Method and apparatus for rate control in a high data rate communication systemBackground
Field of the invention
The present invention relates to communications. More particularly, the present invention relates to a novel method and apparatus for adaptive rate selection in a wireless communication system.
Second, description of related Art
Contemporary communication systems are required to support a variety of applications. One such communication System IS a Code Division Multiple Access (CDMA) System that conforms to the "TIA/EIA/IS-95 Mobile Station-Base Station compatibility Standard for Dual-Mode Wide-Band Spread Spectrum Cellular System" referred to herein as the IS-95 standard. The CDMA system supports voice and data communications over terrestrial links between users. The use of CDMA technology IN MULTIPLE ACCESS COMMUNICATION SYSTEMs is explained IN U.S. Pat. No. 4,901,307, entitled "SPREAD SPECTRUM MULTIPLE ACCESS COMMUNICATION SYSTEM SATELLITE OR TERRESTRIAL REPEATERS," and U.S. Pat. No. 5,103,459, entitled "SYSTEM AND METHOD FOR GENERATING WAVEFORMS IN A CDMA CELLULARTELEPHENE SYSTEM," both assigned to the assignee of the present invention and incorporated herein by reference.
In a CDMA communication system, communication between users is performed through one or more base stations. In a wireless communication system, the forward link refers to the channel through which a signal propagates from a base station to a subscriber station, and the reverse link refers to the channel through which a signal propagates from a subscriber station to a base station. By sending data on the reverse link to the base station, a 1 st user on one subscriber station may communicate with a 2 nd user on a 2 nd subscriber station. The base station receives data from the 1 st subscriber station and transmits the data to the base station serving the 2 nd subscriber station. Both may be served by a single base station or multiple base stations, depending on the location of the subscriber station. In either case, the base station serving the 2 nd subscriber station transmits data on the forward link. Subscriber stations may also communicate with wireline telephones through a Public Switched Telephone Network (PSTN) coupled to the base station, or with the terrestrial internet through a connection with a serving base station, rather than with the 2 nd subscriber station.
Given the growing demand for wireless data applications, the need for very efficient wireless data communication systems has become increasingly important. The IS-95 standard specifies the transmission of traffic data and voice data on the forward and reverse links. In U.S. patent No. 5,504,773, entitled "METHOD AND APPARATUS FOR THE formatting OF DATA FOR TRANSMISSION," assigned to THE assignee OF THE present invention AND incorporated herein by reference, a METHOD OF transmitting traffic DATA in fixed-size code channel frames is described. According to the IS-95 standard, the traffic or voice data IS divided into 20 millisecond wide coded channel frames with data rates up to 14.4 Kbps.
In mobile radio communication systems, there is a significant difference between the requirements for providing voice and data services, i.e. non-voice services such as internet or facsimile transmission. Unlike data services, voice services require strict and fixed delays between speech frames. Typically, the total one-way delay for a speech frame to transmit voice information must be less than 100 milliseconds. In contrast, the transmission delay that occurs during a data (i.e., non-voice information) service may vary and may utilize a delay that is longer than the delay that a voice service can tolerate.
Another significant difference between voice and data services is that voice services require a fixed and common level of service compared to data services. Typically, for digital systems providing voice services, this requirement is met by using a fixed and equal transmission rate for all users and a maximum tolerable error rate for the speech frames. For data services, the level of service may vary from user to user.
Yet another difference between voice services and data services is that voice services require a reliable communication link that is provided using soft handoff in a CDMA communication system. Soft handoff requires redundant transmission of the same voice information from two or more base stations to improve reliability. A method of soft handoff is disclosed IN U.S. patent No. 5,101,501, entitled "soft handoff IN a CDMA cell TELEPHONE SYSTEM". This additional reliability is not required to support the data service because the erroneously received data packets may be retransmitted.
As a mobile station moves through a mobile radio communication system, the quality of the forward link (and the capacity of the forward link upon which data is communicated) will vary. Thus, at some times a given forward link between a base station and a mobile station will be able to support very high data transmissions, at other times the same forward link may only be able to support a substantially reduced data transmission rate. To maximize the throughput of information on the forward link, it is desirable to vary the data transmission on the reverse link so that the data rate is increased during those time intervals in which the forward link can support a higher transmission rate.
When non-voice data is transmitted from a base station to a mobile station on the forward link, it is necessary to transmit control information from the mobile station to the base station. Sometimes, however, even though the forward link signal may be strong, the reverse link signal may be weak, resulting in a situation where the base station is unable to receive the control signals from the mobile station. In such a case, where the forward link and the reverse link are unbalanced, it is undesirable to increase the transmit power on the reverse link to improve the quality of reception of control information at the base station. For example, in a CDMA system, increasing transmit power on the reverse link is undesirable because such a power increase may adversely affect the reverse link capacity seen by other mobile stations in the system. It is desirable to have a data transmission system in which the forward link and reverse link associated with each mobile station are maintained in a balanced state without adversely affecting the reverse link capacity. It is further desirable that such a system maximize the throughput of non-voice data on individual forward links when such links are strong enough to support higher data rates.
One way of the above-mentioned requirements in High Data Rate (HDR) systems is to keep the transmit power fixed and vary the data rate depending on the channel conditions of the users. Thus, in current generation HDR systems, an Access Point (AP) always transmits AT maximum power to only one Access Terminal (AT) in each time slot, and the AP uses rate control to adjust the maximum rate that the AT can reliably receive. An AP is a terminal that allows high data rate transmission to an AT.
As used in this document, a time slot is a time interval of finite length, such as 1.66 milliseconds. A slot may contain one or more packets. A packet is a structure that includes a header, a payload, and a quality metric, such as a Cyclic Redundancy Check (CRC). The header is used by the AT to determine whether a packet is intended for the AT.
An exemplary HDR system defines a set of data rates ranging from 38.4kbps to 2.4Mbps AT which an AP can transmit data packets to an AT. The data rate is selected to maintain a target Packet Error Rate (PER). The AT measures the received signal to interference and noise ratio (SINR) AT regular intervals and uses this information to predict the average SINR over the next packet duration. An exemplary prediction METHOD is explained in pending application serial No. 09/394,980 entitled "SYSTEM AND METHOD for computing algorithm PREDICTING SIGNAL TO INTERFERENCE AND NOISE algorithm TO implementation prediction algorithm" assigned TO the assignee of the present invention and incorporated herein by reference.
Fig. 1 shows a conventional open-loop rate control apparatus 100. The past stream of SINR values at the instances of n-m …, n-1, n are provided to predictor 102, where each SINR value is measured over the duration of the corresponding packet. The predictor 102 predicts the average SINR over the next packet duration according to the following equation:
OL_SINRPredicted＝OL_SINREstimated-K·σc (1)
in equation (1), OL _ SINRPredicedIs the predicted SINR, OL _ SINR, for the next packet from the open loopEstimatedIs the SINR estimated by the open loop from past SINR values, K is a compensation factor, and σcIs the standard deviation of the error metric.
For example, the estimated SINR may be obtained by selecting the output from a set of low pass filters that act on measurements of past SINRs. Selecting a particular filter from the filter bank may be based on an error metric defined as a difference between the particular filter output and a measured SINR immediately after the output for a packet duration. By a step-back from the filter output corresponding to the standard deviation σ of the step-back factor K from the error metriccThe predicted SINR is obtained. The value of the back-off factor K is determined by a back-off control loop which ensures that the tail probability, i.e. the probability that the predicted SINR exceeds the measured SINR, is achieved with a certain percentage of time.
The SINRPredictedThe values are provided to a look-up table 104 that maintains a set of SINR thresholds representing the minimum SINR required to successfully decode the packet at each data rate. The AT (not shown) uses the look-up table 104 to select the highest data rate AT which the SINR threshold is below the predicted SINR and requests the AP (not shown) to send the next packet AT this data rate.
The above method is an example of an open loop rate control method, and the optimal rate at which to receive the next packet is determined based only on a measurement of the channel SINR without any information about the decoder error rate (for each data rate packet) at a given SINR under prevailing channel conditions. Any open-loop rate control algorithm suffers from several drawbacks, some of which are discussed below. First, a certain tail probability, such as 2%, does not mean a PER of 2%. This is because PER is a monotonically decreasing function of SINR, with a finite slope that depends on the coding scheme and channel conditions. Equation (1), however, assumes a "brick wall" PER characteristic, i.e., packets are guaranteed to be decoded as long as the SINR exceeds a threshold for the corresponding rate, and packets are erroneous as long as the SINR falls below the threshold. In addition, the open-loop rate control method uses a fixed set of SINR thresholds, which ensures that under worst-case channel conditions, the packet error rate is close to the target error rate. However, the performance of the decoder depends not only on the SINR but also on the channel conditions. In other words, a method that uses a fixed set of SINR thresholds for all channels achieves different packet error rates on different channels. Thus, while the open-loop method works optimally under the worst-case channel conditions, under typical channel conditions the method results in a much lower error rate than necessary at the expense of reduced throughput. Furthermore, practical rate control methods require a small, limited set of data rates. The rate selection method always selects the closest lower data rate to ensure an acceptable PER. Thus, rate quantization results in a loss of system throughput.
Therefore, there is a need to address the deficiencies of existing approaches.
Summary of The Invention
The present invention is directed to novel methods and apparatus for adaptive rate selection in a wireless communication system. Thus, in one aspect of the invention, the SINR predicted by the open-loop approach is modified with closed-loop correction. The closed loop correction is updated based on the packet error event and the target error rate.
In another aspect of the invention, the closed loop correction is advantageously updated according to the frequency of received packets.
Brief Description of Drawings
The features, objects, and advantages of the present invention will become more apparent from the detailed description set forth below when taken in conjunction with the drawings in which like reference characters identify correspondingly throughout and wherein:
fig. 1 illustrates a block diagram of a conventional open loop rate control apparatus.
Fig. 2 illustrates a block diagram of an apparatus for a rate control method according to one embodiment of the present invention.
FIG. 3 illustrates a flow chart of an exemplary method of updating an outer loop correction.
Detailed description of the preferred embodiments
Fig. 2 illustrates an exemplary communication system 200 capable of implementing embodiments of the present invention. AP 204 transmits signals to AT 202 over forward link 206a and receives signals from AT 202 over reverse link 206 b. The communication system 200 may be operated bi-directionally, with each of the terminals 202, 204 operating as either a transmitter unit or a receiver unit, or both, depending on whether data is being transmitted or received at the respective terminal 202, 204. In a cellular wireless communication system embodiment, transmitting terminal 204 may be a Base Station (BS), receiving terminal 202 may be a Mobile Station (MS), and forward link 206a and reverse link 206b may be of the electromagnetic spectrum.
The AT 202 includes means for a rate control method in accordance with one embodiment of the present invention. The device comprises two control loops, one open loop and one closed loop.
The open loop includes an SINR predictor 208 and a look-up table 210 that controls the forward link data rate based on the difference between the average SINR of the next packet and the SINR threshold for the overall data rate. Forward the direction ofSignals arriving AT the AT 202 in packets from the AP 204 on link 206a are provided to a decoder 212. The decoder 212 measures the average SINR over the duration of each packet and provides the SINR to the SINR predictor 208. In one embodiment, the predictor 208 predicts the SINR of the next packet (OL _ SINR) according to equation (1)Prediction) The value is obtained. However, one of ordinary skill in the art will appreciate that any open-loop method may be used and is not limited to the one expressed by equation (1). Setting the OL _ SINRPredictionThe values are provided to the look-up table 210. The look-up table 210 maintains a set of SINR thresholds representing the minimum SINR required to successfully decode a packet at each data rate. The set of SINR thresholds is adjusted by operation of the closed loop.
In block 214, the closed loop utilizes the PER information provided by the decoder 212 to determine a closed loop correction value L. The closed loop correction value L adjusts the set of SINR thresholds in the look-up table 204 according to the following equation:
CL_SINRprediction＝OL_SINRPrediction+L (2)
In equation (2), L represents the closed loop correction to the open loop prediction of SINR over the next packet duration. Adding L to the SINR predicted by the open-loop algorithm in equation (1) corresponds to subtracting L from the SINR threshold for rate control. Since the correction term L is updated based on PER information, which reflects prevailing channel conditions, the set of SINR thresholds is preferably matched to the prevailing channel conditions.
The AT 202 uses the adjusted set of SINR thresholds in the look-up table 210 to select the highest data rate having an SINR threshold that is lower than the predicted SINR. The AT 202 then requests the AP 204 to send the next packet AT this data rate on the reverse link 206 b.
Although the predictor 208, the decoder 212, and the closed-loop correction block 214 are shown as separate elements, those of ordinary skill in the art will appreciate that the physical differences are for illustration purposes only. The predictor 208, the decoder 212, and the closed-loop correction block 214 may be integrated into a single processor to implement the processing described above. Thus, for example, the processor may be a general purpose processor, a digital signal processor, a programmable logic array, or the like. In addition, the lookup table 210 is a space in memory. The memory may be part of the processor or processors described above, or may be a separate element. The implementation of the memory is a design choice. Thus, the memory may be any medium capable of storing information, such as a magnetic disk, a semiconductor integrated circuit, or the like.
Fig. 3 illustrates a flow chart of an exemplary method of updating L to ensure the best possible throughput is achieved with an acceptable error rate.
In step 300, a normalized Activity Factor (AF) variable is initialized to a 0 or 1 value by an AT (not shown). The AF quantifies the fraction of time the AT receives packets on the forward link. An AF equal to 1 means that the AT 202 receives packets most of the time, while an AF equal to 0 means that the forward link to a given AT is mostly idle. In one embodiment, the AF is initialized AT the time the AT initializes a new communication. In this case, it is advantageous to initialize the AF to 1 because the AT is receiving packets. The AF is updated at the end of each slot according to the following equation:
AFnew＝(1-f)·AFOld age+f， (3)
Or
AFNew＝(1-f)·AFOld age， (4)
Wherein:
f ∈ (0, 1) is a parameter that controls the rate of change of the AF. In one embodiment of the present invention, f is set to 1/50.
Equation (3) is used when the AT finds a packet preamble AT the beginning of a slot, or it is still demodulating a packet whose preamble was detected in the previous slot. This occurs when the AT sends a request for data and the AP (not shown) sends the requested data. Equation (4) is used when the AT is not in the middle of packet demodulation, searching for a packet preamble but fails to find a preamble. This occurs when the AT sends a request for data and the AP fails to receive or ignore the request for data and decides to serve other ATs in the system.
In step 300, the outer loop correction variable L is also initialized by the AT. L may be initialized to LMinimum sizeAnd LMaximum ofAny value in between. L isMinimum size、LMaximum ofAny value can be reached. Example values are referenced below. In one embodiment, L is initialized to 0 dB.
In step 300, the operating mode is also initialized. There are two modes: a normal mode and a fast start mode. The motivation for defining the two modes for the rate control algorithm is based on the knowledge that the optimal step size for the up and down L correction depends on the target PER, the packet arrival process, and the preamble false alarm statistics. While the preamble false alarm statistics are relatively constant and associated with the outer loop term L, the packet arrival process is time-varying and a priori unknown AT the AT. As discussed above, data traffic is prone to burstiness due to rare packet arrivals in the qualified idle state, while data traffic is busy due to frequent packet arrivals. Therefore, the normal mode is used during a steady state. A fast start mode designed to recover quickly from long-term inactivity is used when preamble false alarms tend to drive the rate control algorithm towards the holding mode.
The rules that determine the mode of the algorithm and the rules that update L are based on the detection of good or bad packets. An access terminal is said to receive a good packet if it detects the packet preamble, demodulates and decodes the packet, and recovers a valid CRC. An access terminal is said to receive a bad packet if it detects the preamble but obtains an invalid CRC as soon as it demodulates and decodes the packet.
A transition to the fast start mode occurs if the following conditions are met and the two most recently received packets are good:
L＜LAM threshold， (5)
AF＜AFFree up， (6)
In equations (5) - (6), LAM thresholdIs a threshold for the transition of control of L to the fast start mode. In one embodiment of the invention, L isAM thresholdThe threshold is set to 0 dB. AFFree upIs a threshold value for the transition of the control of AF to the quick start mode. In one embodiment of the invention, AF is appliedFree upThe threshold is set to 10%.
A transition to the normal mode occurs if either of the following conditions is met or the most recently received packet is bad:
L≥LNM threshold， (7)
AF≥AFBusy state， (8)
In equations (7) - (8), LNM thresholdIs a threshold value for the transition of the control of L to the normal mode. In one embodiment of the invention, L isNM thresholdSet to 2 dB. AFBusy stateIs a threshold value for the transition of control of a to the normal mode. In one embodiment of the invention, AF is appliedBusy stateThe threshold is set to 25%.
Upon completion of initialization, the AT waits for a new slot. Once a slot is detected in step 302, the AF is updated in step 304 using equations (3) or (4), and the pattern is updated in step 306 using equations (5) - (6) or (7) - (8).
In step 308 it is tested whether the time slot belongs to a new packet. If no new packets have been detected, the method returns to step 302. If a new packet has been detected, the packet is tested in step 310, and if a bad packet has been detected, the method proceeds to step 312. In step 312, the value of L is updated according to the following equation:
Lnew＝max(LOld age-δ，LMinimum size)， (9)
Where δ is the step size. In one embodiment of the invention, the step size is set to 0.25 dB. Handle LMinimum sizeIs set to the minimum value that L can reach. In one embodiment of the invention, L isMinimum sizeIs limited to-1 dB. The method then returns to step 302.
In step 310, if a good packet is detected, the method proceeds to step 314. In step 314, the pattern is tested. If the AT is in the fast boot mode, the value of L is updated in step 316 according to the following equation:
Lnew＝min(LOld age+δ，LMaximum of)， (9)
Where δ is the step size. In one embodiment of the invention, the step size is set to 0.25 dB. Handle LMaximum ofIs set to the maximum value that L can reach. In one embodiment of the invention, L isMaximum ofIs limited to 3 dB. Once L is updated in step 318, the method returns to step 302.
If a normal mode is detected in step 314, the method continues to step 318, where the value of L is updated according to the following equation:
Lnew＝min(LOld age+ target _ PER · δ, LMaximum of) (10)
In equation (9), δ is the step size. In one embodiment of the invention, the step ofThe length is set to 0.25 dB. Target _ PER is the PER to be maintained. Handle LMaximum ofIs set to the maximum value that L can reach. In one embodiment of the invention, L isMaximum ofIs limited to 3 dB. Once L is updated in step 318, the method returns to step 302.
The previous description of the preferred embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without the use of the inventive faculty. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.