turning to Figure 1, there is shown a diagrammatical representation of a random access satellite communication system of known configuration. The system, generally indicated as 2, comprises a terrestrial hub station 4 having a plurality of CDMA modems 6 connecting to units of different functionality, as indicated. The hub station 4 includes an ISP modem 8, providing a link to a network, such as the internet 10.

The system 2 further comprises a satellite 12. A plurality of user satellite terminals 14, each comprising an indoor unit 16 and an outdoor unit 18, communicating with the hub station 4 in known manner.

The system of Figure 1 may employ the method and systems of the present invention in communications between the hub station 4 and the user terminals 14. It is to be understood that the system represented in Figure 1 is just an example of a communications system and variations to the components and arrangement of the system will be readily apparent to the person skilled in the art. I0

In operation of the system 2 of Figure 1, the hub station 4 assigns a spreading sequence to each of the user satellite terminals 14. The spreading sequence for each user may be assigned upon demand, as and when the user wishes to transmit a signal and communicate via the system. When a user wishes to transmit a signal, the hub station assigns the respective user terminal with a spreading sequence. When the user ends transmission, the spreading sequence is available to be assigned to another user, upon demand. In this way, a given spreading sequence may be assigned to a plurality of different users within the system at different times. Alternatively, each user may be assigned a fixed spreading sequence. In this way, a given spreading sequence is only employed in the system when the user to which the sequence has been assigned is actively transmitting.

As described hereinbefore, the spreading sequences assigned by the hub station 4 are sub-codes of a spreading code, the spreading code having sub-codes that are orthogonal. In particular, for a spreading code of ii bits having orthogonal sub-codes, there are n orthogonal spreading sequences that may be assigned to up to n users of the system. Thereafter, once the n orthogonal spreading sequences have been assigned to users by the hub station 4, further users are assigned non-orthogonal sub-codes as spreading sequences. Interference in the system arises as the number of users assigned non-orthogonal spreading sequences increases.

However, the level of interference is independent of the total number of users.

To exemplify the present invention, a system of the type generally depicted in Figure 1 was analysed, in which each user was assigned a fixed, single distinct spreading sequence derived from a spreading code. As the net interference power level changes from data bit to data bit, due to the combined effects of the transmissions from other users, error correction coding was employed. For maximum performance, soft decision decoding was employed. In this example, the double circulant quadratic residue error correcting code based on the prime number 67, the [138, 68, 24] binary code, was used in conjunction with near maximum likelihood soft decision decoding using ordered reliability decoding, applying the techniques of Tomlinson, M. et al., Extending the Dorsch decoder towards achieving maximum-likelihood decoding for linear codes', lET Communications, Vol. 1, Issue 3, June 2007, pages 479 to 468.

For the analysis, spreading sequences were derived from the following code: (108, 150, 44) quadratic residue based code.

In each case, the first n users (where n is the number of bits in the code) were each assigned one of the n orthogonal sub-codes as a spreading sequence. Once the n orthogonal sub-codes had been assigned to users, additional users were each assigned a non-orthogonal sub-code as the spreading sequence. The efficiency of the system was determined at an energy per bit to noise power spectral density (FIN0) ratio of 5dB.

The maximum efficiency of the codes was determined in terms of the maximum number of users achievable without exceeding a packet error rate of 1 4* The percentage efficiency of the system operating with each code was obtained by normalising, by dividing the maximum number of users by the number of users of the system able to operate without interference. In the case of the above codes, this number is the length n of the spreading sequence.

For comparison purposes, the performance of the same system operating with spreading sequences derived from the following circulant codes was also determined: [88, 8, 40] circulant code; [112,8, 52] circulant code; and [56, 7, 26] circulant code.

The results are set out in Table I below.

Table I I0

Code Code Type Maximum number Efficiency (%) of users [88, 8, 40] Circulant 53 60.3 [112,8,52] Circulant 82 73.2 [56,7,26] Circulant 64 114.3 (108, 150, 44) Quadratic residue 131 121.3 The results set out in Table I are represented graphically in Figure 2.

As can be seen, by assigning to users a spreading sequence derived from the quadratic residue code, a spectral efficiency significantly in excess of 100 % can be achieved. The efficiency achieved is also significantly higher than the comparable circulant codes. As shown in Table I, the [56, 7, 26] circulant code can be used as a basis for spreading sequences and can achieve an efficiency of greater than 100%.

However, this code is limited to a maximum of 64 users. To increase the number of users above 64, a longer circulant code is required to be used. However, such longer circulant codes have a significantly poorer efficiency.

By way of further comparison, a corresponding analysis carried out using known Gold and Kasami codes resulted in an efficiency of less than 25 %.

As a further example, the efficiency of the system based on the following spreading codes was analysed: (108, 150, 44) quadratic residue based code; (132, 170, 56) quadratic residue based code; (128, 220, 56) enlarged Reed Muller code.

The method as outlined above was followed.

For comparison purposes, the performance of the same system operating with spreading sequences derived from the following circulant codes was also determined: [112,8, 52] circulant code; and [56, 7, 26] circulant code.

The results are set out in Table II below.

Table II

Code Code Type Maximum number Efficiency (%) of users [112,8,52] Circulant 82 73.2 [56,7,26] Circulant 64 114.3 (108, 150, 44) Quadratic residue 131 121.3 (132, 170, 56) Quadratic residue 154 116.7 (128, 220, 56) Reed-Muller 144 112.5 The results set out in Table II are represented graphically in Figure 3.

As can be seen, by assigning to users a spreading sequence derived from the quadratic residue codes and the enlarged Reed-Muller code, a spectral efficiency significantly in excess of 100 % can be achieved. The efficiency achieved is also significantly higher than the comparable circulant codes. As noted above, the [56, 7, 26] circulant code can provide an efficiency of greater than 100 % However, the maximum number of users is limited to 64. II is to be noted that the quadratic residue and Reed-Muller codes provide both an efficiency significantly greater than 100 % and a significantly higher number of maximum users than the 156. 7, 26J circulant code.

In the preceding examples, the system was analysed with each user being assigned a fixed spreading sequence. Greater efficiency can be achieved if the spreading sequences are assigned to each user on an on demand basis, that is as the user requests the hub station to transmit. To demonstrate this improvement in efficiency, the general system of Figure 1 was again analysed using the following codes: (108, 150, 44) quadratic residue based code; (132, 170, 56) quadratic residue based code; (128, 220, 56) enlarged Reed Muller code.

The method as outlined above was followed, with the exception that each spreading sequence was assigned to a user on demand.

The results are set out in Table Ill below.

Table Ill

Code Code Type Maximum number Efficiency (%) of users (108, 150, 44) Quadratic residue 137 126.9 (132, 170, 56) Quadratic residue 168 127.3 (128, 220, 56) Reed-Muller 160 125.0 The results set out in Table Ill are represented graphically in Figure 4.

As can be seen, by assigning to users a spreading sequence derived from the quadratic residue codes and the enlarged Reed-Muller code, a spectral efficiency significantly in excess of 100 % can be achieved. Further, by assigning the spreading sequences in a dynamic manner, that is an on demand basis, the spectral efficiency is further increased, compared with the same system with spreading sequences assigned on a fixed basis to each user.

tO The analysis described above assumes that all users of the CDMA system are synchronised within one bit period at the spreading code bit rate, as the transmitted signals are received by the satellite.

Claims (1)

1. <claim-text>CLAIMS1. A method for transmitting signals in a CDMA system, the method comprising: assigning to up to n users an orthogonal spreading sequence derived from a spreading code having n bits and n orthogonal spreading sequences; spreading the signal of each of the n users by applying the respective orthogonal spreading sequence; and tO transmitting from each user the spread signal; wherein users in excess of n are assigned a non-orthogonal spreading sequence.</claim-text> <claim-text>2. The method according to claim 1, wherein each user is assigned a fixed spreading sequence.</claim-text> <claim-text>3. The method according to claim 1, wherein a user is assigned a spreading sequence dynamically on demand.</claim-text> <claim-text>4. The method according to any preceding claim, wherein each user is assigned a single spreading sequence.</claim-text> <claim-text>5. The method according to any of claims 1 to 4, wherein each user is assigned a plurality of spreading sequences.</claim-text> <claim-text>6. The method according to any preceding claim, wherein the number of users assigned a spreading sequence from the n orthogonal spreading sequences is greater than the number of users assigned a non-orthogonal spreading sequence.</claim-text> <claim-text>7. The method according to claim 6, wherein the number of users assigned a non-orthogonal spreading sequence is less than 75% of the number of users assigned an orthogonal spreading sequence.</claim-text> <claim-text>8. The method according to any preceding claim, wherein users assigned a non-orthogonal spreading sequence transmit at a higher power level than the users assigned an orthogonal spreading sequence.</claim-text> <claim-text>9. The method according to claim 8, wherein the users assigned a non-orthogonal spreading sequence transmit at a power level from 3 to 6dB higher than the users assigned an orthogonal spreading sequence.</claim-text> <claim-text>10. The method according to any preceding claim, wherein the spreading tO sequences are derived from binary error correcting codes.</claim-text> <claim-text>11. The method according to claim 10, wherein the spreading sequence is derived from a linear code.</claim-text> <claim-text>12. The method according to claim 11, wherein the linear code is a Reed-Muller code or an extended Reed-Muller code.</claim-text> <claim-text>13. The method according to claim 10, wherein the spreading sequence is derived from a non-linear code.</claim-text> <claim-text>14. The method according to claim 13, wherein the non-linear code is a quadratic residue code.</claim-text> <claim-text>15. The method according to claim 14, wherein the spreading sequence is derived from an extended orthogonal codeword of the quadratic residue code.</claim-text> <claim-text>16. The method according to any preceding claim, wherein the spreading sequence is at least 64 bits in length.</claim-text> <claim-text>17. The method according to claim 16, wherein the length of the spreading sequence is from 64 to 256 bits.</claim-text> <claim-text>18. The method according to any preceding claim, wherein the carrier frequencies of users having a high timing uncertainty are offset from each other in multiples of the information bit rate.</claim-text> <claim-text>19. The method according to any preceding claim, wherein the bit length of the spreading sequence is increased using orthogonal frequency division multiplexing (OFDM).</claim-text> <claim-text>20. The method according to any preceding claim, wherein the spreading tO sequences are multiplied by a covering sequence.</claim-text> <claim-text>21. A method for assigning spreading codes to multiple users in a CDMA system, comprising: providing a spreading code having n bits and n orthogonal spreading sequences; assigning each of up to n users a different spreading sequence selected from the n orthogonal spreading sequences; and thereafter assigning each user in excess of n users a non-orthogonal spreading sequence.</claim-text> <claim-text>22. The method according to claim 21, wherein each user is assigned a fixed spreading sequence.</claim-text> <claim-text>23. The method according to claim 21, wherein a user is assigned a spreading sequence dynamically on demand.</claim-text> <claim-text>24. The method according to any of claims 21 to 23, wherein each user is assigned a single spreading sequence.</claim-text> <claim-text>25. The method according to any of claims 21 to 14, wherein each user is assigned a plurality of spreading sequences.</claim-text> <claim-text>26. The method according to any of claims 21 to 25, wherein the number of users assigned a spreading sequence from the n orthogonal spreading sequences is greater than the number of users assigned a non-orthogonal spreading sequence.</claim-text> <claim-text>27. The method according to claim 26, wherein the number of users assigned a non-orthogonal spreading sequence is less than 75% of the number of users assigned an orthogonal spreading sequence.</claim-text> <claim-text>28. The method according to any of claims 21 to 27, wherein the spreading tO sequences are derived from binary error correcting codes.</claim-text> <claim-text>29. The method according to claim 26, wherein the spreading sequence is derived from a linear code.</claim-text> <claim-text>30. The method according to claim 29, wherein the linear code is a Reed-Muller code or an extended Reed-Muller code.</claim-text> <claim-text>31. The method according to claim 28, wherein the spreading sequence is derived from a non-linear code.</claim-text> <claim-text>32. The method according to claim 31, wherein the non-linear code is a quadratic residue code.</claim-text> <claim-text>33. The method according to claim 32, wherein the spreading sequence is derived from an extended orthogonal codeword of the quadratic residue code.</claim-text> <claim-text>34. The method according to any of claims 21 to 33, wherein the spreading sequence is at least 64 bits in length.</claim-text> <claim-text>35. The method according to claim 34, wherein the length of the spreading sequence is from 64 to 256 bits.</claim-text> <claim-text>36. The method according to any of claims 21 to 35, wherein the bit length of the spreading sequence is increased using orthogonal frequency division multiplexing (OFDM).</claim-text> <claim-text>37. The method according to any of claims 21 to 36, wherein the spreading sequences are multiplied by a covering sequence.</claim-text> <claim-text>38. A CDMA system, the system comprising: spreading code assigning means operable to assign to up to n users an orthogonal spreading sequence derived from a spreading code having n bits and n orthogonal spreading sequences; spreading means operable to spread the signal of each of the n users by applying the respective orthogonal spreading sequence; and a transmitter for transmitting from each user the spread signal; wherein the spreading code assigning means is further operable to assign to each user in excess of n a non-orthogonal spreading sequence.</claim-text> <claim-text>39. The system according to claim 38, wherein the spreading code assigning means is operable to assign a fixed spreading sequence to each user.</claim-text> <claim-text>40. The system according to claim 38, wherein the spreading code assigning means is operable to assign a user a spreading sequence dynamically on demand.</claim-text> <claim-text>41. The system according to any of claims 38 to 40, wherein the transmitter is operable such that users assigned a non-orthogonal spreading sequence transmit at a higher power level than the users assigned an orthogonal spreading sequence.</claim-text> <claim-text>42. The system according to claim 41, wherein the transmitter is operable to transmit signals of users assigned a non-orthogonal spreading sequence at a power level from 3 to 6dB higher than the users assigned an orthogonal spreading sequence.</claim-text> <claim-text>43. The system according to any of claims 38 to 42, further comprising means to identify users having a high timing uncertainty, wherein in use the carrier frequencies of users having a high timing uncertainty are offset from each other in multiples of the information bit rate.</claim-text> <claim-text>44. A system for assigning spreading codes to multiple users in a CDMA system, comprising: spreading code assigning means operable to provide a spreading code having n bits and n orthogonal spreading sequences and assigning each of up to n users with one of the n orthogonal spreading sequences; and thereafter operable to assign each user in excess of n users with a non-orthogonal tO spreading sequence.</claim-text> <claim-text>45. A method for transmitting signals in a CDMA system substantially as hereinbefore described having reference to any of the accompanying figures.</claim-text> <claim-text>46. A method for assigning spreading codes to multiple users in a CDMA system substantially as hereinbefore described having reference to any of the accompanying figures.</claim-text> <claim-text>47. A ODMA system substantially as hereinbefore described having reference to any of the accompanying figures.</claim-text> <claim-text>48. A system for assigning spreading codes to multiple users in a CDMA system substantially as hereinbefore described having reference to any of the accompanying figures.</claim-text>