Circuit for simultaneously performing path search and antenna directivity control

BACKGROUND OF THE INVENTION

1. Field of the Invention:

0001The present invention relates to an automobile telephone/portable telephone system (cellular system) based on direct sequence - code division multiple access (DS-CDMA) principles, and more particularly to a path search (path timing detection) circuit for use in a base station apparatus having an adaptive array antenna.

2. Description of the Related Art:

0002One conventional path search circuit is disclosed in <patcit id="pcit0001" dnum="JP10032523A"><text>Japanese laid-open patent publication No. 10-32523</text></patcit> entitled "Reception timing detecting circuit for CDMA receiver". <patcit id="pcit0002" dnum="EP806844A"><text>EP-A-806 844</text></patcit> discloses a diversity receiver and a control method therefrom. The disclosed reception timing detecting circuit detects reception timing according to the following procedure: <ol id="ol0001" compact="compact"><li>(1) A cross correlation signal indicative of a cross correlation between a signal sequence in which a known pilot symbol sequence is spread and a received signal is determined.</li><li>(2) In order to increase the accuracy of detection of reception timing without increasing the amount of calculations for determining the correlation, the cross correlation signal is interpolated to obtain a cross correlation signal having a short sampling period.</li><li>(3) The power of the cross correlation signal is determined.</li><li>(4) The power of the cross correlation signal is averaged over a given period of time.</li><li>(5) A peak of the averaged power of the cross correlation signal is detected, and the timing of the detected peak is regarded as the reception timing of the CDMA receiver.</li></ol>

0003For determining a cross correlation signal, a received signal and a sequence in which a pilot symbol sequence is spread may be stored in a memory and read therefrom a plurality of times that are shifted from each other thereby to determine cross correlation values (cross correlation signals) with shifted delays.

0004According to the above conventional reception timing detecting circuit, it is possible to perform stable detection of reception timing (path search) even if the ratio (Eb/Io) of signal energy (Eb) per bit to interference power (Io) per Hz. Furthermore, the interpolation of the cross correlation signal allows highly accurate reception timing to be calculated without increasing the amount of calculations for determining the correlation.

0005One application of the present invention is in the art of Wideband Code Division Multiple Access (W-CDMA) for the third-generation mobile communication system. While the W-CDMA technology has not yet been put to practical use, efforts are being made to standardize its radio interface. See Specifications of Air-Interface for 3G Mobile Systems, Volume 3, Ver. 1.0 (-1.2), ARIB. According to the W-CDMA technology, known pilot symbols are added to an individual channel for each mobile terminal. Since the pilot symbols and a spreading code sequence are known by a receiver, they can be used as a known signal sequence for a correlation processor in carrying out a path search.

0006The W-CDMA technology is based on the assumption that an adaptive array antenna or smart antenna will be introduced in order to increase the capacity of communications in the future. The adaptive array antenna comprises a plurality of antenna elements arranged at a constant interval, e.g., one half of the wavelength, and the amplitude and phase of signals transmitted and received by each of the antenna elements are controlled to establish an antenna directivity optimum for the mobile terminal with which to communicate, for thereby reducing multiple access interference (MAI) at the same frequency. The adaptive array antenna or smart antenna is described in "<nplcit id="ncit0001" npl-type="s"><text>Smart Antenna Arrays for CDMA systems", IEEE Personal Communications, Vol. 3, No. 5</text></nplcit>, for example.

0007The conventional path search circuit described above is not assumed to be applicable to a radio base station apparatus using an adaptive array antenna, and performs independent processing per antenna (element). Therefore, while the conventional path search circuit is capable of effecting a normal path search if applied to a radio base station apparatus using an antenna which comprises one element, it suffers problems if applied to a radio base station apparatus using an adaptive array antenna.

0008More specifically, if the path search circuit is controlled so that the Eb/Io of a signal after it has obtained a directivity gain has a required quality as in a CDMA cellular system using an adaptive array antenna, then the Eb/Io per antenna element is of a value much lower than the Eb/Io having the required quality. Therefore, the path search circuit finds it difficult to perform a path search.

0009If an adaptive array antenna has Na elements, then the Eb/Io after the Na elements have been combined is ideally Na times the Eb/Io of each element. Therefore, if the transmitted power is controlled in a quick period of 0.625 ms so as to enable a receiver to obtain a necessary and sufficient Eb/Io as in the W-CDMA system, then a path search has to be carried out with the Eb/Io per antenna element being 1/Na as compared with the case where no adaptive array antenna is used.

0010A specific numerical example will be described below. If the number Na of antenna elements is Na = 8 and the Eb/Io after the antenna elements are combined is Eb/Io = 5 dB, then since the Eb/Io per antenna element is ideally 1/8 (- 9 dB) of the Eb/Io after the 8 antenna elements are combined, the path search circuit needs to operate with the Eb/Io = - 4 dB per antenna element. If the Eb/Io after the 8 antenna elements are combined is the same as the conventional Eb/Io with no adaptive array antenna used, i.e., if the reception quality is the same as the conventional reception quality, then the path search is required to be performed with a Eb/Io that is 9 dB lower than the conventional Eb/Io. Therefore, the conventional path search circuit fails to achieve a sufficient reception quality. Furthermore, if there are four paths having the same level, then the Eb/Io per path is Eb/Io = - 10 dB, which is a very strict condition to satisfy, making it difficult for the conventional path search circuit to detect a path with sufficient accuracy.

0011The CDMA system cannot receive a signal unless it detects proper path timing. If a path search is conducted using a signal after the directivity of an adaptive array antenna is established, i.e., after received signals from respective antenna elements are multiplied by a suitable weighting coefficient and added together, then a signal having the same Eb/Io as the conventional Eb/Io may be used to carry out the path search. The path search and the appropriate antenna directivity depend on each other in that if the path search is not carried out properly, no signal can be received, and hence the appropriate antenna directivity cannot be established, and if the appropriate antenna directivity cannot be established, then no proper path can be detected. Consequently, once path timing or antenna directivity loses a proper value due to some disturbances, no normal reception condition can subsequently be regained autonomously.

SUMMARY OF THE INVENTION

0012It is therefore an object of the present invention to provide a path search circuit capable of performing a stable path search even if applied to a radio base station apparatus using an adaptive array antenna, i.e., a smart antenna.

0013In order to achieve the above object, there is provided a path search circuit comprising an antenna having a plurality of elements, a plurality of radio receivers, a plurality of A/D converters, a plurality of correlation processors, a weighted-mean-value processor, a correlation peak detector, and a weighting controller.

0014The radio receivers frequency-convert radio frequency signals received respectively by the elements of the antenna into respective baseband signals. The A/D converters convert the respective baseband signals into digital data. The correlation processors calculate cross correlations between the digital data converted from the baseband signals and a signal known at a reception side, and output respective correlation signals. The weighted-mean-value processor weights and adds the correlation signals output from the correlation processors based on indicated weighting coefficients, and averages the weighted and added correlation signals for a predetermined number of times. The correlation peak detector detects at least one peak from the weighted and averaged correlation signals output as delay profiles from the weighted-mean-value processor, and outputs a reception level and reception timing corresponding to the detected peak as a reception level and reception timing of a reception path. The weighting controller controls the weighting coefficients to determine a directivity of the antenna and generates a plurality of weighting coefficients to establish a plurality of general antenna directivities for dividing a sector where a mobile terminal with which to communicate is present, when a communication session starts.

0015With the above arrangement, the weighting controller controls weighting coefficients to establish an antenna directivity in a plurality of directions in which a party to communicate with is possibly present, when a communication session starts. The weighted-mean-value processor calculates delay profiles for the respective antenna directivities. The correlation peak detector detects a peak from the calculated delay profiles for thereby carrying out general antenna directivity control and a path search. At the time of starting the communication session when the direction of a mobile terminal is unknown, a path search can be conducted using a signal with a Eb/Io increased by an antenna gain for increased path search accuracy.

0016Since the antenna directivity control and the path search are simultaneously carried out, an element of instability heretofore experienced, i.e., a condition in which if the antenna directivity is lost, then the path search is disabled and no signals can be received, with the result that the antenna directivity cannot autonomously be recovered, can be eliminated.

0017The above processing can be performed without increasing the amount of correlation calculations that is the greatest among other calculations in carrying out a path search. Since weighting calculations for the respective antenna elements may be effected on a relatively small sampled number of correlation signals, any increase in the amount of calculations is small even if delay profiles corresponding to a plurality-of antenna directivities are simultaneously determined.

0018The weighted-mean-value processor comprises a first memory, a plurality of second memories, and a processing means.

0019The first memory temporarily stores the correlation signals output from the correlation processors. As many second memories as the number of antenna directivities to be established store interim calculated results of the weighted and added correlation signals for the respective antenna directivities. The processing means multiplies the correlation signals read from the first memory by the weighting coefficients indicated by the weighting controller, adds the multiplied correlation signals into a sum signal, determines the power of the sum signal, updates the interim calculated results stored in the second memories based on the determined power, averages the updated interim calculated results for a predetermined number of times to obtain weighted means of correlation values, and outputs the obtained weighted means as delay profiles.

0020The weighted-mean-value processor may further comprise an interpolation filter for sampling again the weighted correlation signals at a frequency higher than a sampling frequency of the A/D converter to increase a sampling rate.

0021By passing the weighted correlation signals through the interpolation filter, it is possible to obtain correlation signals that have been sampled again with an arbitrary accuracy for thereby increasing the accuracy of a delay time.

0022The weighting controller may comprise means for generating combinations of weighting coefficients for establishing an antenna directivity in a plurality of directions in which a party to communicate with is possibly present, when the communication session starts, and reducing the number of combinations to be generated during the communication session so as to be smaller than the number of combinations to be generated when the communication session starts.

0023The weighted-mean-value processor may further comprise means for averaging the power of the correlation signals according to a moving average method or a method of generating an exponentially weighted mean in time with forgetting coefficients.

0024The above and other objects, features, and advantages of the present invention will become apparent from the following description with reference to the accompanying drawings which illustrate examples of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

0025<ul id="ul0001" list-style="none" compact="compact"><li><figref idref="f0001">Fig. 1</figref> is a block diagram of a path search circuit according to a first embodiment of the present invention;</li><li><figref idref="f0002">Fig. 2</figref> is a block diagram of a weighted-mean-value processor in the path search circuit shown in <figref idref="f0001">Fig. 1</figref>;</li><li><figref idref="f0003">Fig. 3</figref> is a diagram illustrative of the principles of operation of the path search circuit shown in <figref idref="f0001">Fig. 1</figref>; and</li><li><figref idref="f0004">Fig. 4</figref> is a block diagram of a weighted-mean-value processor in a path search circuit according to a second embodiment of the present invention.</li></ul>

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

1st Embodiment

0026Referring now to <figref idref="f0001">Fig. 1</figref>,there is shown a path search circuit according to a first embodiment of the present invention which comprises an array antenna 101 having Na antenna elements 10₁ - 10_Na that are arranged as a linear array at half wavelength intervals, Na radio receivers 102₁ - 102_Na connected respectively to the antenna elements 10₁ - 10_Na, Na A/D converters 103₁ - 103_Na connected respectively to the radio receivers 102₁ - 102_Na, a replica generator 105, Na correlation processors 104₁ - 104_Na connected respectively to the A/D converters 103₁ - 103_Na and the replica generator 105, a weighted-mean-value processor 106 connected to the correlation processors 104₁ - 104_Na, a correlation peak detector 107 connected to the weighted-mean-value processor 106, and a weighting controller 108 connected to the weighted-mean-value processor 106.

0027The A/D converters 103₁ - 103_Na convert received signals supplied from the radio receivers 102₁ - 102_Na into digital data. The replica generator 105 generates a replica of a known signal sequence contained in the received signals, e.g., a sequence in which pilot symbols in a W-CDMA system are spread.

0028The correlation processors 104₁ - 104_Na calculate correlations (complex number values) between the replica signal generated by the replica generator 105 and the digital data generated by the A/D converters 103₁ - 103_Na, and output the calculated correlations as Na correlation signals. The weighted-mean-value processor 106 weights the Na correlation signals output from the correlation processors 104₁ - 104_Na based on weighting coefficients indicated by the weighting controller 108, and adds the weighted correlation signals to each other, and averages the weighted and added signal over a given period of time.

0029A specific arrangement of the weighted-mean-value processor 106 will be described below with reference to <figref idref="f0002">Fig. 2</figref>.

0030As shown in <figref idref="f0002">Fig. 2</figref>, the weighted-mean-value processor 106 comprises memories 201₁ - 201_Na, multipliers 202₁ - 202_Na, an adder 203, a power calculator 204, an averaging unit 205, a selector 206, memories 207₁ - 207_Nb, and a selector 208.

0031The memories 201₁ - 201_Na, which have as many blocks as the number Na of antenna elements, store the Na correlations calculated by the correlation processors 104₁ - 104_Na. The multipliers 202₁ - 202_Na multiply the correlations read from the memories 201₁ - 201_Na by respective weighting coefficients W₁, W₂, ···, W_Na indicated by the weighting controller 108. The adder 203 adds the weighted Na correlations to each other.

0032The power calculator 204 calculates the power of the sum signal output from the adder 203, i.e., the sum of squares of real and imaginary parts of the sum signal output from the adder 203. The memories 207₁ - 207_Nb, which have as many blocks as the number Nb of antenna directivities, store interim calculated results of delay profiles. The selectors 206, 208 select the writing of data into the memories 207₁ - 207_Nb, and the reading of data from the memories 207₁ - 207_Nb, respectively. The averaging unit 205 averages the delay profiles over a given period of time.

0033The multipliers 202₁ - 202_Na, the adder 203, the power calculator 204, the averaging unit 205, and the selectors 206, 208 jointly make up a processing circuit.

0034The correlation peak detector 107 detects a plurality of peaks from the weighted and averaged correlation signal which corresponds to the delay profile for each antenna directivity, and outputs reception levels and reception timing corresponding to the detected peaks as reception levels and reception timing of reception paths.

0035The weighting controller 108 controls the weighting coefficients W₁, W₂, ..., W_Na to be set in the weighted-mean-value processor 106 for thereby controlling the directivity of the antenna 101. At the start of communications, the weighting controller 108 generates a plurality of weighting coefficients W₁, W₂, ..., W_Na to form a plurality of general antenna directivities for dividing a sector where a mobile terminal with which to communicate is present, and supplies the generated weighting coefficients W₁, W₂, ..., W_Na to the weighted-mean-value processor 106.

0036The principles of operation of the path search circuit according to the first embodiment will be described below with reference to <figref idref="f0003">Fig. 3. Fig. 3</figref> shows a cellular system where one cell is divided into three sectors each having an angular extent of 120 degrees. It is assumed that the base station is aware of the sector where a mobile terminal with which to communicate is present, but does not know the whereabouts of the mobile terminal in the sector.

0037The base station successively forms four antenna directivities BF1 - BF4 which divide and covers the sector where the mobile terminal with which to communicate is present, calculates delay profiles for the respective antenna directivities BF1 - BF4, and detects peaks of the obtained delay profiles, for thereby simultaneously determining which of the antenna directivities BF1 - BF4 the mobile terminal is located in and detecting path timing. Since one sector is divided by four antenna directivities, i.e., beams, in this example, an antenna gain of about 6 dB can be expected. The weighting controller 108 sets the combination of calculated weighting coefficients W₁, W₂, ···, W_Na in the multipliers 202₁ - 202_Na of the weighted-mean-value processor 106 for thereby forming the antenna directivities BF1 - BF4.

0038The path search circuit according to the first embodiment forms a plurality of delay profiles for limiting the directivity to a plurality of directions where the mobile terminal with which to communicate is possibly present, and then detects peaks from the delay profiles for simultaneously roughly determining the direction of the mobile terminal and detecting path timing. Since the path search circuit generates delay profiles after having limited the antenna directivity, it can calculate delay profiles at a higher Eb/Io than the Eb/Io of the signal per element.

0039However, when four antenna beams are generated and a delay profile is calculated with respect to each combined signal, as shown in <figref idref="f0003">Fig. 3</figref>, the amount of calculations required is very large. This is because if a path search is carried out with received signals after antenna directivities are formed, then weighting and adding calculations need to be carried out at a clock rate that is twice the chip rate or higher, resulting in the need for a considerably large amount of calculations.

0040For conducting a path search with received signals after antenna directivities are formed, if it is assumed that a W-CDMA system is employed, a chip rate = 4 MHz, pilot symbols = 6 symbols × 256 chips = 1536 chips, the number Na of antenna element = 8, a path search range L = 128 chips (reciprocal propagation delay = 31 microseconds, radius = about 4.7 km), and the accuracy of calculations of a delay profile = 1/8 chip, then the number of calculations per slot is given as follows: <ol id="ol0002"><li>(1) Calculations for generating a combined signal for one beam: 1536 × 8 × Na complex product and sum calculations.</li><li>(2) Calculations of delay profiles per beam: 1536 × 128 × 8 = 1,572,864 correlation calculations.</li></ol>

0041For four beams, calculations four times the above calculations need to be repeated per slot = 0.625 ms. No multiplications are required in the correlation calculations because the replica signal is a code signal having no amplitude.

0042It follows from the foregoing example that about 630M complex product and sum calculations per second and about 10G correlation calculations per second are necessary. However, this amount of calculations is huge and not realistic.

0043The path search circuit according to the first embodiment performs a smaller amount of calculations to achieve results that are equivalent to those of the above process.

0044Operation of various circuit elements in the first embodiment will be described in detail below.

0045In <figref idref="f0001">Fig. 1</figref>, signals received by the elements 10₁ - 10_Na of the antenna 101 are frequency-converted into baseband signals, which are then converted into digital signals. In the W-CDMA system, since the chip rate is 4 MHz and the amplitude range is 5 MHz (one side of the range is 2.5 MHz), if the received signals are sampled at a clock frequency of 5 MHz or higher, the information represented by the received signals can fully be retained according to the Nyquist's sampling theorem. It is assumed that the received signals are sampled at 8 MHz which is twice the chip rate because the sampling frequency would not be easily handled unless it is an integral multiple of the chip rate. The baseband signals are expressed by a complex number in order to handle passband signals as equivalent baseband signals. Specifically, the baseband signals are expressed by a complex number whose real part represents an in-phase component (I component) and imaginary part represents a quadrature component (Q component).

0046The replica generator 105 is a means for generating a known signal sequence in a receiving base station. In the W-CDMA system, the replica generator 105 generates a pilot symbol sequence. The correlation processors 104₁ - 104_Na calculate a cross correlation between the complex baseband digital signals equivalent to the signals received by the respective antenna elements and the replica of the pilot symbols.

0047If the number of an antenna element is represented by n (n = 1 - Na), the chip number by t (t = 1 - Nc), the lag by 1 (1 = 0 - 2L-1), the signal received by the antenna element n by Rn(2t+i), i = 0, 1, and the pilot replica signal by P(t), then the cross correlation signal Xn of the antenna element n is expressed by the following equation (1): <maths id="math0001" num="(1)"><math display="block"><mi>Xn</mi><mfenced><mn mathvariant="normal">1</mn></mfenced><mo mathvariant="normal">=</mo><mstyle displaystyle="true"><munderover><mo mathvariant="normal">∑</mo><mrow><mi mathvariant="normal">t</mi><mo mathvariant="normal">=</mo><mn mathvariant="normal">1</mn></mrow><mi>Nc</mi></munderover></mstyle><mi>Rn</mi><mo>⁢</mo><mfenced><mn mathvariant="normal">2</mn><mo>⁢</mo><mi mathvariant="normal">t</mi><mo mathvariant="normal">+</mo><mn mathvariant="normal">1</mn></mfenced><mo mathvariant="normal">×</mo><mover><mrow><mi mathvariant="normal">P</mi><mfenced><mi mathvariant="normal">t</mi></mfenced></mrow><mo mathvariant="normal">‾</mo></mover></math><img file="EP1069708B1_D0001.tif" /></maths> where P(t) with an overbar represents a complex conjugate signal of P(t).

0048Since the mobile terminal transmits an upstream signal (from the mobile terminal to the base station) in synchronism with a downstream signal (from the base station to the mobile terminal), the reception timing at the base station is delayed from the transmission timing of the downstream signal by a reciprocal propagation delay between the base station and the mobile terminal. Consequently, the base station may search reception paths in a certain delay range determined from the radius of the service area. If the path search range is 128 chips, then the reciprocal propagation delay is 31 microseconds, and the service area radius is about 4.7 km. The cross correlation may be calculated up to L = 128 chips at maximum.

0049Operation of the weighted-mean-value processor 106 will be described below. A weighted correlation signal XW_k(l) corresponding to the antenna directivity k (k = 1 - Nb) is expressed by the following equation (2): <maths id="math0002" num="(2)"><math display="block"><msub><mi>XW</mi><mi mathvariant="normal">k</mi></msub><mfenced><mn mathvariant="normal">1</mn></mfenced><mo mathvariant="normal">=</mo><mstyle displaystyle="true"><munderover><mo mathvariant="normal">∑</mo><mrow><mi mathvariant="normal">n</mi><mo mathvariant="normal">=</mo><mn mathvariant="normal">1</mn></mrow><mi>Na</mi></munderover></mstyle><msub><mi mathvariant="normal">W</mi><mrow><mi mathvariant="normal">k</mi><mo mathvariant="normal">,</mo><mi mathvariant="normal">n</mi></mrow></msub><mo mathvariant="normal">×</mo><msub><mi mathvariant="normal">X</mi><mi mathvariant="normal">n</mi></msub><mfenced><mn>1</mn></mfenced></math><img file="EP1069708B1_D0002.tif" /></maths> where W_k,n (n = 1, ..., Na, k = 1, ..., Nb) represents a weighting coefficient corresponding to the antenna directivity k.

0050The Na correlation signals (complex signals) for the respective antenna elements are used a plurality of times for calculating the delay profiles of a plurality of antenna directivities, the Na correlation signals are temporarily stored in the memories 201₁ - 201_Na. After the stored Na correlation signals are read from the memories 201₁ - 201_Na, they are multiplied by weighting coefficients indicated by the weighting controller 108 by the multipliers 202₁ - 202_Na, and then added to each other by the adder 203. The sum signal produced by the adder 203 is converted into a power value by the power calculator 204. Since the correlation signals are complex signals, the power can be calculated by calculating the sum of the squares of the real and imaginary parts of the correlation signals.

0051The memories 207₁ - 207_Nb for storing interim calculated results are initially cleared to zero, and then controlled by the selectors 206, 208 to add a correlation power value calculated in each slot and read interim calculated results to each other, and store again updated interim calculated results.

0052The above calculations are repeated as many times as the number Nb of antenna directivity patterns during one slot, thus updating Nb correlation signals (delay profiles). Since the signal to noise power ratio for correlation signals during one slot is small, resulting in a high probability to detect peaks in error, correlation signals are averaged over a given period of time until the probability to detect correct peaks becomes sufficiently high, whereupon delay profiles corresponding to a plurality of antenna directivities are output. As the averaging time increases, the signal to noise power ratio increases, improving the ability to detect correct peaks, but a time delay occurs until the peaks are detected. Particularly, while the mobile terminal is moving at a high speed, a change in the propagation during the averaging process cannot be ignored. For ordinary mobile communications, therefore, the averaging time is limited to about 50 ms to 100 ms.

0053The correlation peak detector 107 detects peaks from delay profiles corresponding to the Nb antenna directivities. For the sake of brevity, the detection of one maximum path will be described below. Nb delay profiles corresponding to the Nb antenna directivities are determined, each of the delay profiles having correlation value information in a delay range from 0 to 2L - 1. The correlation peak detector 107 determines a maximum value from Nb × 2L correlation values, and determines an antenna directivity index k (k = 1 - Nb) and a delay 1 (1 = 0 - 2L) at the time. For detecting a plurality of paths, a plurality of peaks may be detected in a similar manner. Of course, it is necessary to preclude base regions of already detected peaks (paths) from peaks to be detected.

0054If it is assumed that a W-CDMA system is employed, a chip rate = 4 MHz, pilot symbols = 6 symbols × 256 chips = 1536 chips, the number Na of antenna element = 8, a path search range L = 128 chips (reciprocal propagation delay = 31 microseconds, radius = about 4.7 km), and the accuracy of calculations of a delay profile = 1/8 chip, then the number of calculations per slot is given as follows: <ol id="ol0003" compact="compact"><li>(1) Calculations of delay profiles per antenna element: 1536 × 128 × 2 = 393,216 correlation calculations. For 8 antenna elements, it is necessary to repeat calculations that are 8 times the above calculations for each slot = 0.625 ms. </li><li>(2) Weighting calculations per beam: 8 × 128 × 2 = 2,048 complex product and sum calculations.</li><li>(3) If a 4-tap filter is used to interpolate 1/2 chip sampling into 1/8 chip sampling: 128 × 6 × 4 × 2 = 6,144 real number product and sum calculations.</li></ol>

0055For 4 beams, it is necessary to repeat calculations that are 4 times the above calculations for each slot = 0.625 ms. No multiplications are required in the correlation calculations because the replica signal is a code signal having no amplitude.

0056It follows from the foregoing example that about 13M complex product and sum calculations per second, about 39M real number product and sum calculations, and about 5G correlation calculations per second are necessary. The number of product and sum calculations is much smaller than if antenna weighting calculations are carried out before correlation signals are determined, and can be reduced to a range that can be achieved by a general-purpose DSP.

0057Using the calculated delay profiles corresponding to the plural antenna directivities, the correlation peak detector 107 detects peaks over the plural antenna directivities, so that it can simultaneously detect an antenna directivity where the reception level (correction value) is maximum and a path delay.

0058The path search circuit according to the present embodiment can perform a path search in low Eb/Io environments where the direction of a mobile terminal has not been determined and the directivity of the adaptive array antenna has not yet been established.

0059Furthermore, because the path search circuit according to the present embodiment simultaneously performs a path search and detects the general direction of a mobile terminal, their dependency on each other, i.e., the problem that once one of the path search and the direction detection fails, the other is disabled and will not be recovered autonomously, is reduced. Therefore, if the path search circuit according to the present invention is applied to a base station apparatus using an adaptive array antenna, then the stability of the path search and the antenna directivity control is increased.

0060The number of complex calculations is prevented from increasing greatly by effecting antenna weighting calculations on received signals after correlation signals are determined.

0061Once communications can be started, the direction of the mobile terminal can be monitored continuously. Inasmuch as there is a limitation on the speed of travel of the mobile terminal, the probability that the direction in which the received signal arrives suddenly changes greatly is considered to be small. Though the possibility that the direction in which the received signal arrives suddenly changes cannot be ignored if the mobile terminal is located near the base station antenna, any changes in the direction in which the received signal arrives are considered to be small in most cases. Therefore, the amount of calculations may be reduced by reducing combinations of antenna directivities calculated in a path search as compared with the case when the direction of the mobile terminal upon start of communications is unknown.

2nd Embodiment

0062A path search circuit according to a second embodiment of the present invention will be described below. The path search circuit according to the second embodiment is similar to the path search circuit according to the first embodiment except that it has a weighted-mean-value processor 106a shown in <figref idref="f0004">Fig. 4</figref>, rather than the weighted-mean-value processor 106. Those parts of the weighted-mean-value processor 106a which are identical to those of the weighted-mean-value processor 106 are denoted by identical reference characters and will not be described in detail below.

0063The weighted-mean-value processor 106a shown in <figref idref="f0004">Fig. 4</figref> differs from the weighted-mean-value processor 106 shown in <figref idref="f0002">Fig. 2</figref> in that it has an interpolation filter 301 between the adder 203 and the power calculator 204, and a multiplier 302 between the selector 208 and the averaging unit 205.

0064The interpolation filter 301 samples again the weighted correlation signal output form the adder 203 at a frequency higher than the sampling frequency of the A/D converters for thereby increasing the sampling rate.

0065The multiplier 302 multiplies the value output from the selector 208 by a certain value, and then outputs the resultant product signal to the averaging unit 205.

0066If the delay time accuracy (sampling time) of correlation signals is to be increased using an interpolation filter, then the interpolation filter may be inserted immediately in front of a stage for performing power calculations that are a nonlinear operation.

0067In the equation (2) shown above, all calculations up to the determination of XW_k(l) are a linear operation. Therefore, if a path delay is to be determined with an accuracy finer than the 1/2 chip, then it is possible to obtain correlation signals sampled at finer delay timing by interpolating XW_k(l) and passing the interpolated XW_k(l) through a low-pass filter. For interpolating the 1/2 chip into the 1/8 chip, "0" may be inserted into XW_k(l) and it may be passed through a low-pass filter, as indicated by the equation (3) shown below. For details, reference should be made to <patcit id="pcit0003" dnum="JP10032523A"><text>Japanese laid-open patent publication No. 10-32523</text></patcit> entitled "Reception timing detecting circuit for CDMA receiver".

0068If i = 0, then <maths id="math0003" num="(3)"><math display="block"><msub><mi>XW</mi><mi mathvariant="normal">k</mi></msub><mo>⁢</mo><mfenced><mn mathvariant="normal">41</mn><mo mathvariant="normal">+</mo><mi mathvariant="normal">i</mi></mfenced><mo>⁢</mo><mi mathvariant="normal">ʹ</mi><mo mathvariant="normal">=</mo><msub><mi>XW</mi><mi mathvariant="normal">k</mi></msub><mfenced><mn>1</mn></mfenced></math><img file="EP1069708B1_D0003.tif" /></maths>

0069If i = 1, 2, 3, then <maths id="math0004"><math display="block"><msub><mi>XW</mi><mi mathvariant="normal">k</mi></msub><mo>⁢</mo><mfenced><mn mathvariant="normal">41</mn><mo mathvariant="normal">+</mo><mi mathvariant="normal">i</mi></mfenced><mo>⁢</mo><mi mathvariant="normal">ʹ</mi><mo mathvariant="normal">=</mo><mn>0</mn></math><img file="EP1069708B1_D0004.tif" /></maths>

0070The path search circuit according to the present embodiment serves to increase the accuracy of a delay time by passing weighted correlation signals through the interpolation filter 301. Conversely, the path search circuit may be understood as being capable of minimizing the amount of correlation calculations required to achieve a certain accuracy of a delay time. Specifically, irrespective of the accuracy of a delay time, correlation calculations may be effected on a received signal that has been sampled at a rate that is twice the chip rate, and correlation signals sampled again with an arbitrary accuracy can be obtained by passing the correlation signals through the interpolation filter.

0071The path search circuit according to the present embodiment, which has the interpolation filter, can greatly reduce the amount of calculations for a path search because the accuracy of a delay time (sampling accuracy) required for correlation calculations may be minimum (one half of the chip rate).

0072The weighted-mean-value processor 106 which comprises the memories and the processing circuit can easily be implemented by a general-purpose digital signal processor (DSP).

0073In the path search circuits according to the first and second embodiments, as described above, the clearing of the memories which store interim calculated results, the averaging of the correlation power, and the outputting of delay profiles are repeated. However, the correlation power may be averaged according to a moving average method or a method of generating an exponentially weighted mean in time with forgetting coefficients. If the moving average method or the method of generating an exponentially weighted mean is used, then the frequency of the outputting of delay profiles may be smaller than the averaging time interval. Therefore, it is possible to improve, to some extent, a trade-off between the averaging time and the averaging delay time. If the coefficient of the multiplier 302 shown in <figref idref="f0004">Fig. 4</figref> is set to a value smaller than 1, then the averaged results in the past can exponentially be forgotten, and an average value that has been exponentially weighted in time can be determined.

0074In the second embodiment, both the interpolation filter 301 and the multiplier 302 are added to the weighted-mean-value processor 106 according to the first embodiment. However, either one of the interpolation filter 301 and the multiplier 302 may be added to the weighted-mean-value processor 106.