GB2140254A - Mobile radio system - Google Patents

Description

SPECIFICATION Mobile Radio System The present invention relates to a mobile radio system.

Radio signals are always subject to fading due to natural phenomena, but when one station of a radio link is mobile and moving at variable speeds through various and unpredictable environments, the situation is seriously compounded. In such a situation there are two types of received signal level variations observed. First there is the rapid multipath Rayleigh type fading due to different path cancellations and then there is a slower variation in the mean signal level due to gross path variations from building shadowing and other terrain effects. Both types of signal level variations are functions of the speed of the mobile.

Space diversity has been found to provide one of the best solutions to mobile radio fading. One analog mobile radio system employing space diversity is disclosed in U.S. Patent 3,693,088 issued to A. J. Rustako, Jr., et al on September 1 9, 1972. There, diversity transmission from the base to the mobile is provided by switching between two spaced base transmitting antennas on command from the mobile. More particularly, means are provided at the mobile station for determining when the signal level then being received by the mobile from a given base station antenna falls below a level which depends upon the nature of the fade itseif. When this occurs, the mobile transmits an out of message band signal back to the base which causes the base to switch to a different antenna.

A similarly operated digital mobile radio system is disclosed in U.S. Patent 4,057,758 issued to T. Hattori et al on November 8, 1 977.

There, a plurality of receiving antenna systems are switched at a constant frequency higher than the signaling rate of the digital baseband signal but less than the frequency shift width of the frequency modulated wave or iess than a product of the maximum phase shift of the phase modulated wave and the signaling rate, so that average-power dispersion in a signal element of the digital baseband signal received at the receiving antenna system is effectively compressed. Alternatively, the plurality of antennas may be transmitting antennas which are simultaneously switched to achieve compression of average power dispersion in the baseband signal elements.

The above-described analog and digital systems, however, require the flow of feedback information to control antenna switching, necessitating the use of complex and expensive apparatus at the mobile. Co-phasing of the antenna elements in an analog system has been found to provide transmission from a diversity station by means of a multi-element array, as disclosed in U.S. Patent 3,717,814 issued to M.

J. Gans on February 20, 1 973. Phase corrected intelligence signals are transmitted from a diversity array transmitter and received in-phase at a monochannel receiver. An individual piiot associated with each diversity branch and frequency separated from the pilots of the other branches is received along with the in-phase intelligence. All of the pilots are fed back, as part of the return modulation, to the diversity transmitter where they are used to establish the proper phase correction for the modulated intelligence transmission. However, the signal processing at the diversity transmitter as taught by Gans occurs at i. f., thus requiring the use of expensive RF hardware to achieve relatively accurate phase correction.

In the claimed digital mobile radio system which employs the techniques of space diversity and time-division retransmission, all the required adaptive signal processing is performed at baseband at a base station location.

It is preferred to employ digital co-phasing techniques at the base station of a digital cellular mobile radio system in conjunction with the above-mentioned space diversity and retransmission properties to overcome the mobile radio transmission-related problems of intercell interference, shadow fading and Rayleigh fadings.

A preferred embodiment of the present invention will now be described, by way of example, with reference to the accompanying drawings, in which: Fig. 1 illustrates the transmission of a signal from a mobile to a base station including a plurality of antenna elements and the random phase associated with each element; Fig. 2 illustrates the co-phased transmission of signals from the base station back to the mobile of Fig. 1, where the signals are pre-phased at the base so that they will be coherent upon reception by the mobile; Fig. 3 illustrates an exemplary signaling frame including two reference intervals and two message intervals for implementing time-division retransmission between a mobile and a base station as illustrated in Figs. 1 and 2;; Fig. 4 contains a detailed embodiment of an exemplary diversity branch of the plurality of branches employed at a base station embodying the present invention.

Mecanical constraints, cost, and maintenance requirements all suggest that the equipment employed at the mobile stations of a mobile radio system should be kept as simple as possible. This goal has been achieved with time-division retransmission, while retaining the advantages of space diversity processing. Instead of using a different frequency for each direction of communication, two-way communication between mobile and base station is conducted on a single time-shared channel. These basic principles of operation may be understood by reference to Figs. 1 and 2, which illustrate mobileto-base and base-to-mobile communication, respectively.

In the mobile-to-base communication scheme illustrated in Fig. 1, a mobile station 10 containing a single antenna element 1 2 transmits a message to a base station 14 which contains a plurality of antenna elements 1 6i, 162... 1 6M, where the plurality of antenna elements are employed to provide for space diversity at base station 14. The detailed structure of the message transmitted from mobile station 10 to base station 1 4 will be explained in greater detail hereinafter in association with Fig. 3, but in general, mobile radio reception is characterised by large fluctuations in received signal power, P, at base station 1 4 as mobile 10 travels along a street.

This variabiiity can be modeled as the product of three factors, as shown by P(r)=lrl "S(r)R2(r), (1) wheres is the position vector denoting the location of mobile 10 relative to base station 14.

The first factor, lrl~n, represents the general reduction in signal strength as mobile 10 recedes from base station 14. In free space, n=2, but in an urban environment it can be shown that n is in the range of 3 to 4. The second factor, S(r), represents shadow fadings, which is primarily the result of blockage due to large objects such as buildings and hiils. It has been found by measurement of S in several cities that it is approximately a log-normal random variable. The third factor, R2(r), in equation (lYrnprnsents Rayleigh fading, a phenomenon caused by the random addition of signals arriving at an antenna via multiple paths.The amplitude of the received envelope, R, may be modeled as a random variable with a probability density function p(R)-=2Re-R2, (2) Therefore, in accordance with the above described random properties of signal strength, such as shadow fading and Rayleigh fading associated with transmission from mobile-to base, each signal received at a separate antenna element 1 6i through 16M located at base station 14 will possess an independent random phase,0 through oM, respectively.As will be described in greater detail hereinafter, each antenna element processes its associated received signal in its associated retransmission branch 1 8i through 18M, respectively, to delete the random phase so that the M received signals may be added coherently in combiner 20 at base station 1 4. In addition to adjusting the phases of the M received signals, each retransmission branch 1 8i through 1 8M also functions to adjust the respective weight of the received signal passing therethrough to achieve the optimum net signal-to-interference ratio (SIR) at base station 1 4. For equal power Gaussian interference at each retransmission branch, it can be shown that the best net SIR is achieved with a maximal-ratio combiner.

Synchronization of reception of the plurality of antenna elements is achieved by employing a clock 31.

The reverse transmission operation is illustrated in Fig. 2, where the message to be transmitted back to mobile 1 0 from base station 14 originates from a message source 21 and is applied via clock 31 as an input to each retransmission branch 1811 8M. The base-tomobile message signal is adapted by applying thereto the conjugate (negative) of the abovedescribed random phases at each retransmission branch 1 811 8M associated with antenna elements 1611 6M, respectively.More specifically, retransmission branch 1 8i applies a phase shift of -8, to the base-to-mobile message signal, retransmission branch 1 82 a phase shift of 02, and so on, with retransmission branch 1 8M applying a phase shift or 0-8M to the base-to- mobile message signal. These excitation phases O1 through 8M exactly compensate for the different phase delays experienced by the baseto-mobile message signals so that the transmission medium "undoes" the conjugate phase shift applied at each retransmission branch, thereby allowing the M signals to be received coherently at mobile 10.Therefore, since reception at mobile 10 will always be coherent as shown in the vector diagram of Fig. 2, associated with mobile 10, the receiver employed by the mobile may be extremely simple in form and yet provide adequate reception of the signal transmitted by base station 14.

A single frame in mobile-to-base and base-tomobile transmission illustrating time-division retransmission is shown in Fig. 3. The basic frame consists of four time intervals including two reference intervals and two message intervals.

Starting at time T=O, and assuming that the base station and all mobiles communicating with that base station are synchronised, a carrier burst is transmitted from the mobile 10 to the base station 14 during a reference interval Rl1. The carrier burst transmitted by the mobile 10 enables the base station 14 to identify the mobile and cophase its antenna elements accordingly. The burst repetition rate is chosen to be rapid enough to ensure that the multipath conditions do not change significantly during the subsequent message transmission.

During the carrier burst transmission it is important that interference from unwanted mobiles be minimized so that the co-phasing of the antenna elements can be performed accurately. This minimization of interference may be accomplished, for example, by a time-division reference transmission method or a frequency offsett reference transmission method. An illustration of a particular time-division scheme is included in the expanded version of Rl1 included in Fig. 3. Here, the reference interval is divided into a plurality of unique time slots, labeled in this example, A through H, which are associated in a one-to-one relationship with eight separate pairs of communicating mobile and base stations, which are capable of interfering with each other.

Therefore, if an exemplary base station is adapted to receive communication during, for example, a reference sub-interval B, the mobile desiring to communicate with that base will transmit its reference signal during the same sub-interval B.

Thus, since each base station gates its receiver "on" only during its pre-assigned time slot in the reference interval, interference from other mobiles will be minimal. In a frequency-offset reference transmission scheme, each mobile and corresponding base station in a particular area is assigned a unique offset frequency which is a multiple (0,+1,+2,+3) of a lowfrequency Q=2t1/T, where T is the duration of reference interval Rl1. During the reference interval, the transmitting frequency of a mobile and the local oscillator at the base station are shifted from the carrier frequency 9c by the offset assigned to its associated base station.The use of different reference frequencies allows the base station to select the desired reference signal and suppress the interference. The choice of Q=2f1/T allows the various reference signals to be orthogonal; unwanted signals do not contribute to the cophasing operation of the base station.

Once the base station identifies the mobile by its associated reference signal, the mobile then transmits its message to the base during message interval Ml1. For purposes of illustration only, to achieve, for example, a 32 kbit/sec transmission rate, which is necessary for speech transmission, 64 bits must be sent during the message interval of, in this example, 790 usec, implying a baud rate of 81 kbaud/sec. This exemplary message interval of 790 jusec was determined by assuming that the entire mobile-to-base transmission interval is 1 msec, with 210 zbsec reserved for reference interval Rl1. Depending on filtering and tolerable dB penalty, this exemplary rate would require 80-120 kHz bandwidth with binary PSK modulation.

In a cellular mobile radio system employing the above-described space diversity properties, cochannel interference from unwanted mobiles is effectively rejected and the same frequency channel may therefore be used in cells much closer together than is the case with existing analog systems. Therefore, fewer distinct channel sets are required and each cell is able to occupy a larger share of the total system bandwidth. Thus, for the example above, the number of mobiles that can be served in the 40-MHz bandwidth of the 850 MHz mobile radio band by employing the digital transmission techniques of the present invention is approximately 1 30. This high capacity of 1 30 mobiles/base illustrates the advantage of the digital system of the present invention over existing analog systems which have a much smaller capacity.

At the completion of message interval Ml1 the mobile transmits a second carrier burst during reference interval Rl2 to update the location information of the mobile with respect to the base, where the second carrier burst may be either one of the time-division or frequency-offset forms described hereinbefore. Once the location information has been updated, the message from the base to the mobile is transmitted during message interval Ml2. Like the time interval associated with Rl1 and Ml1, the reception of location update during Rl2 and base-to-mobiie transmission during Ml2 also occurs, in this example, during a 1 msec period.Assuming, for example, that Rl2 is also 210 Xusec in duration, the same baud rate of 81 kbaud/sec is associated with transmission during Ml2. At the end of 2 msec, therefore, an entire message cycle has occurred and the entire process starts again.

The signal processing circuitry for an exemplary antenna element 1 6j and its associated retransmission branch 18 is illustrated in Fig. 4 and may be analyzed in conjunction with the timing sequence illustrated in Fig. 3. In the following discussion, any reference to the timedivision or frequency-offset signaling schemes will be omitted for the sake of clarity however, the use of these schemes to avoid co-channel interference is an obvious extension of the principles described hereinafter.

The reference signal received at an exemplary antenna element 1 6j from a mobile (not shown) is of the form Ricos(w,t+Bi), where Wc is the carrier frequency, Rj is the Rayieigh amplitude and Oj is the previously described random phase. Although both Rj and Oj are functions of time, they vary slowly and may be considered as remaining constant during reference interval Rl1.

The reference signal received by an antenna element 16j passes through a circulator 1 7 and into a switch 19, where switch 1 9 controls the transmit and receive modes of operation of retransmission branch 18j. During reference interval Rl1 message interval Ml1 and reference interval Rl2, switch 1 9 remains in its "receive" position. After passing through switch 1 9 the reference signal is transmitted along two distinct signal paths of retransmission branch 181, and Irail and a Q-rail. The signal on the I-rail is applied as one input of a mixer 23, where the other input to mixer 23 is a local oscillator 22 which generates a cos ct signal.The output of mixer 23 is one quadrature component of the reference signal Rjcos(csct+O;), specifically, RjcosOp In a like manner the signal on the O-rail is applied to one input of a mixer 25, where the remaining input to mixer 25 is a local oscillator 24 which generates a sino,t signal. The output of mixer 25 is, therefore, the remaining quadrature component of reference signal Rjcos(w,t+Bi), specifically, --RisinBi.

Phasing of antenna element 16j to receive a message signal possessing random phase 0 is accomplished by passing the down-converted signals through separate reference coefficient generator circuits via a switch 27, the signal RjOOS8j through a reference coefficient generator circuit 26 and the signal RsinQ.thrnugha reference coefficient generator circuit 28.

Generators 26 and 28 produce reference coefficients aRjcosOj and -o'R1sin Q, respectively, where for the above-described time-division scheme reference coefficient generators 26 and 28 can be sample-and-hold circuits, which samples the carrier burst transmitted from e mobile during its preassigned sub-interval, as described hereinbefore. In accordance with the frequency-offset scheme, reference coefficient generators 26 and 28 may be of the form of wellknown integrator circuits which integrate over the entire reference interval.Thus, reference coefficients aRicos8j and -αRjsin#j modify the signal received by antenna element 16j by phasing element 16j to receive a message signal with random phase Oj, where a is a constant generated during the reference coefficient process.

At the completion of reference interval Rl1, a switch 27 is activated from a first to a second position by a clock signal from clock 31 to switch the outputs of mixers 23 and 25 from the inputs of reference coefficient generators 26 and 28 to the inputs to a pair of multipliers 30 and 32. In Fig. 4, reference coefficient generator 26 has its output coupled to an input to multiplier 30 and likewise, reference coefficient generator 28 has its output coupled to an input to multiplier 32.

The message transmitted from mobile-to-base during message interval Ml1 comprises, for example, 64 bits, where the kth bit and accompanying interference may be written as AkRjcos(#ct+#i)+Iccos#ct+lssin#ct, (3) where Ak=+l represents the transmitted bit and lc and Is are Gaussian random variables with zero mean and variance 52 The message bit is a down-converted in retransmission branch 1 through mixers 23 and 25 in a like manner as the above-described reference signal to form its quadrature components.The down-converted quadrature components of the kth bit appearing at the outputs of mixers 23 and 25 are then applied via switch 27 as inputs to multipliers 30 and 32, respectively, where the remaining inputs to the multipliers are its associated reference coefficient appearing at the outputs of reference coefficient generators 26 and 28, respectively. Specifically, the down-converted message bit on the I-rail from mixer 23 and reference coefficient aRjcO5O from reference coefficient generator 26 are applied as inputs to a multiplier 30, and the down-converted message bit on the Q-rail from mixer 25 and reference coefficient --RisinBi from reference coefficient generator 28 are applied as inputs to a multiplier 32.The output signals of multipliers 30 and 32 may be represented respectively by: a:AkR2jcos2flj+a:1cRJcos0J for the I-rail and (4) aAkR2jsin20 lsRjsinS for the Q-rail.

These two signals are subsequently applied to an adder 34, resulting in an output signal of aAR2i and a mean-square noise term of a2R2js2.

Note that multiplication of the quadrature components of the message bit by their associated reference coefficients and subsequent summation thereof produces a demodulated, i.e., baseband, signal which is independent of the random phase OJ. Therefore, the demodulated signals produced by the remaining retransmission branches 1 811 8M (not shown) will likewise be independent of their respective random phases f MT and further, each signal possesses a magnitude proportional to R2. Therefore, combiner 20 of base station 14 may comprise only a simple adder circuit to achieve the abovementioned optimal, maximal-ratio combination of the signals produced at the base station by retransmission branches 1811 8M.

Fig. 4 may also be used to illustrate the signal flow from the base station back to the mobile. At the completion of the mobile-to-base message during message interval Ml1 switch 27 is reactivated and the outputs of mixers 23 and 25 are switched back to their first positions as inputs to reference coefficient generators 26 and 28, respectively. The mobile then transmits a second carrier burst during reference interval Rl2 to update its propagation location information with respect to the base station.In general, this reference signal will travel through slightly different propagation conditions than the signal transmitted during Rl1. The up-dated random phase value Oj stored in a like manner as the preceding 0 by reference coefficient generators 26 and 28 will enable the base station to compensate for this new value of Op At the completion of reference interval Rl2, switch 27J is once again activated and the outputs of mixers 23 and 25 are switched back to their alternative positions as the inputs to multipliers 30 and 32, respectively. Also, switches 19 and 35 are activated at this time to switch retransmission branch 18J from its "receive" mode to its "transmit" mode.The base 14 is then prepared to begin transmitting a message signal during message interval Ml2 back to the mobile. The message signal originates from message source 21, and the identical signal is introduced to the Iand Q-rails of each retransmission branch 1 81 1 8M via switch 35 and adder 34. For transmission back to the mobile, the signal flow along the Iand Q-rails is reversed from that hereinbefore described with the reception of signals during periods Rl1, Ml1, and Rl2, with the phase conjugation described hereinabove in association with Fig. 2 accomplished by inverting the sign of the reference coefficient produced by reference coefficient generator 28. More particularly, the input message signal is directed by adder 34 to flow along the I- and Q-rails, where on the I-rail the message signal is applied as an input to multiplier 30, the other input to multiplier 30 being the updated reference coefficient stored in reference coefficient generator 26. On the Q-rail, the message signal is applied as an input to multiplier 32, the other input to multiplier 32 being the negative of the updated reference coefficient stored in reference coefficient generator 28, where the use of the negative, as described hereinabove, in conjunction with the processed signal on the i-rail, will allow the signal to be received coherently by the mobile station by "undoing" the effects of the environment. The outputs of multipliers 30 and 32 are then applied as inputs to mixers 23 and 25, where local oscillators 22 and 24 are also applied as inputs to mixers 23 and 25, respectively, to upconvert the quadrature components of the message signal.

The upconverted signals are then combined and passed via switch 1 9 through circulator 17 and transmitted by antenna element 16. This procedure gives the same SIR at the mobile as if all the transmitted power were radiated from a single antenna and the mobile included the same degree of diversity (M-branch in this example) as the base station.

Claims (6)

1. A time-division retransmission method of transmitting digital communication signals between at least one base station and at least one mobile station, the method comprising the steps of (a) transmitting a first reference signal from the at least one mobile station during a first reference interval, (b) transmitting message information from said at least one mobile station to said at least one base station during a first message interval, (c) transmitting a second reference signal from the at least one mobile station to the at least one base station during a second reference interval, and (d) transmitting message information from said at least one base station to said at least one mobile station during a second message interval.

2. A method according to claim 1, wherein the at least one base station comprises a plurality of base stations and said plurality of base stations are capable of communicating with the at least one mobile station, the method comprising the further steps of (e) in performing step (a) transmitting a first carrier burst from the at least one mobile station which is associated with one of the plurality of base stations during a predetermined sub-interval of the first reference interval, each base station of said plurality of base stations being associated with a separate and distinct sub-interval of said first reference interval; and (f) in performing step (c) transmitting a second carrier burst from the at least one mobile station which is associated with one of said plurality of base stations during a predetermined sub-interval of the second reference interval, each base station of said plurality of base stations being associated with a separate and distinct sub-interval of said second reference interval.

3. A method according to claim 1, wherein the at least one base station comprises a plurality of base stations and said plurality of base stations are capable of communicating with the at least one mobile station, the method comprises the further steps of (e) in performing step (a) transmitting a first carrier burst from the at least one mobile station to the plurality of base stations at a predetermined off-set frequency, each base station associated with a separate and distinct off-set frequency; and (f) in performing step (c) transmitting a second carrier burst from the at least one mobile station to the plurality of base stations at a predetermined off-set frequency, each base station associated with a separate and distinct off-set frequency.

4. A method as claimed in claim 1, including (a) receiving at said base station a reference signal transmitted by the mobile station; (b) down-converting the reference signal into a reference signal first quadrature component and a reference signal second quadrature component; (c) forming a first reference coefficient associated with the reference signal first quadrature component and a second reference coefficient associated with the reference signal second quadrature component;; (d) receiving at said base station a message signal transmitted by said mobile station, (e) down-converting the message signal into a message signal first quadrature component and a message signal second quadrature component, (f) multiplying said first reference coefficient with said message signal first quadrature component to form a first baseband component, and multiplying said second reference coefficient with said message signal second quadrature component to form a second baseband component, and (g) adding the first and second baseband components to form a baseband digital communication signal.

5. A method according to claim 4, wherein in performing frequency-offset reference transmission the method comprises the further step of (j) in performing step (f), integrating each of the first reference signal first and second quadrature components to form the first and second reference coefficients, respectively.

6. A method according to claim 4, wherein in performing time-division reference transmission the method comprises the further step of (j) in performing step (f), sampling and holding each of the reference signal first and second quadrature components to form the first and second reference coefficients respectively.