SE516355C2 - Receiver for code division multiple access communication system, has processors transforming digital signals into sufficiently high target power level representing short term dynamic range for each signal input path - Google Patents

Abstract

Processors (12,13) in each input path (17) converts received input signals to baseband signals and A/D converter (14) converts baseband signals into sequence of digital signal samples. Another processor (16) in one input path transforms digital signal into a sufficiently high target power level representing short term dynamic range that is common for each signal input path and forwards signal samples to demodulator (11).

Description

25 30 516 355 šïï * Ilšï- Ilšï 2 Another aspect, specific to the multiple access used in CDMA, relates to the fact that all users transmit broadband signals at potentially the same time and use the same bandwidth. Due to this fact and the propagation effects described above, a base station will receive broadband signals from different users within a wide dynamic range, for example up to 80 dB, depending on the received signal power. In addition, since the signal power received from an individual user may change during transmission and vary depending on the receiving antenna, each user transmission is exposed to an extended dynamic range. Therefore, in order to be able to receive signals from all users within the coverage area for a base station, for example, it is important for this base station to have means for monitoring and controlling the total received uplink transmission power per user as well as the total received uplink transmission power from all users for to ensure both that signals from all transmitting users, regardless of signal propagation effects, arrive at the base station with practically the same average power and that the total interference level does not become too high with the resulting reduced system capacity. This is achieved by estimating the total power received at each base station antenna and by using this estimate to regulate the output power of the transmitting user equipment for each base station.

Another measure to be able to handle signals with a large dynamic range is to apply an automatic gain control unit (AGC unit) in the analog part of the receiver. JP-A-8/335 928 discloses, for example, a receiver in a CDMA-type mobile communication system comprising, inter alia, an AGC unit (30) which controls the gain in accordance with the maximum amplitude of the analogue baseband signals and outputs a constant output signal (S30).

THE INVENTION IN BRIEF A first object of the present invention is to provide a device and a method for simplifying and improving the processing of the signals in the demilter part of a receiver unit, preferably in a CDMA-based communication system.

Another object of the present invention is to simplify and improve this processing for digital signal sample sequences by reducing the number of bits used to encode the samples without losing any information transmitted by these signals. Yet another object of the present invention is to enable the demodulator to use substantially identical signals received from your signal inputs, such as base station antennas, to achieve a reliable signal to be used for demodulation. Yet another object of the present invention is to achieve an improved digital automatic gain control unit (AGC unit) which is less sensitive to external noise interference and which requires a minimum of hardware.

Briefly, these and other objects of the present invention are achieved by signal processing means arranged in front of the demodulator in at least one of the signal input paths, which means comprise at least one digital AGC unit and a subsequent quantizer unit. The digital AGC unit rates the sequences of signal samples to a common and constant power level. The quantizer unit can then transform the graded signal samples into samples using a reduced number of bits. To use the AGC effectively, a control error is determined, which is used to maintain the signal quality of the demodulator input.

The invention is based on the insight that the input signals from the different antennas on a receiver have a large long-term dynamic range, ie large differences in the signal power due to varying reception characteristics, but normally a much smaller short-term dynamic range, especially in CDMA-based communication system. This means that the corresponding digitized input signals must be coded 10 20 25 30 nnn nnn nn nn nn n nn n nn nnn nnnnnnnnn I nn nn nn nn be able to represent the large area. This number of coding bits can be reduced by transforming the input signals to a common power level that only represents the short-term dynamic range.

As a first advantage, the present invention comprises an apparatus and method that reduces the number of bits needed to represent signal samples fed to a demodulator.

Another advantage of the present invention is thus that the required transmission bandwidth, i.e. the number of transmitted bits per second, is reduced in the receiver. Yet another advantage of the present invention is that the processing of the signal samples in the subsequent demodulator is considerably less complex, which means reduced power consumption and saved chip area. Yet another advantage of the present invention is an improved digital AGC unit that is less sensitive to external noise interference. In a CDMA-based communication system, the AGC can be implemented by reusing existing power estimators for received signals. Yet another advantage of the present invention is that the demodulator can be quickly switched to antennas which momentarily best receive a signal, for example in cases with a softer hand-off.

Other objects, advantages and novel features of the invention will become apparent from the following detailed description of the invention when considered in conjunction with the accompanying drawings and claims. BRIEF DESCRIPTION OF THE DRAWINGS For a better understanding, reference is made to the following drawings and preferred embodiments of the invention.

Fig. 1 shows an overview view of a base station receiver and in particular the uplink paths of the uplink to the demodulator of the base station.

Fig. 2 shows in more detail the signal processing means for one of the signal paths according to a preferred embodiment of the present invention.

Figs. 3a and 3b show two examples of assignment functions that evaluate the relationship between the power estimate for a signal sample and the grading factor that is necessary to transform the signal sample to a certain target power level.

Fig. 4 is a flow chart illustrating the main steps of the method according to the preferred embodiment of the present invention.

DETAILED DESCRIPTION Fig. 1 relates generally to a base station, for example in a CDMA-based communication system, and in particular to the part of a base station receiver 10 which is responsible for forwarding the received signals to its demodulator 11. The base station generates communication services for the user equipment within its cell, which is usually divided into a number of sectors and is equipped with at least one antenna per sector or preferably two antennas to improve signal reception by utilizing scattering effects.

Transmissions from a user equipment within one of the sectors are mainly received by the antennas directed at this specific sector, but also from antennas for the neighboring sectors. The base station can thus use various received copies of a transmitted signal from a user equipment. This is an important precondition, for example with respect to a softer hand-off, ie when a base station detects a better quality of the received signals from another sector antenna and switches the reception to this antenna. The received signals differ considerably in signal quality and are received at different power levels that cover a large dynamic range depending on, for example, different propagation and attenuation characteristics of the current radio channel. In addition, said differences may also be due to the fact that the signal is received from an area outside the appropriate coverage area of the antenna.

The present invention therefore aims to enable the base station receiver 10, and in particular the demodulator part 11 in the receiver 10, to use different received signal copies and to extract an optimized signal which can be used for demodulating the original user information.

Accordingly, the receiving part 10 of the base station for each of its antennas comprises a separate input path 17 for uplink signals to be forwarded to the demodulator 11. Initially, the received high frequency signals are pretreated by amplification and filter means 12, downconverted to an intermediate frequency (IF) 13. and are represented as sequences of 16-bit signal samples by means of an analog-to-digital converter 14 with a subsequent low-pass filter 15. Using the signal processing means 16 according to the present invention, the digitized input signals are then transformed to a sufficiently high target power level, which is common to all the signal input paths 17 and thus make the signals suitable for the demodulator 11 as described above. For each of the signal input paths, the signal processing means 16 consists of at least one digital AGC unit 161 for grading the sequences of the input signal samples and a quantizer unit 162 for reducing the number of bits necessary to encode the signal samples. The adapting unit may optionally also comprise a conventional analog AGC unit 163, which is arranged, for example, as a previous signal attenuation unit for the analog signal.

Fig. 2 describes in more detail the various parts of the signal processing means 16 according to the present invention and how these parts cooperate in accordance with the preferred embodiment according to the invention. This signal processing means can be divided into a number of blocks: the main blocks are a grading block 21, which grades the input signals depending on signal power estimates from a power estimating unit 22, and a quantizer block 23 which outputs a graded output signal. Furthermore, the fitting unit comprises displacement compensating means 24 and, optionally, a conventional AGC unit 25.

The central part of the signal processing means consists of the grading block 21, which comprises means 211 for determining a grading factor from the estimated power values of a received input signal X [n] and means 212 for grading this input signal X [n] to an input signal Y [n]. which still carries the same information. When these input signals are considered as a sequence of digital signal samples Xk [n], where k denotes the time interval index, the grading means that each of these signal samples is adjusted to a common power target level ozmf by multiplication by a suitably determined grading factor otk, i.e. Yk [n] = ak - Xk [n]. The gradation block 21 thus compensates for the varying average power levels of the input signal samples, ie reduces the long-term dynamic range of a signal which does not contain any user information but mainly represents the stop of the radio channel. This results in a graded output signal Y [n] with a total dynamic range corresponding to the short-term dynamic range of the input signal X [n], which is generally much smaller than the long-lasting dynamic range.

Consequently, the coding of the signal samples no longer has to represent the redundancy caused by the significantly different signal power levels. Instead, the quantizer unit 23 can reduce the number of coding bits per sample to a number sufficient to represent signal variations within the short-term dynamic range.

Assuming a sequence of input signal sampler Xk [n] with an approximately constant variance 02k per sample, the grading can be performed by a multiplication by a grading factor which is constant for the time interval of the kzte sample. Then 2 _ O-ref ak _ __? 'Uk Calculation of the grading factor ork for an input signal Xk [n] requires buffer storage of the fi nier grading factor as the input data stream during the entire time interval while the variance 02k is calculated. However, if it can be assumed that the probability of significant changes in the average power level of the received signal, ie the variance, between two consecutive frames is small, the buffer storage can be dispensed with using the variance estimate from the previous frame to calculate Yk [n]. Therefore, in the method according to the preferred embodiment of the present invention, the following equation is used to recover the graded output signal Yk [n]: Yk [n] = ork_1 - Xk [n] In the equation above, the gradation function has been described as a multiplication. As the grading in the AGC is performed on a sample basis, strong demands will be placed on the multipliers that perform the current grading. However, these requirements can be substantially reduced if the grading factor is quantized to a value that can be represented by the power of two. In this case, the grading factor ork can be interpreted as a shift factor ßk, ie ak = 2/3 * <=> ßk = log2 ak.

Grading takes place by means of a left shift of a signal sample in accordance with the shift factor ßk. To avoid excessive steps of the power factor grading due to this quantization, it is possible to quantize this factor to a value that can be represented as the sum of powers of two. In this case, a grading factor is represented by a set of shift values Him) consisting of the exponents 10 20 25 30 516 3 55 ÉQÄÉÉ- fÃÉf 9 of the sum terms. Grading is done by left-shifting a signal sample for the number of positions indicated by each shift factor in the set and adding these shifted parts.

The gradation block as described above assumes that the input signals X [n] have a mean value of zero. This is necessary because the grading operation takes place by means of a multiplication. In practice, however, the radio receiver's hardware would introduce a bias voltage to the input signal, which can be considered as a time varying process with a very large time constant. This deviation is eliminated by the introduction of an offset compensating block 24, which comprises means 241 for estimating the signal average value by using a low pass filter with a socket and a large time constant. Any signal shift is eliminated by subtracting the estimated average value from the input signal. An analog AGC unit 25 can optionally be inserted to reduce the total dynamic range of an incoming signal. This unit is designed as an attenuation unit that is controlled by power estimates from the power estimator 22.

A control unit 251 activates a certain attenuation level in response to the estimated signal power level. The analog AGC unit should be designed as a step attenuation unit comprising active modes with a certain attenuation and an inactive mode with low attenuation. The control unit 251 activates an attenuation level if the signal power level exceeds a certain threshold level and deactivates when the signal power level drops below the threshold. By means of the attenuation, incoming signals covering a very large dynamic range can be attenuated, for example to avoid an undesired clipping in the subsequent analog / digital converter.

As described above, the grading factor for each signal sample is defined using estimation of the variance O 2k of the input signal sample Xk [n]. However, the grading factor is calculated using power estimates obtained from a power estimate block 22. To reduce the hardware complexity required for this operation, however, the abnormal variance Pk is calculated in 10 15 20 25 30 5 1 5 5 5 5 gi: llÉï- 10 instead of said variance estimates å:, ie f; = 2Nó-f. The power estimation block 22 thus consists of the estimating means 221 for calculating a power estimation and filter means 222, namely a low-pass filter with a filter outlet, in order to avoid sharp changes. The hardware complexity for calculating power estimates can be further reduced by reusing an already implemented power estimate function block: Since it is crucial to minimize the total interference between sectors and between different base stations in a CDMA-based communication system, means are provided for estimating power estimates. the total power received at the antenna in each sector and to control and minimize the output power from the mobile units in the sector. These means needed for good sister function can also be combined with the device and the method according to the present invention and thus replace said estimating means 221.

Figs. 3a and 3b show logarithmic presentations of examples of assignment functions for a suitable grading factor and with respect to a calculated variance level Pk for a certain signal sample Xk [n]. These functions are fed to approximate means 211 for determining and assigning discrete grading factors in accordance with certain ranges of the estimated variance levels Pk. The diagram in Fig. 3a represents a first approximation to a grading factor assignment based on the calculated variance value within certain intervals. If it is assumed that the common signal power target level has been chosen higher than the levels of the received signal power, it is obvious that the grading factor must be chosen larger the lower the received power level. Since the power estimate values are sensitive to noise, two values at the boundaries between two adjacent power ranges and thus cause an undesired oscillation in the graded signal.

Therefore, according to another preferred solution, as shown in Fig. 3b, the assignment of grading factors also comprises a relay control at the boundary between the power ranges which prevents the assignment of oscillating grading factors. The power jump value for changing the grading factor varies depending on which side the power value approaches the ordinary jump value, ie the boundary line between the two adjacent 10 20 25 30 ll power intervals. For decreasing power values, the jump value is slightly smaller than the ordinary jump value, and correspondingly for increasing power values, the jump value is slightly larger than the ordinary jump value. Within a narrower power range around the limit, the change of the grading factor is thus delayed in order to prevent an oscillation between two grading factors. The insensitivity to power fl uctuations depends on the width of the narrow power range.

Fig. 4 shows a flow chart illustrating the main steps of the method according to the invention for transforming an input signal sequence Xk [n], block 41, to a common and constant target variance level ozref. In a first step, block 42, the received signals can be adjusted, if necessary, to a common zero average value in order to be able to perform a grading operation by a multiplication. Then, for each signal sample Xk [n], a power estimate Pk, block 43, is calculated, and a grading factor otk is determined depending on said calculated power estimate Pk and on the target power level ozæf, block 44. Thereafter, each of the received signal samples Xk [ n] to a signal Yk [n] = ak-, - Xk [n] at said target power level, block 45, and fed to the demodulator, block 46. In an optional step, block 47, the estimated power level Pk in AGC the unit is arranged to initiate an attenuation of the analog received signal if the power estimates exceed the threshold values.

From what has been described above, it is obvious that the graded signal is disturbed by a signal error. This error is firstly due to the fact that the grading factor is only updated at regular time intervals and secondly to the fact that the grading factor is quantized with respect to distinct intervals for signal power estimates and, possibly, quantized to a value that can be represented as a sum of the powers of two. This means that a subsequent unit, for example the demodulator, which assumes signal samples that are graded to a common variance level, receives and processes a signal that can have a considerable deviation from the common variance level. As a first solution, the grading factor can be updated more quickly. However, as this increases the complexity of the subsequent demodulator, in the preferred alternative embodiment of the present invention, the grading block 21 of the digital AGC unit of Fig. 2 generates the side information 26 relating to the present invention. the quality of the grading and quantization, which is updated faster than the grading factor and fed to the demodulator ll together with the graded signal samples. This side information 26 is typically the control error of the AGC algorithm due to the quantization of the grading factor. When this side information is used, the demodulator can easily and quickly connect to the antenna that currently best receives the input signals, for example when a softer hand-off is performed.

Claims (19)

10 15 20 25 516 355 13 Patent claims A device in a receiver unit (10) for a communication system, which receiver unit (10) is equipped with at least one signal input (17) for receiving signals from a transmitter in the communication system, which signals have a substantially less short-lived dynamic range compared to the long-lasting dynamic range, the receiver unit (10) comprising means (1 1) for demodulating received signals, a separate input path from each signal input (17) to the demodulator (11), first processing means (12, 13) in each of the input paths comprising at least means (13) for converting the received signals into baseband signals and means (14) for converting the baseband signals into sequences of digital signal samples, characterized by other processing means (16, 20) in at least one of the input paths. to adapt the signal samples corresponding to a signal power adjustment to a common power level that only represents the short-term dynamic range, and to forward the signal samples to the demodulator (11). Device according to claim 1, characterized in that the second processing means (16, 20) comprise at least one digital automatic gain control unit (AGC unit) (161) and a subsequent quantizer unit (162). Device according to claim 2, characterized in that the digital AGC unit (161) comprises first means (211) for determining and assigning grading factors from power estimates for each of the digital signal samples and other means (212) for grading. of a digital sample towards a common power level using the assigned grading factor. 10 20 25 30 516 355; :: = išfgåï- 14 Device according to claim 2, characterized in that the digital AGC unit (161) comprises means (26) for supplying side information regarding the quality of the gradation directly to the demodulator (11). Device according to claim 2, characterized in that the second processing means (16, 20) comprise means (221) for determining power estimates for the digital signal samples and means (222) for low-pass filtering of the power estimates. Device according to claim 2, characterized in that the second processing means (16, 20) comprise means (24) for adjusting the received sequences of digital signal samples to a zero average value. Device according to claim 2, characterized in that the second processing means (16) comprise an attenuation unit in the form of an analog AGC unit (163), which attenuation unit is controlled in response to power estimates for the digital signal samples. Communication system comprising a number of transmitter and receiver units, characterized in that at least one receiver unit comprises a device according to one of claims 1 to 5. Communication system according to claim 8, characterized in that said at least one receiver unit is part of a radio base station. Communication system according to claim 8 or 9, characterized in that the system is a CDMA-based communication system. Communication system according to claim 10, wherein each base station is equipped with a mandatory function block for power estimation of the received uplink signals, characterized in that the mandatory function block for power estimation is reused for extracting the signal power estimates for the digital AGC. the device. A method for receiving signals from a transmitter in a communication system in a receiver unit of a communication system through at least one signal input, comprising the steps, for each of the signal inputs, to process received signals at least by converting the signals into baseband signals and converting these signals into a sequence of digital signal samples, characterized in that for at least one of said signal inputs converting the digital signal samples into signals at a common and constant power level, resulting in a reduced dynamic range, to quantize the converted signal samples into signals with a reduced number of bits and to feed the converted and quantized signal samples to a demodulator. Method according to claim 12, characterized in that the conversion of a digital signal sample takes place by determining a grading factor from power estimates of the signal samples, quantizing the grading factor with respect to distinct intervals for the power estimates and grading by multiplying the signal sample by the quantized grading. - the factor. Method according to claim 12, characterized in that the conversion of a digital signal sample takes place by determining a grading factor from power estimates of the signal sample, quantizing the grading factor with respect to distinct intervals of the power estimates to a value which can be represented as a sum of powers of two, determination of shift values representing the exponents of the sum meters and 10 20 516 3 55 gzt fi ïf, 16 grading by left-shifting the signal sample according to the shift values and summing the shifted parts. Method according to Claim 13 or 14, characterized by determining one grading factor per signal sample. Method according to one of Claims 12 to 15, characterized by feeding side information regarding the quality of the gradation directly to the demodulator (11). Method according to one of Claims 12 to 16, characterized in that the quantized grading factors are assigned with an overlap for increasing and decreasing values of the power estimates at the boundaries between the distinct power ranges. Method according to one of Claims 12 to 17, characterized by adjusting received sequences of digital signal samples to a zero average value by subtracting a calculated average estimate value from the signal samples. Method according to one of Claims 12 to 18, characterized by attenuation of the received signals in the analog part of the receiver in response to power estimates of signal samples of the received signals in the digital part.