Phone standby time in cdma2000: the quick paging ... - Google Sites

5 downloads 100 Views 217KB Size Report
supported by Qualcomm Inc. The authors are with Qualcomm Inc., San Diego, 92121 CA USA (e-mail: ssarkar@qualcomm.com; bbutler@qualcomm.com; ...
1240

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 50, NO. 5, SEPTEMBER 2001

Phone Standby Time in cdma2000: The Quick Paging Channel in Soft Handoff Sandip Sarkar, Member, IEEE, Brian K. Butler, Member, IEEE, and Edward G. Tiedemann, Jr., Senior Member, IEEE

Abstract—cdma2000 is one of the proposals currently being reviewed for the 3G systems. This paper addresses one of the areas of improvement of cdma2000 over the IS-95 1 -based systems—the standby time of the phone. This paper proposes a novel scheme for the paging channel by splitting it into three parts. It analyzes the design of the quick paging channel (QPCH) and characterizes its performance for various conditions. This paper also looks at a way to put this common channel in soft handoff, thereby improving the reliability and leading to a better standby time for the cdma2000 phones. Both the physical layer and network implications are analyzed in detail. Index Terms—cdma2000, quick paging channel, paging indicator, soft handoff, standby time, 3G.

I. INTRODUCTION

T

HE paging channel is used to communicate from the base station to the mobile station when the mobile station is not assigned to a dedicated channel, i.e., the traffic channel. The paging channel carries overhead messages, pages, acknowledgment to messages sent by the mobile station on the access channel, and channel assignments. A. IS-95 Paging Channel Procedure The paging channel is transmitted at a constant rate, either 9600 b/s or 4800 b/s. This is in contrast to the traffic channel, which is variable rate. The paging channel is also transmitted at a constant power of about 10–15% of the total transmitted power is used for a 9600 b/s channel. This is because the paging channel does not know the required power to reach the mobile station [1]. There are basically the following operating scenarios for the paging and access channel: call origination, call termination, registration, status request, shared secret data (SSD) update and unique challenge. For call origination, the mobile station autonomously enters the system access state and sends the origination message. The base station then sends an acknowledgment. When the channel is set up, the base station sends the channel assignment message to the mobile station. This provides the information on the dedicated channels that the mobile station is to use. The mobile station then begins using the dedicated channels. Manuscript received May 30, 1999; revised April 20, 2001. This work was supported by Qualcomm Inc. The authors are with Qualcomm Inc., San Diego, 92121 CA USA (e-mail: [email protected]; [email protected]; [email protected]). Publisher Item Identifier S 0013-9545(01)08252-8.

For call termination, the base station sends the general page message to the mobile station. The mobile station responds with the page response message which indicates the cell in which the mobile station is located. When the channel is set up, the base station sends the channel assignment message to the mobile station. This provides the information on the dedicated channels which the mobile station is to use. The mobile station then begins using the dedicated channels. When performing an origination or termination, the base station may send a status message, perform an SSD update, or send a unique challenge to the mobile station. The operations may also be performed without doing a call termination by having the base station send a general page message to the mobile station to move the mobile station into the system access state to perform these operations. When the base station has completed these operations, the base station may send a release to the mobile station or just let the mobile station time out of the system access state. For registration, the mobile station sends the registration message; the base station sends an acknowledgment and then may later send a registration accepted order, a registration rejected order, or a service redirection message. B. IS-95 Paging Channel Slotted Operation The IS-95 paging channel can be operated in a way that reduces the power consumption of a mobile station. This is called slotted mode. In slotted mode, the mobile station wakes up at a pre-determined interval to check whether there is a page. If there is no page, the mobile station is permitted to stop monitoring the paging channel until its next assigned slot. As the time spent in demodulating a single slot is far less than the time required to continuously monitor the paging channel, there are considerable power savings in the mobile station resulting in longer battery life. Paging channel slotted operation is described in the IS-95 standard [2]. II. STANDBY TIME ANALYSIS One of the features that is extremely desirable to cell phone users is how long is the time before the phone needs to be recharged. The new paging channels is cdma2000 are aimed to increase this as much as possible over IS-95 phones. To get an estimate of the standby time goals of the phone, the following assumptions are made: Consider no coverage holes, registration transmits as included in talking budget, everyone using the same type of phone and battery for comparison purposes and

0013–9545/01$10.00 © 2001 IEEE

Authorized licensed use limited to: Univ of Calif San Diego. Downloaded on September 17, 2009 at 20:34 from IEEE Xplore. Restrictions apply.

SARKAR et al.: PHONE STANDBY TIME IN cdma2000

Fig. 1.

1241

Generation of the QPCH.

the phone being left on for 24 hours a day. In a two-state battery consumption model, battery life is broken down into talk time (TT) and standby time (ST). Let be the fraction of time using talk time, be the fraction of time using standby time, and UT be the utility time of phone. Then

If the standby time was unlimited, the asymptotic case could be . The defined as a limit: achieves an utility time of 90% of the target standby time h. asymptotic value. For a heavy user (120 min/day), h, and for a light For a medium user (25 min/day), h. Hence, improvements of up to user (5 min/day), 1020 h of standby time per 2 h of talk time are very useful to the customer. The utility times for the heavy, medium, and light user would be 0.98, 4.32, and 15.4 days, respectively. Using a two-state battery consumption model for standby be the fraction of time the MS is awake, and time, let the fraction of time the MS is asleep, i.e., . If denotes the battery capacity, denote the current drawn by the MS when awake, and the current drawn when asleep, then

Thus, for the two schemes, the standby time gain puted as

can be com-

Now, defining (1) Equation (1) defines the gain in standby time of the phone, and is directly proportional to its battery life improvement. III. STRUCTURE OF THE PAGING CHANNEL FOR cdma2000 cdma2000 has been developed as a standard in the United States as a third generation technology for wireless phones [3], [4]. The organization of cdma2000 common channels will be a structure consisting of three forward link physical channels: the quick paging channel (QPCH), the broadcast channel (BCCH), and the common control channel (CCCH). The QPCH carries indications of pages directed to the mobile station. The base station transmits on the QPCH whenever the base station needs to contact a mobile station operating

in slotted mode. The BCCH carries overhead information and broadcast short messages. Overhead information is not required to be continually transmitted; broadcast messages are only required to be transmitted when a broadcast message is to be sent. The CCCH is transmitted when the base station responds to the mobile station. All messages on the CCCH can be transmitted to the mobile station in a soft handoff mode. Furthermore, all messages can be transmitted to the mobile station with approximately the amount of power needed to contact the mobile station. These techniques result in greater overall system capacity and greater reliability. In this paper, we will analyze only the QPCH. Further details can be found in [5], [6]. A. Operation of the QPCH The quick paging channel contains single bit messages to direct slotted-mode mobile stations to monitor their assigned slot on the paging channel. The QPCH is generated according to Fig. 1. From the figure, it is seen that the QPCH data rate is 9600, 4800, or 2400 b/s, and is specified in an overhead message. Each single bit message is transmitted twice per 80-ms slot. i.e., each QPCH bit corresponds to two QPCH symbols. The QPC protocol provides for scheduling the transmission of the paging indication symbols for a specific mobile station in certain assigned slots. The mobile station is hashed into two groups: A or B. (The hashing function is defined in [5].) Let denote the start time of the paging channel or the forward common control channel. Group A mobile station receives the first paging indims and ms and second cation sysbol between ms and ms. Group B mosymbol between bile station receives the first paging indication symbol between ms and ms and second symbol between ms and ms. The timeline of QPC is shown in Fig. 2. , usually 0, 1 The term SCI refers to slot cycle index ( or 2). The length of the paging channel slot cycle is s. B. Modulation Selection The QPCH symbol detection is pilot aided. Hence, coherent on-off keying (OOK) and coherent binary phase shift keying (BPSK) are considered as possible modulation schemes. De. (see [7]). In AWGN, for fine coherent maximum likelihood (ML) detection, the symbol error , where is rate is given by the noise spectral density. If denotes the average energy per bit . Let denote the probability of a one for BPSK: . Thus, OOK being transmitted. Then, for OOK, . As an example, for conserves average energy when . Fig. 3 eight pages per 80 ms at 9600 b/s, compares the performance of OOK versus BPSK in AWGN.

Authorized licensed use limited to: Univ of Calif San Diego. Downloaded on September 17, 2009 at 20:34 from IEEE Xplore. Restrictions apply.

1242

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 50, NO. 5, SEPTEMBER 2001

Fig. 2. QPCH timeline.

and data services together, eight pages per paging channel are a very significant amount of the CCCH throughput [8]. Denote the collision probability, to be the probability that the mobile finds that a one has been transmitted in its slot although it was not paged. Take care to distinguish this from the error probability of a false alarm that occurs when a bit 0 is incorrectly detected as bit 1. That is a decoder error, while is a function of network traffic. Let denote the number of pages in a given slot, denote the total number of pages, and denote the number of available bit positions to hash the bits to. Then we . Assume that the hashing function have is good, i.e., uniformly distributes the mobiles to the number of available bit positions. Then

Thus, the following is obtained:

Fig. 3. Performance in AWGN.

It is clear that the average power used is far less in the OOK scheme. It is estimated that we need 5% more total power per base station to support BPSK. Yet another concern in a practical communication system is the peak to average ratio of the envelope. Experiments were performed on moderately loaded channels (25 voice users) with the OOK QPCH operating at 3 dB below the pilot transmit level. No discernable effect was observed at the base station. Even for extremely lightly loaded channels, there was a very small effect as seen in Fig. 4. Thus, OOK was chosen to benefit from the power saving.

Each 80 ms slot is divided into two equal halves of 40 ms each for hashing the bits. Hence, if the data rate is b/s, the number . This evaluates to 384 for of available bit positions are 9600 b/s, 192 for 4800 b/s, and 96 for 2400 b/s. The previous formula holds for each 40 ms subslot. The probability that it is falsely paged for both slots can be approximated to be the square . Fig. 5 plots the results for . of , if we assume Realistically, using Little’s formula 30 Erlangs of traffic per sector and an average call duration of 120 s, we get call setups per second per sector. Assuming 30% of them to be mobile terminated and two pages per call setup, we have 0.15 calls per second, or one call every 6–7 s on an average. This excludes data and short message service (SMS).

C. Collision Analysis For any particular mobile, the probability of it being falsely paged depends on the network load. If the network handles half a million calls per hour, it corresponds to roughly 11 pages per 80 ms slot. Such a case is definitely one of the worst cases. Even if we include voice calls, short message service (SMS)

IV. DETECTION OF THE QPCH BIT Each QPCH bit consists of two symbols, decoded as shown in Fig. 6. Each symbol is decoded coherently. The pilot is used to provide the phase reference for doing so. Multiple paths may be

Authorized licensed use limited to: Univ of Calif San Diego. Downloaded on September 17, 2009 at 20:34 from IEEE Xplore. Restrictions apply.

SARKAR et al.: PHONE STANDBY TIME IN cdma2000

1243

Fig. 4. Peak-to-average ratio with QPCH in OOK.

enough that the fading envelope is a constant during one signaling interval (see Fig. 7). For each symbol, the received signal can be written as

Fig. 5. Single bit collision probabilities.

combined using the maximal ratio combining as described in [9]. To obtain a reliable estimate, the decision is tri-valued as mentioned before (erasure, one, or zero). The sysbol is declared an erasure based solely on the pilot strength. If the pilot level falls below a certain threshold, the sysbol is declared an erasure, else the symbol is detected in the regular way. This threshold is tuned to keep the miss probability within acceptable limits. For cdma2000 the target erasure probability is 1 10 . The performance in AWGN is the same as outlined in Section III-B. Next, the channel is considered in a fading scenario.

Here, the transmitted signal is is the complex is the valued Gaussian noise corrupting the signal, and time varying channel gain. Let us assume that the signal fades perfectly. For a fixed slowly enough to estimate the phase , the channel is fixed and the error statistics can be calculated at time . Then, averaging as a function of , the value of yields the final error statistics. over the pdf of be a zero mean Gaussian random Let , and be the signal energy. The variable with variance detection problem can be set up as

For a threshold , declare if , otherwise. Let and denote the miss and false alarm probabilities, respectively. In terms of the complementary error function, this is written as (see [9])

A. Rayleigh-Fading Channels The main results are presented in this section, with an outline of the approach. It is assumed that the fading process is slow

Authorized licensed use limited to: Univ of Calif San Diego. Downloaded on September 17, 2009 at 20:34 from IEEE Xplore. Restrictions apply.

1244

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 50, NO. 5, SEPTEMBER 2001

Fig. 6. Decoding the QPCH.

Evaluating the integral

The corresponding false alarm probability is found as

Thus, the design reduces to a proper choice of , followed by the calculation of the false alarm probability. If the second symbol is far enough temporally to assume that it is independent from the first, the same analysis holds for that, too, as is the case here. B. Performance Curves Fig. 7.

Rayleigh-fading channel.

The channel is estimated using the pilot transmitted from the base station. It is assumed that the channel fading is slow enough such that the QPCH symbol signal-to-noise ratio (SNR) is pro. Note portional to the estimated pilot energy. So, let . For . that Next, let be a random variable with a Rayleigh probability . Then, has a chi-square pdf with two density function . The miss degrees of freedom. Define probability can now be found as

Substitution of makes the integral a function of and . Let . the SNR of the QPCH symbol

The system design here is similar to a Neyman–Pearson deand try to minimize tection problem in the sense that we fix . For the present system, we target . The simulation results closely matches the theoretical predictions. In the figures, the numbers in parentheses refers to the value and curves to obtain of . One can reverse the roles of the the curves for thresholds between 0.5 and 1. The curves for and overlap for . Note that these are per symbol error probabilities. It is now possible to get the whole system performance based on the state diagram presented in Fig. 6. The cost of a false alarm for the phone is to lose some battery life by monitoring the paging channel. However, a miss could mean a lost call. Since the pilot is used to determine the fading condition, the symbol is declared an erasure when the pilot is in a deep fade, and the phone monitors the CCCH. Typically, the pilot is sent at 3-dB higher power than the QPCH. A threshold is chosen such that such that the miss probabilities will still meet

Authorized licensed use limited to: Univ of Calif San Diego. Downloaded on September 17, 2009 at 20:34 from IEEE Xplore. Restrictions apply.

SARKAR et al.: PHONE STANDBY TIME IN cdma2000

the design criteria in a deep fade. This is governed by the receiver noise floor. When the received pilot SNR is not within the required confidence interval, the symbol is called an erasure. V. SOFT HANDOFF ON QPCH An MS at a distance of from a BS suffers a signal attenuation . Here is the shadowing proportional to effect, which is often modeled as a lognormal process with standard deviation 8 dB. Experimental data shows that a choice of models the realistic scenario pretty well [10]. Due to the susceptibility of call drops at cell boundaries due to too much signal attenuation and other cell interference, it is often advantageous to do soft handoff (SHO) over hard handoff [11], [12]. This section analyzes such a condition in detail. In order to motivate the discussion, we define the following. denote the received signal strength, the thermal noise Let to be the interference due to other cells. The geometry and of a MS is defined to be . The utility of this parameter stems from the fact that a higher geometry implies that the MS is well within the cell of a BS. If we assume that is dominated by , a 0 dB geometry implies that the MS is at a cell boundary (in reality a slightly negative geometry due to ). A very negative geometry usually means a deep fade, or that the MS is not listening to the optimal BS or is going out of coverage. Assume the same conditions as in the single path case, except that we have independently faded paths, each having statistics similar to the single path case. Let path have a Gaussian . Let the energy in path be and noise the channel gain . Then, the hypothesis problem is set up as follows:

The detector with threshold

has the following performance:

1245

Let the received SNR per symbol for path be: , and the total SNR . It is assumed are iid random variables with Rayleigh that distribution. A. Equal Strength Base Stations equally strong BSs with energy per If path are received, the average SNR per channel is . Then, the total SNR has pdf: . , and Let . The error probabilities are given by

For two paths, the performance is plotted in Fig. 8. and, hence, the 3-dB Note that this plots the combined combining gain is already included in the graphs. It is observed that the diversity gain due to combining is almost 1.7 dB for and . Thus, in a practical system, where the combining gain is not quite 3 dB, one can expect a diversity gain of about 1 dB for the miss probabilities. This is confirmed by actual simulations. B. Unequal Strength Base Stations The general solution is found from calculating the pdf of the SNR for the particular configuration [13]. As an example, consider the case when the BS receives signals from two BSs, one and two weaker paths with energy arriving with energy each. If they are combined as described above with each path being faded identically and independently, the pdf of the total SNR can be similarly computed to obtain Fig. 9. It shows some diversity gain over the previous case. While this is great from a physical layer perspective, it leads to certain issues of network management. These are addressed next. VI. CONTROLLING SOFT HANDOFF

Set the threshold as a function of the sum of received pilot . Then, we have SNRs, i.e.,

In the overhead messages the BS lists the BSs which are permitted to transmit to the MS in SHO on the CCCH and QPCH. It should be noted that the IS-95-B access handoff and access probe handoff may be performed with some BSs and soft handoff performed with other BSs. In particular, IS-95 has a flag for every member of the neighbor list for which Access handoff and access probe handoff are permitted. A flag is introduced for every member of the neighbor list in which soft handoff is permitted on the CCCH. A separate flag is introduced for every member of the QPCH. This is illustrated for five cells with three sectors each in Table I. In the table, ACS refers to the access handoff flag, CCS to the CCCH SHO allowed flag, QCS to the QPCH SHO allowed flag, and AcHO implies that access handoff is allowed. The MS is located in sector A1. New MSs are permitted to perform SHO on the CCCH with all sectors in cells A and B. The MS is not

Authorized licensed use limited to: Univ of Calif San Diego. Downloaded on September 17, 2009 at 20:34 from IEEE Xplore. Restrictions apply.

1246

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 50, NO. 5, SEPTEMBER 2001

Fig. 8.

Peformance of QPCH in SHO at cell boundary.

Fig. 9.

Unequal strength signals. TABLE I ILLUSTRATION OF THE SHO FLAGS

of cells may be used for SHO on the QPCH since paging can be more readily handled from a central controller. However, this is not required and there is total flexibility in set of cells that permit SHO on the CCCH. Note that SHO on the QPCH is not permitted for cell D. This may occur since cell D is in a different registration zone. No forms of handoff are permitted to cell E. This may be because it is controlled by a different BSC. Fig. 10 illustrates the scenario. To show the operation of the CCCH, we use a call origination, as shown in Fig. 11. It uses the BSs and the various flags previously described. It is assumed that soft handoff is permitted on the CCCH between BTSs. The MS sends the origination message while located in sector A1 of the corresponding BTS. Two additional BSs have pilots strong enough to combine: B1 and C1. The flags that the MS receives in the overhead messages indicate that soft handoff is permitted with B1 and that access handoff is permitted with C1. The origination message is received at a BTS and forwarded to the BSC. At the BSC, the message is processed. At the BSC, an acknowledgment is generated and sent to the BTSs for transmission to the MS on the CCCH. The BTSs selected to be sent the message are those corresponding to pilots reported to be strong by the MS and have CCS set to 1. The BSC send the layer 2 acknowledgment to these BTSs and these BTSs send the layer 2 acknowledgment to the MS in soft handoff. In the example, the MS reports B1 and C1 to have strong pilots, and since CCS set to 1 in B1, the BSC sends the acknowledgment to BTS B and A1. After setting up the channel, the BSC then sends the channel assignment message (or information to determine the channel assignment message) to BTSs A, B, and C. The BSC includes BTS B since B1 was reported by the MS, and CCS was set to 1. The BSC includes BTS C since C1 was reported and ACS was set to 1. The channel assignment message is transmitted in soft handoff mode from BSs A1 and B1. As in IS-95-B, the channel assignment message is also transmitted from BS C1. This transmission does not have to be in a soft handoff mode since the MS is not combining the transmission with the transmissions from other BSs. While this has been shown for an origination, the same methods work for all other exchanges which are begun by the MS. A. QPCH Operation: Call Termination

permitted to perform SHO on the CCCH with sectors from other cells. It should be noted that it may be desirable to restrict SHO on the CCCH to only sectors of the same BS. In this case, SHO of the CCCH would be only to A2 and A3. This is because it is easier to synchronize and control the SHO from a cell where only single processor is involved. It also permits layer 2 to be fully run from the BTS. The example also shows that SHO on the QPCH may be done in cells A, B, and C. A wider number

To show the operation of the F-QPCH, we use a call termination as shown in Fig. 12. It uses the BSs and the various flags that have been previously described. As was previously discussed, soft handoff may be restricted to a single BTS. In what follows, it is assumed that soft handoff is permitted on the QPCH between BTSs. The BS sends a quick page at a particular time that is given by the IMSI of the MS and the configuration of the BS. Typically, the configuration of all BSs in a paging region are configured the same so all BSs would be transmitting the quick page at the same time. For those base stations that transmit the quick page at the same time, the MS can combine these BSs in soft handoff. For a particular BS, the overhead messages would have the QCS flag set to 1 for a neighboring BS when the neighboring BS sends the same quick pages and at the same time. If the above example, BS A1 would set QCS to 1 for BSs A2, A3, B1, B2,

Authorized licensed use limited to: Univ of Calif San Diego. Downloaded on September 17, 2009 at 20:34 from IEEE Xplore. Restrictions apply.

SARKAR et al.: PHONE STANDBY TIME IN cdma2000

Fig. 10.

SHO on QPCH.

Fig. 11.

Illustration of call origination.

B3, C1, C2, and C3. Base station D1, D2, and D3 do not have QCS set to 1. This is since these base stations are not in the same paging area or are otherwise unable to transmit the quick page at the same time. When monitoring the appropriate slot on the QPCH, the MS combines the transmissions from multiple BSs in a soft handoff mode. The MS then determines whether it has received a quick page. It should be noted that the MS is not required to combine signals from other cells in the same slot that are not in soft handoff. This is because the MS would pick up the strongest BS when receiving the quick page. However, there would be some benefit if the MS knew the configuration

1247

of the neighboring BS and, thus, could also process the quick page transmitted by a neighboring BS. However, the MS may not have received the configuration information from the neighboring cell and thus may not know the time that the quick page is being transmitted. It should be noted that the neighboring BS would typically use the soft handoff mode with a neighboring BS if the neighboring BS were sending the power control bits in the same slot. After the MS has been alerted by the quick page, the MS begins to monitor the CCCH. The BS will then send a page message, similar to a normal page, on the CCCH. This message contains the full address of the MS. It should be noted

Authorized licensed use limited to: Univ of Calif San Diego. Downloaded on September 17, 2009 at 20:34 from IEEE Xplore. Restrictions apply.

1248

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 50, NO. 5, SEPTEMBER 2001

Fig. 12.

Illustration of call termination.

that full page step is not required and the MS could respond directly with the page response message after receiving the quick page. However, the false alarm rate on the quick paging channel may be sufficiently high so that it is preferable for the MS to wait for the page message before sending the page response message. B. Standby Time Comparison To provide a comparison, consider eight pages per 80-ms slot. Assume 5 ms warmup time, and a cost of 54 ms to monitor full paging slot. Using the present design at 4800 b/s, the amount of improvement depends on hardware platform. Define to be the ratio of the current drawn when the phone is awake to that when the phone is asleep. Presently, for many commercially available to is common. This may go up a phones, a value of little in the near future. Fig. 13 summarizes the results. In Fig. 13, the axis represents the percentage improvements in the phone standby time with the QPCH over the IS-95–based systems. SCI was defined at the end of Section III-A.

Fig. 13.

Standby time improvement.

VII. CONCLUSION The design of the new paging channels in cdma2000 allows for seamless interoperability between all combinations of base stations and mobile stations that do and do not support this new

Authorized licensed use limited to: Univ of Calif San Diego. Downloaded on September 17, 2009 at 20:34 from IEEE Xplore. Restrictions apply.

SARKAR et al.: PHONE STANDBY TIME IN cdma2000

feature. It does not significantly impact call setup time, or increase the missed page probability. However, it greatly reduces the amount of time the phone spends monitoring the slots of the paging channel, if this feature is supported by the phone. The introduction of soft handoff allows the message to be transmitted at lower power levels. This significantly reduces the interference to the other mobiles, and specially leads to a far better performance on the cell boundaries, where the problem of call drops are most acute. In the paper, first the traditional paging channel of IS-95 was described. Next, the design criteria for the QPCH was presented and its performance analyzed. The use of soft handoff was investigated and found to be useful. However, controlling soft handoff involves significant signaling requirements and these were looked at in detail. Finally, the standby time improvements were quantified. Thus, the QPCH scheme causes considerable improvements in standby time at very little cost. The required transmit power level can be reduced if this channel is put in soft handoff [14], [15]. Thus, it has been incorporated into the standards for cdma2000. It is a distinctive feature of cdma2000 when compared to the other 3G wireless CDMA proposals. ACKNOWLEDGMENT The authors would like to thank K. Gilhousen, Y. C. Jou, R. Rezaiifar, and Y. C. Lin for their help with this work. REFERENCES [1] S. Willenegger, “cdma2000 physical layer: An overview,” J. Commun. Networks, vol. 2, pp. 5–17, Mar. 2000. [2] Mobile Station-Base Station Compatibility Standard for Dual-Mode Wideband Spread Spectrum Cellular Systems, 1998. [3] Y. S. Rao and A. Kripalani, “cdma2000 mobile radio access for imt2000,” in Proc. ICPWC, Jaipur, Rajastham, India, Feb. 1999, pp. 6–15. [4] D. N. Knisely, Q. Li, and N. S. Ramesh, “cdma2000: A third-generation radio transmission technology,” Bell Labs Tech. J., pp. 63–78, July–Sept. 1998. [5] Mobile Station-Base Station Compatibility Standard for Dual-Mode Wideband Spread Spectrum Cellular Systems, 1999. [6] S. Sarkar and E. Tiedemann, “The paging channel in cdma2000,” in Proc. ICON, Brisbane, Queensland, Australia, Sept. 1999, pp. 257–264. [7] A. Papoulis, Probability, Random Variables, and Stochastic Processes, 2nd ed. Sydney, Australia: McGraw-Hill, 1984. [8] S. Sarkar and B. Butler, “Phone standby time and the quick paging channel,” in Proc. PIMRC, Osaka, Japan, Sept. 1999, pp. 1341–1346. [9] J. G. Proakis, Digital Communications, 4th ed. New York: McGrawHill, 2000. [10] A. J. Viterbi, CDMA Principles of Spread Spectrum Communication. Reading, MA: Addison-Wesley, 1995. [11] W. C. Y. Lee, Mobile Cellular Telecommunications, 2nd ed. New York: McGraw-Hill, 1995.

1249

[12] T. S. Rappaport, Wireless Communications: Principles and Practice. Englewood Cliffs, NJ: Prentice-Hall, 1999. [13] M. Schwartz Jr., W. R. Bennet, and S. Stein, Communication Systems and Techniques. New York: McGraw Hill, 1966. [14] S. Sarkar, B. Butler, and E. Tiedemann, “Soft handoff on the quick paging channel,” in Proc. IEEE GLOBECOM, Rio de Janeiro, Brazil, Dec. 1999, pp. 2794–2798. [15] S. Sarkar and E. Tiedemann, “Common channel soft handoff in cdma200,” IEEE Trans. Microwave Theory Tech., vol. 46, pp. 938–950, June 2000.

Sandip Sarkar (M’97) was born in Calcutta, India. He received the B.Tech degree in electrical engineering from the India Institute of Technology (IIT), Kanpur, in 1992 and the M.A. and Ph.D. degrees in telecommunications and signal processing from Princeton University, Princeton, NJ, in 1994 and 1996, respectively. Since then, he has been a Member of the Technical Staff in the Corporate Research and Development, Qualcomm Inc., San Diego, CA. He has been involved in the design of the physical layer of 3G systems as an active member of the ITU. He has also been involved in the system design of Globalstar and played a major part in the design of the forward and reverse link chipsets for these phones. His research interests include wireless communications, error-control coding, information theory, and associated signal processing systems.

Brian K. Butler (S’90–M’90) received the B.S. degree from Harvey Mudd College, Claremont, CA, in 1989, and the M.S.E.E. degree in from Stanford University, Stanford, CA, in 1990. He has been with Qualcomm, Inc., San Diego, CA, for over ten years and has made contributions to the physical layer design specification and the ASIC designs used in IS-95 CDMA digital cellular. He was involved in early CDMA field tests, particularly in analyzing the performance and system capacity. He was also involved in the Globalstar LEO communicaiton system and handset ASIC design. For the last several years, he focused on 3G cdma2000 and increasing CDMA standby time by such commercial projects as MSM3000 and MSM5000. He is a coinventor of 12 issued U.S. patents and has several pending. He currently supervises the ASIC division’s main communication systems engineering department.

Edward G. Tiedemann, Jr. (S’72–M’86–SM’91) received the B.S. degree from Virginia Polytechnic Institute and State University, Blacksburg, the M.S. degree from Purdue University, West Lafayette, IN, and the Ph.D. degree from the Massachusetts Institute of Technology (MIT), Cambridge, in 1975, 1977, and 1986, respectively. While at Purdue University, he worked on bandwidth efficient modulation and at MIT, he worked in the areas of queueing theory and communications networks. He leads Qualcomm’s System Engineering Group, San Diego, CA, developing technology for wireless standards and also directs Qualcomm’s worldwide standardization activities. He was very instrumental in the design and development of the TIA/EIA-95 CDMA system, also called cdmaOne, and in the development of the third generation cdma2000 system. Prior to joining Qualcomm in 1988, he spent 12 years at MIT Lincoln Laboratory. Dr. Tiedemann chaired the standardization group in the Telecommunications Industry Association (TIA), responsible for the TIA/EIA-95 physical layer, and chaired the Joint Technical Committee on Wireless Access (JTC) group, responsible for J-STD-008, the PCS version of TIA/EIA-95.

Authorized licensed use limited to: Univ of Calif San Diego. Downloaded on September 17, 2009 at 20:34 from IEEE Xplore. Restrictions apply.