US 20080102873 A1 Abstract A power controller in a communication system adjusts the transmitted power level using adaptive prediction of channel fading for maintaining the power level for improved communications by using an adaptive nonlinear predictor that explicitly takes into account the propagation delay and channel fading where the transmitted power is determined on the basis of the measurement of the received power levels over the individual diversity branches for providing communication through a fading channel with effective performance equivalent to that of a channel without any signal fading.
Claims(14) 1. A system for maintaining constant power levels of received uplink signals communicated through an uplink channel having a time variable channel gain, the system comprising,
an uplink transmitter for transmitting uplink signals through the uplink channel, the uplink signals having an adjusted power level adjusted by an adaptive power reference, an uplink receiver for receiving the uplink signals as the received uplink signals and for generating power measurement signals, a downlink transmitter for transmitting downlink signals through a downlink channel, the downlink signals containing the power measurement signals, a downlink receiver for receiving the downlink signals and for generating received power measurement signals, and a power controller for receiving a power reference and the received power measurement signals and for generating the adaptive power reference through a prediction of the time variable channel gain at a future time defined by a round trip time delay between transmitting the uplink signals and receiving the downlink signals. 2. The system of the uplink channel and downlink channel are the downlink channel and the uplink channel, respectively. 3. The system of the round trip delay time is predetermined. 4. The system of 5. The system of a controller delay for delaying the adaptive power level reference by the round trip time delay for generating a delayed adaptive power reference signal, an adder for subtracting the delayed adaptive power reference from the received power measurement signals for generating adaptive power measurement signals, a predictor for receiving the adaptive power measurement signals and generating a predicted adaptive power reference signal, and an adder for subtracting the predicted adaptive power reference signal from the power reference for generating the adaptive power reference. 6. The system of a first delay line for receiving the adaptive power measurement signals and generating first delay line adaptive power measurement signals, a predictor delay for delaying by the round trip time delay the adaptive power measurement signals into delayed adaptive power measurement signals, a second delay line for receiving the delayed adaptive power measurement signals and generating second delay line adaptive power measurement signals, a calculator for receiving the adaptive power measurement signals and second delayed adaptive power measurement signals for generating parameters modeling gain variations of the time varying channel gain, and a nonlinear predictor for receiving the parameters and the first delay line adaptive power measurement signals for generating the predicted adaptive power measurement signals. 7. The system of a smoothing filter for filtering the adaptive power measurement signals into filtered adaptive power measurement signals, a first delay line for receiving the filtered adaptive power measurement signals and generating first delay line adaptive power measurement signals, a predictor delay for delaying by the round trip time delay the filtered adaptive power measurement signals into delayed adaptive power measurement signals, a second delay line for receiving the delayed adaptive power measurement signals and generating second delay line adaptive power measurement signals, a calculator for receiving the adaptive power measurement signals and the filtered adaptive power measurement signals and the second delayed adaptive power measurement signals for generating parameters modeling gain variations of the time varying channel gain, and a nonlinear predictor for receiving the parameters and the first delay line adaptive power measurement signals for generating the predicted adaptive power measurement signals. 8. The smoothing filter of the smoothing filter is a first order recursive filter. 9. The predictor of the calculator applies an exponentially data weighted recursive least square algorithm upon a state vector comprising the exponential powers of the second delay line adaptive power measurement signals and the filtered adaptive power measurement signals and the adaptive power measurement signals for generating the parameters for defining in a logarithmic scale the time varying channel gain in logarithmic scale, the parameters being recursively updated over time. 10. The nonlinear predictor of the calculator applies an exponentially data weighted recursive least square algorithm upon a state vector comprising the exponential powers of the second delay line adaptive power measurement signals and the filtered adaptive power measurement signals and the adaptive power measurement signals for generating the parameters for defining in a logarithmic scale the time varying channel gain in logarithmic scale, the parameters being recursively updated over time, and the predictor predicts the predicted adaptive power reference level by computing an inner product of a parameter vector defined by the parameters and a state vector comprising exponential powers of the first delay line adaptive power measurement signals. 11. The system of a controller delay for delaying the adaptive power level reference by the round trip time delay for generating a delayed adaptive power reference signal, adders for respectively subtracting the delayed adaptive power reference from the received power measurement signals for generating respective adaptive power measurement signals, a bank of predictors for respectively receiving the adaptive power measurement signals and generating respective predicted adaptive power reference signals, a selector for selecting one of the predicted adaptive power reference signals as a selected predicted adaptive power reference signal, and an adder for subtracting the selected predicted adaptive power reference signal from the power reference for generating the adaptive power reference. 12. The system of the selected predicted adaptive power reference signal is a one of the predicted adaptive power measurement signals, the selection a selection function selected from the group consisting of maximum, minimum, average, and weighted average. 13. The bank of predictors of a first delay line for receiving the adaptive power measurement signals and generating first delay line adaptive power measurement signals, a predictor delay for delaying by the round trip time delay the adaptive power measurement signals into delayed adaptive power measurement signals, a second delay line for receiving the delayed adaptive power measurement signals and generating second delay line adaptive power measurement signals, a calculator for receiving the adaptive power measurement signals and second delayed adaptive power measurement signals for generating parameters modeling gain variations of the time varying channel gain, and a nonlinear predictor for receiving the parameters and the first delay line adaptive power measurement signals for generating the predicted adaptive power measurement signals. 14. The bank of predictors of a smoothing filter for filtering the adaptive power measurement signals into filtered adaptive power measurement signals, a first delay line for receiving the filtered adaptive power measurement signals and generating first delay line adaptive power measurement signals, a predictor delay for delaying by the round trip time delay the filtered adaptive power measurement signals into delayed adaptive power measurement signals, a calculator for receiving the adaptive power measurement signals and the filtered adaptive power measurement signals and the second delayed adaptive power measurement signals for generating parameters modeling gain variations of the time varying channel gain, and Description The invention was made with Government support under contract No. F04701-00-C-0009 by the Department of the Air Force. The Government has certain rights in the invention. The invention relates to the field of communication systems. More particularly, the present invention relates to power control of communication signals for improved reception in fading channel communication systems. Existing closed loop power control algorithms in the Code Division Multiple Access (CDMA) and Universal Mobile Telephone System (UMTS) ignore the roundtrip delay incurred in the measurement of the received signal power. Such an assumption of zero round trip delay may be valid when the round trip delay is negligible in comparison to the inverse of the channel fading bandwidth. However, in the case of satellite channels such a condition is not satisfied. The round trip delay for geostationary satellites, for example, may be in the range of 250 msec to 500 msec, including any terrestrial delay, for a single hop or two-hop system. When compared to the inverse of the channel fading delay in the range of 1.0 to 10.0 sec, corresponding to the bandwidth in the range of 0.1 to 1.0 Hz, such a delay is not negligible. Hence, any power control algorithm based on such a zero delay assumption will not have satisfactory performance. Power control is necessary for frequency division multiple access (FDMA) and time division multiple access (TDMA) systems to control adjacent channel interference and mitigate unexpected interference caused by a near-far problem. In these systems, power control is needed to improve the quality of voice delivered to a user. However, in CDMA systems that are self-interfering, the capacity of the system depends directly on the power control and an accurate power control mechanism is needed for maximizing the number of users that can operate simultaneously in the system. Improved power control can also save the amount of transmitted power of the mobile station (MS), which increases the life of the battery. In case of fading channels, diversity-combining techniques are generally employed to improve system performance. For example, in case of the CDMA systems, implicit diversity gain is exploited with the help of a rake receiver. In such a diversity system employing either an explicit or an implicit diversity, the power control in the terrestrial systems is based on the estimated power level at the diversity combiner output. In CDMA systems the power control on the reverse link, for example, is based on a combination of open loop control along with a closed loop correction implemented in the base station and the MS. For open loop probing on the access channel with closed loop correction inactive, the mobile station transmits the first probe signal at an output power level, defined by a mean output power in dBm is equal to k minus the mean receive input power in dBm plus 0.5, times, the nominal power in dB plus 0.5, times, the initial power in dB. The mean power is referred to a nominal CDMA channel bandwidth of 1.23 MHz and k, nominal power, and initial power are system parameters. The initial power parameter is any margin, positive or negative, above the required level. When the initial power is zero, then the nominal power is the correction to provide the correct received power at the base station. Essentially the power control, in both open loop and closed loop, is based on the assumption of identical channel gain on both the reverse and forward link, thus ignoring both the roundtrip propagation delay and the difference in the transmission frequency on the two links. Subsequent probes in an access probe sequence are sent at an increasing power level until a response is obtained from the base station. The initial transmitted power in the reverse traffic channel is at the power level given by the mean TX output power equation plus all the access probe corrections. Uncontrolled differences in the forward and reverse channels, such as opposite fading that may result due to the frequency difference plus mismatches in the mobile station's transmit and receive chains result in the transmit power, which is different than the desired level. To reduce these residual errors, a close loop control is used wherein the mobile station measures the received E The closed loop correction in the CDMA systems ignores the round trip delay in comparison to the inverse of the channel fading bandwidth. For example, typically the fading bandwidth for the terrestrial cellular communication networks may be of the order of 50.0 Hz having a time constant of 20.0 msec compared to a round trip delay smaller than 0.1 msec for a cell radius of 10 Km and thus such assumption is satisfied. However, in the case of satellite channels with a roundtrip delay of 500 to 600 msec such an assumption is not valid even when the fading bandwidth is less than 1 Hz. In the UMTS system, two loops for power control are involved. The inner loop is based on a bandwidth of about 1500 Hz with a period of 0.66 msec, measures the received signal to interference ratio and compares to the desired signal to interference ratio. The loop period is much higher compared to the roundtrip delay and thus the roundtrip delay is not too important in the loop design. This loop is similar to the closed loop in the CDMA system. An outer loop measures a service metric such as the frame error rate and adjusts the desire signal to interference ratio to account for any unmodeled uncertainties. The outer loop bandwidth is in the range of 10.0 to 100.0 Hz. Similar to the CDMA power control, a satisfactory operation of the UMTS system power control requires that the roundtrip delay be negligible compared to the channel fading bandwidth and the loop bandwidth. However, in the case of satellite channels such a condition is not satisfied. In case of the satellite channels, the round trip delay may vary between 250 to 550 msec, depending upon whether a single hop or two hop system is used, compared to 1.0 μsec to 100 μsec variation in UMTS system. When compared to a fading channel bandwidth of 0.1 to 1.0 Hz with a time constant of 1.0 to 10.0 seconds, the roundtrip delay is not negligible and hence the power control approach will not work satisfactorily. It is therefore necessary to take into account the roundtrip propagation delay explicitly in the design of the power control algorithm. A satisfactory operation of both the CDMA and the UMTS power control algorithm requires that the roundtrip delay be negligible compared to the channel fading bandwidth and the loop bandwidth. When compared to a fading channel bandwidth and time constant, the roundtrip delay is not negligible and hence the prior power control algorithms may not work satisfactorily. These and other disadvantages are solved or reduced using the invention. An object of the invention is to provide a system for improving the reception of the communication signals. Another object of the invention is to provide a system for improving the reception of the communication signals by adjusting the power level of a transmitted signal. Yet another object of the invention is to provide a system for improving the reception of the communication signals by adjusting the power level of a transmitted signal communicated through a fading channel. Still another object of the invention is to provide a system for improving the reception of the communication signals by adjusting the power level of a transmitted signal by predicting the amount of fading in a fading channel. The present invention is directed toward a power control system in a communication system for adjusting the power level of a transmitted signal through prediction of the amount of the amount of fading in a fading channel. A power control algorithm explicitly takes into account the propagation delay by including an adaptive N-step ahead nonlinear predictor. The fading channel power gain in dB is modeled in terms of a parameterized nonlinear model with memory. The model parameters are adaptively estimated by a recursive least squares algorithm from which a power controller derives the required prediction of the channel power gain and the requisite transmitted power level. The system can be used for improved reception through satellite communication channels where the roundtrip propagation delay are incurred when obtaining the measurements of the received power level at the transmitter site for the purpose of controlling the transmitted power. Such a roundtrip delay is significant when compared to the inverse of the channel fading bandwidth. Thus, the power control algorithms take into account such a delay for use with communication systems where the roundtrip delay is relatively significant. The power control algorithm is based on an adaptive nonlinear predictor that explicitly takes into account a long propagation delay. The transmitted power is determined on the basis of the measurements of the received power levels over the individual diversity branches for improving performance. Using a predictive and adaptive algorithm, the overall performance is shown to be very close to the performance of the ideal additive white Gaussian noise (AWGN) channel without any fading. For example, when the received power from all diversity branches is combined, as in case of equal gain diversity combining, and compared to the power level required for the nonfading channel to achieve the same probability bit error, the residual loss due to fading is about one dB for a probability of bit error equal to 10 An embodiment of the invention is described with reference to the figures using reference designations as shown in the figures. Referring to When the channel The downlink signals communicate composite information including the payload data and the power measurement signals ψ The power controller Referring to The transmitted power is determined by the power controller The system reduces the penalty due to fading to a negligible value, especially when some form of implicit or explicit diversity is available. Such a diversity may involve just a second channel whose average power may be only a fraction of the main channel average power and is for the purpose of removing any outliers left by the adaptive algorithm. For the CDMA systems such an implicit diversity is mostly available as the coherence bandwidth is generally smaller than the wideband CDMA signal bandwidth. Even in the absence of any diversity, the algorithm makes it possible to achieve a probability of bit error (P Referring again to The noisy estimate of the received power level at the receiver at the sampling time instance (k+N In the adaptive predictor, the term {circumflex over (z)}
_{m }is minimized with respect to θ where
where λ is the exponential data weighting factor with 0<λ<1. The recursive solution of the optimization problem known as the exponentially data weighted recursive least squares algorithm is given by {circumflex over (θ)}(k)={circumflex over (θ)}(k−1)+[P(k−1)x(k)/μ(k)] [z(k)−{circumflex over (θ)} The measurement of the received signal power level is at any discrete time instance k. The receiver has a time constant equal to the sample period which may be 10.0 msec or higher in the case of the geostationary satellite channels for providing an accurate estimate of the average received power level during that period. Thus, in the measurement z For the performance evaluation, channel gain variations are modeled by a random process obtained by a complex white noise process filtered by a Butterworth filter of some specified bandwidth and order. Thus, the real envelope of the channel gain a From the simulations the time average of a The power control algorithm can be combined with diversity combining techniques with L diversity branches to further improve the power control performance. The diversity techniques may be classified into three main classes. In the explicit transmit-receive diversity, the transmit power in each of the L diversity channels can be individually controlled. Examples of such techniques are the frequency diversity and polarization diversity techniques. The second class of techniques is the receive only diversity wherein there is only one transmitter but L receivers. In this second class, the power in different diversity channels cannot be individually controlled, however each diversity signal is individually available from which the received power level over different diversity paths can be ascertained. An example of this second class of techniques is L receive antennas that are physically separated by a predetermined distance. The third class of the diversity techniques is the class of implicit diversity techniques. An implicit diversity is a special case of the receive only diversity case with the difference that no explicit resources, such as multiple receive antennas, are required. The diversity is achieved by separating the single received signal on the basis of signal parameters. For example, in the rake receiver applicable for the spread spectrum signals, the signal is separated on the basis of propagation delays encountered, such as when the signal propagates over multiple paths, which have significant difference in respective propagation delays. The power control techniques can be combined with various diversity methods in order to achieve high performance. In the case of explicit transmit-receive diversity, the power control algorithm operates independently on the various diversity channels. The signals from various diversity channels under power control are then combined at the receiver using any one the traditional combining techniques such as switched, equal gain combining, or the maximal ratio combining technique. For the explicit or implicit receive only diversity, the receiver measures the power level in each of the diversity channels and transmits power information to the transmitter. The transmitter then applies the adaptive prediction algorithm to each of the channels individually and computes the required control input for each of the diversity channels. The actual control input, that is, the transmit power, is then determined from the computed controls according to predetermined criteria, such as, u Both the transmit and receive explicit or implicit diversity techniques can be used, such as the dual implicit diversity. In the example of dual diversity, the two channel gain variations are independent random processes and have the same statistics. The control is based on the minimum of the transmit power levels required individually for the two implicit channels so that at least one of the diversity channels has the predicted received power level equal to the desired level u For the case of an explicit dual diversity with u For an implicit triple switched diversity with power control based on the minimum channel gain, the rms power level variation is equal to 0.73 dB with mean power level equal to −14.9 dBW. The mean transmitted power is equal to 4.14 dBW and is smaller than the difference between the mean channel power gain of −23.36 dB and the desired received power level of −15 dBW, which is equal to 8.36 dBW. The difference is the result of the diversity gain and the power control algorithm. In all of these examples, both the parameters N Effect of the measurement noise on power control can be determined. There are several possible algorithms for the measurement of the received power level. In one such method, the inphase and quadrature components of the complex envelope of the received signal, after removing the data modulation in a decision directed manner, are averaged over the period T The performance of the power control algorithm is a function of the delay parameter N With the same average received power level in both the cases of nonfading and fading channels, wherein for the case of fading diversity channels the received power is measured at the combiner output, the performance of the power control system for L≧2 is nearly the performance of the nonfading channel. However, due to fading, the power control, and different combining techniques, the average transmitted power level may be different even for the same received power level in the two cases. A power control and diversity power table lists various power levels for the case of L=1 and L=2, and L=3.
For the case of AWGN channel, the transmit power is 5.65 dBW, which is equal to the required received power of −15 dBW minus the average of the channel power gain equal to −20.65 dB. In comparing the performance in terms of the transmitted power level, the average transmitted power level for the power control system is selected to be the same for the fading channel with and without any power control and for the AWGN channel. Thus, in making the transmit power the same as for the AWGN channel, the transmit power is reduced below those given by Table 1 and brought to the same 5.65 dBW level in all cases. With implicit diversity, the average transmit power is measured in a single channel as there is only one transmitter for the case of implicit diversity, and is set equal to the transmit power required for the ideal AWGN channel. The diversity combining technique is selected to be the optimum maximal ratio combining for the case of no power control whereas switched or equal gain non-optimum combining is used for the case of power control in this example. The comparison shows that for the case of L=3, the residual degradation is negligible when using power control. However, when no power control is used, about 7.0 dB degradation results at a probability error of 10 When the maximal ratio combining is used along with power control, the performance of the diversity technique with power control is better than the AWGN of a nonfading channel by about 1.7 dB for a probability error of 10 Unequal power in the diversity channels provides other examples of the power control. When different diversity channels have unequal channel gains, the average relative gains for the three diversity channels are selected as [0, −3, −5] dB in one of the examples. Even in this case, the performance of the diversity system with power control stays close to that of an ideal channel for a probability error equal to 10 or worse. In the absence of power control, the performance is worse by about 7 dB for a probability error equal to 10
The power control algorithm can be evaluated after averaging the performance over the complete duration of fading. However, under fading situations, it may be adequate to achieve a specified probability error only over a certain specified percentage of time. Such percentage of time is termed as the availability and may be specified to be between 99% and 99.9%, that is, less than 100% availability. With P In the case of CDMA systems, the performance is determined primarily by the self noise of the system and a few dB increase in the transmitted power level may be insignificant. In this case, the results should be compared on the basis of the total received power level rather than in terms of the transmitted power. In that case, there is no significant performance degradation in comparison to AWGN channel when L≧2. Also for the case of L=1, the degradation is smaller than a dB for P The power control system of the invention can be applied to all fading communication channels involving significant propagation delays, such as satellite channels. The power control system also provides improvement to the performance of the terrestrial communication systems such as CDMA cellular networks by taking into account the propagation delay and other nonideal effects in the adaptive algorithm. Besides communication systems, power control can be used in other systems, such as, command and tracking systems. The power control determines the transmitted power on the basis of a measurement of the received power levels over the individual diversity branches. Using power control, the overall performance is very close to the performance of a system without any fading. For example, when dual or triple diversity is used with equal power over all the diversity branches, the performance is within a fraction of a dB of the performance of a nonfading channel for a P In TDMA and FDMA systems, which are not self-interference limited, performance improvements may be on the basis of the total power transmitted. Such a comparison naturally includes the loss in the fading channel, where higher transmission power is necessary to counter higher losses in the channel in order to receive the same average receive power in both the fading and nonfading channels. In this comparison for the case of equal average power over all diversity channels, the average transmitted power over any of the diversity channels is set equal to the power transmitted over the nonfading channel and then the P With power control and explicit or implicit diversity, the residual loss due to fading can be made negligible. The residual loss occurs due to the outliers present in the probability density function of the received power level. The received power level averaged over a sampling period T For the TDMA and FDMA systems, wherein the comparison is performed in terms of the transmitted power rather than the received power, triple diversity results in the same or better performance than the AWGN channel case for a P Referenced by
Classifications
Legal Events
Rotate |