Zhang, Guodong (Patchogue, NY, US)

Plaque It! Sponsored by: Flash of Genius

10/918889

01/20/2005

08/16/2004

Images are available in PDF form when logged in. To view PDFs, Login  or  Create Account (Free!)

View patents that cite this patent

Click for automatic bibliography generation

InterDigital Technology Corporation (Wilmington, DE, US)

370/328

(IPC1-7): H04Q007/00

Volpe And, Koenig Dept Icc P. C. (UNITED PLAZA, SUITE 1600, PHILADELPHIA, PA, 19103, US)

1. A method of performing call admission control (CAC) in a network including at least one target cell and a plurality of neighboring cells, the method comprising: (a) receiving a request to setup or reconfigure a radio access bearer; (b) measuring the signal-to-interference ratio (SIR) of the request; (c) measuring the SIR of existing coded composite transport channels (CCTrCHs) in the target cell and the neighboring cells; and (d) determining whether to accept or reject the request based upon the measurements performed in steps (b) and (c).

2. The method of claim 1 further comprising: (e) determining the probability that the instantaneous load of a time slot used to fulfill the request will exceed a maximum allowed value; and (f) further determining whether to accept or reject the request based upon the determination of step (e).

3. A method of performing call admission control (CAC) in a network including at least one target cell and a plurality of neighboring cells, the method comprising: (a) receiving a request to setup or reconfigure a radio access bearer; (b) employing dedicated measurements to determine whether to accept or reject the request; (c) if the dedicated measurements are not available, employing common measurements to determine whether to accept or reject the request; (d) if both the dedicated and common measurements are not available, determining the probability that the instantaneous load of a time slot used to fulfill the request will exceed a maximum allowed value; and (e) determining whether to accept or reject the request based upon the determination of step (d).

4. The method of claim 3 wherein the dedicated measurements are performed by a user equipment (UE).

5. The method of claim 3 wherein the common measurements include uplink time slot interference signal code power (ISCP).

6. The method of claim 3 wherein the common measurements include downlink carrier power.

7. The method of claim 3 wherein the common measurements are performed on at least one of the target cell and the neighboring cells.

8. The method of claim 3 wherein the determination of step (d) is based on a time variant signal-to-interference ratio (SIR).

9. The method of claim 8 wherein the time variant SIR comprises a required SIR and a SIR range of the radio access bearer.

10. The method of claim 8 wherein step (d) further comprises measuring the SIR of existing coded composite transport channels (CCTrCHs) in the target cell and the neighboring cells.

11. In a network in which codes of existing coded composite transport channels (CCTrCHs) are assigned to different time slots, a method of performing call admission control (CAC) to optimize load or carrier power of a target cell and neighboring cells of the network, the method comprising: (a) assigning a code of a CCTrCH to a particular time slot; (b) determining the load of the target cell and the neighboring cells in the particular time slot; (c) determining whether the load of the target cell and the neighboring cells is below at least one predetermined threshold; (d) selecting the time slot that yields the lowest average load; and (e) repeating steps (a)-(d) until all of the codes are assigned.

12. In a network in which codes of existing coded composite transport channels (CCTrCHs) are assigned to different time slots, a method of performing call admission control (CAC) to optimize load or carrier power of a target cell and neighboring cells of the network, the method comprising: (a) for each time slot in the target cell and the neighboring cells, determining the probability that the instantaneous load of the time slot will exceed a maximum allowed value; (b) assigning a code of a CCTrCH to a particular one of the time slots having the lowest probability as determined in step (a); (c) updating the instantaneous load probability of the particular time slot; (d) repeating steps (a)-(c) until all of the codes are assigned; and (e) determining whether or not to admit a new user to the network based on whether or not a total instantaneous load probability associated with the new user is less than a predetermined threshold.

CROSS REFERENCE TO RELATED APPLICATIONS

This application a continuation of U.S. patent application Ser. No. 10/331,442, filed Dec. 30, 2002, which claims priority from U.S. Provisional Patent Application No. 60/383,273, filed May 24, 2002, which is incorporated by reference as if fully set forth.

BACKGROUND

The present invention is directed to strategies and algorithms by which CDMA networks perform call admission control (CAC) in three different situations: 1) when only common-measurements are available; 2) when no measurements are available; and 3) based on outage probability requirements. In particular, it is applicable to Universal Mobile Telephone System-Time Division Duplex (UMTS-TDD) systems.

Call admission control (CAC) is a function responsible for deciding whether to accept or reject a request to setup or reconfigure a radio access bearer in the radio access network (RAN). CAC is performed at the Controlling Remote Network Controller (CRNC). Sometimes, although UE dedicated measurements are not available, common measurements such as uplink timeslot ISCP and downlink carrier power are available. Thus, the CRNC must have the ability to perform CAC properly with only common measurements.

Sometimes, no measurements are available. In this case, the CRNC must have the ability to perform CAC properly in absence of measurements.

In a UMTS-TDD system, the required signal-to-interference ratio (SIR) of a user changes with time because of fading and imperfect power control. In such a system, outage probability is a good measure of system quality of service (QoS). CAC should be designed to provide the required outage probability to the system.

SUMMARY

The present invention is able to perform CAC under a variety of conditions. When only common measurements are available, CAC will accept or reject a request to setup or reconfigure a radio access bearer (i.e., radio link) based on: 1) common measurements of the target cell and neighboring cells; 2) estimated common measurements after admission; and the measure of the loading of the target cell and neighboring cells, (which are also estimated from common measurements). When the CAC assigns codes of the CCTrCH to different time slots, it will try to optimize the load or carrier power of the target cell and neighboring cells; whereby average or weighted average load/carrier power of the target cell and neighboring cells can be used.

When no measurements are available, CAC accepts or rejects the request based on estimated load of the target cell and neighboring cells. The load may be estimated using the following information: 1) required SIR of the request, (this represents the load of the request); and 2) required SIR of existing coded composite transport channels (CCTrCHs) in the target cell and neighboring cells, (this represents the current load of the target cell and neighboring cells). When CAC assigns codes of the CCTrCH to different time slots, it will try to optimize the load of the target cell and neighboring cells; whereby average or weighted average load of the target cell and neighboring cells can be used.

Finally, an alternative embodiment of the present invention can accept or reject a request based on the estimated outage probability of the target cell and neighboring cells. The estimated outage probability is based on assumption of the time-variant SIR. One alternative for estimating the outage probability is to use: 1) required SIR and SIR range of the radio link setup/reconfiguration request, (this represents the range of load of the request); 2) required SIR and SIR range of existing CCTrCHs in the target cell and neighboring cells, (this represents the current range of load of the target cell and neighboring cells); and 3) the outage probability, which is defined as the probability that the instantaneous load in a time slot exceeds a maximum allowed value. When CAC assigns codes of the CCTrCH to different time slots, it will try to minimize the total outage probability of the CCTrCH, ensuring that the outage probability of assigned timeslots in neighboring cells also meets the requirements.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a flow diagram of a slot selection procedure for the uplink and downlink for a first embodiment in accordance with the present invention.

FIG. 2 is a flow diagram of a slot selection procedure for the uplink and downlink for a second embodiment in accordance with the present invention.

FIG. 3 is a flow diagram of a slot selection procedure for the uplink and downlink for a third embodiment in accordance with the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

The present invention will be described with reference to the drawing figures wherein like numerals represent like elements throughout.

A first embodiment of the present invention is directed to CAC with only common measurements available. Referring to FIG. 1, the basic assignment procedure 10 for CAC in the uplink is shown. The procedure 10 commences with the first code in the code set (step 12 ). The load of an uplink time slot in cell i is defined as the load generated by the codes assigned in the same time slot in this cell and in first tier cells and second tier cells, since the load generated from cells beyond second tier is negligible. Preferably the load from neighboring cells is measured using the uplink time slot ISCP. Uplink time slot Interference Signal Code Power (ISCP) contains inter-cell interference only. For the target cell, the load after admission can be calculated as follows:
The load from its own cell is called the intra-cell load Load _{UL — Intra} (i) and is given by: $\begin{array}{cc}{\mathrm{Load}}_{\mathrm{UL\_Intra}}\left(i\right)={\alpha }_{\mathrm{UL}}·\sum _{h\in \Omega \left(i\right)}SI{R}_h& \mathrm{Equation}1\end{array}$
where Ω(i) is the set of codes assigned in this time slot in the cell i. Since the load is defined based on noise rise, then intra-cell interference I _intra may be given by: $\begin{array}{cc}\mathrm{Intra}\text{-}\mathrm{cell}\mathrm{noise}\mathrm{rise}=\frac{I_{\mathrm{intra}}}{N_0}=\frac{1}{1-{\mathrm{Load}}_{\mathrm{UL\_Intra}}\left(i\right)}& \mathrm{Equation}2\end{array}$
Which can be rewritten as: $\begin{array}{cc}{I}_{\mathrm{intra}}=\frac{N_0}{1-{\mathrm{Load}}_{\mathrm{UL\_Intra}}\left(i\right)}& \mathrm{Equation}3\end{array}$
The total interference can be found by: $\begin{array}{cc}\mathrm{Total}\mathrm{noise}\mathrm{rise}=\frac{I_{\mathrm{Intra}}+\mathrm{ISCP}}{N_0}=\frac{1}{1-{\mathrm{Load}}_{\mathrm{UL}}\left(i\right)}& \mathrm{Equation}4\end{array}$
Then, the total load is given by: $\begin{array}{cc}{\mathrm{Load}}_{\mathrm{UL}}\left(i\right)=1-\frac{1}{\mathrm{Total}\mathrm{noise}\mathrm{rise}}=1-\frac{N_0}{I_{\mathrm{Intra}}+\mathrm{ISCP}}& \mathrm{Equation}5\end{array}$

For neighboring cells, (tier one or tier two cells only), the load after admission can be calculated as follows:
The load of a time slot in cell j before admission is calculated using Equations 1-5, and denoted by Load _{UL — Before} (j). If SIR _t is the required SIR target of the code to be assigned in this time slot in the target cell i, the load of a time slot in cell j after admission, denoted by Load _UL (j), is given by: $\begin{array}{cc}{\mathrm{Load}}_{\mathrm{UL}}\left(j\right)=\{\begin{array}{cc}{\mathrm{Load}}_{\mathrm{UL\_Before}}\left(j\right)+{\beta }_{\mathrm{UL}}·{\mathrm{SIR}}_t& \mathrm{if}j\in \mathrm{Tier}\mathrm{One}\left(i\right)\\ {\mathrm{Load}}_{\mathrm{UL\_Before}}\left(j\right)+{\sigma }_{\mathrm{UL}}·{\mathrm{SIR}}_t& \mathrm{if}j\in \mathrm{Tier}\mathrm{Two}\left(i\right)\end{array}& \mathrm{Equation}6\end{array}$
Where Tier One(i) is the set of codes assigned in this time slot in tier-one neighboring cells of the cell i, and Tier Two(i) is the set of codes assigned in this time slot in tier-two neighboring cells of the cell i.

For CAC, a code will be admitted into a time slot in cell i only if, after admission, the following conditions can be satisfied in this time slot:
Load _UL ( i )≦ CAC _— Target _— Load _— Thres _— UL Equation 7
and
Load _UL ( j )≦ CAC _— Neighbor _— Load _— Thres _— UL, ∀jεTier One(i)∪Tier Two(i) Equation 8
Where CAC_Target_Load_Thres_UL is the admission threshold of load in the target cell, and CAC_Neighbor_Load_Thres_UL is the admission threshold of load in the neighboring cells.

A measure of the quality after admission is the average load, {overscore (Load _UL )}, which is defined as the average load of target cell and neighboring cells. It is given by: $\begin{array}{cc}\stackrel{\_}{{\mathrm{Load}}_{\mathrm{UL}}}=\frac{1}{N}\sum _{j=1}^N{\mathrm{Load}}_{\mathrm{UL}}\left(j\right)& \mathrm{Equation}9\end{array}$

An alternative measure is the weighted average load in the uplink, {overscore (Load _{UL — Weighted} )}, which is similar to Equation 9, but gives priority to the target cell load by using a weight factor w (w>1) for the target cell. It is given by: $\begin{array}{cc}\stackrel{\_}{{\mathrm{Load}}_{\mathrm{UL\_Weighted}}}=\frac{1}{N}\left(\sum _{j=1j\ne 1}^N{\mathrm{Load}}_{\mathrm{UL}}\left(j\right)+w·{\mathrm{Load}}_{\mathrm{UL}}\left(i\right)\right)& \mathrm{Equation}10\end{array}$

Slot Selection Procedure for Call Admission Control in the Uplink

Suppose that the new CCTrCH seeking admission has M codes in its code set to be assigned. These M codes are arranged in the order of increasing spreading factor, (decreasing required SIR target). The slot selection follows the procedures below:

• • 1. Start with the first code in the code set (step 12 —FIG. 1).
  • 2. For each uplink time slot, compute the load of target cell and neighboring cells in this time slot using Equation 1 as if this code is assigned into this time slot (step 14 ).
  • 3. For each uplink time slot, judge if this code can be assigned to this time slot by checking if load of target cell and neighboring cells in this time slot after assignment satisfies conditions in Equations 7 and 8 (steps 16 , 17 ).
    • a. If yes, this time slot will be considered as possible time slot for assignment in step 22 (step 18 ).
    • b. Otherwise, this time slot will not be considered as possible time slot for assignment in step 22 (step 20 ).
  • 4. Check if there are any possible time slots for assignment (step 22 ).
    • a. If there is at least one possible time slot for assignment, among all possible time slots, select the time slot that yields the lowest average load as defined in Equation 9 or the lowest weighted average load as defined in Equation 10 (step 24 ). This code will be assigned to this selected time slot. Go to step 28 .
    • b. Otherwise, this code cannot be assigned in the target cell (step 26 ). It also means the CCTrCH with this code set cannot be accommodated in the target cell. The slot selection procedure ends (step 33 ).
  • 5. Check if there are any more codes to be assigned (step 28 ).
    • a. If yes, select the next code in the code set (step 30 ) and go to step 14 .
  • b. Otherwise, the assignment of the CCTrCH is done (step 32 ). The slot selection procedure ends (step 33 ).

Basic Assignment Procedure for Call Admission Control in the Downlink

Let CaPwr(i) denote the carrier power of a downlink time slot in the cell i. Let SIR _t denote the required SIR target of the code to be assigned in this time slot in the target cell i. Let PL(k) denote the pathloss of this UE between BS of cell k, k=1, 2, . . . , N. The code TX power for this new code, denoted by TX _code , is given by:
TX _code =SIR _t ·PL ( i )· I _total Equation 11
and I _total is given by: $\begin{array}{cc}{I}_{\mathrm{total}}={\alpha }_{\mathrm{DL}}·\mathrm{CaPwr}\left(i\right)/\mathrm{PL}\left(i\right)+\sum _{j\in \mathrm{Tier}-\mathrm{One}\left(i\right)}\mathrm{CaPwr}\left(j\right)/\mathrm{PL}\left(j\right)+\sum _{j\in \mathrm{Tier}-\mathrm{Two}\left(j\right)}\mathrm{CaPwr}\left(j\right)/\mathrm{PL}\left(j\right)+N_0& \mathrm{Equation}12\end{array}$

When the system is at moderate or high load, (where call admission control is really put in use), background noise N ₀ is negligible. Therefore, Equation 11 is rewritten as: $\begin{array}{cc}{\mathrm{TX}}_{\mathrm{code}}={\mathrm{SIR}}_t·\left({\alpha }_{\mathrm{DL}}·\mathrm{CaPwr}\left(i\right)+\sum _{j\in \mathrm{TierOne}\left(i\right)}\mathrm{CaPwr}\left(j\right)·\mathrm{PL}\left(i\right)/\mathrm{PL}\left(j\right)+\sum _{j\in \mathrm{TierTwo}\left(j\right)}\mathrm{CaPwr}\left(j\right)·\mathrm{PL}\left(i\right)/\mathrm{PL}\left(j\right)\right)& \mathrm{Equation}13\end{array}$

If $X=\sum _{j\in \mathrm{TierOne}\left(i\right)}^{}\mathrm{CaPwr}\left(j\right)·\mathrm{PL}\left(i\right)/\mathrm{PL}\left(j\right),$
for UE at different locations, (which implies different pathloss to BS), X is a random variable. Define ω _DL as: $\begin{array}{cc}{\omega }_{\mathrm{DL}}=\underset{\min }{\arg }\left\{{\omega }_{\mathrm{DL}}:\mathrm{Pr}\left(X\le {\omega }_{\mathrm{DL}}·\sum _{j\in \mathrm{TierOne}\left(i\right)}^{}\mathrm{CaPwr}\left(j\right)\right)>\theta \right\}& \mathrm{Equation}14\end{array}$
Where θ is a predefined percentage, for example, 90%.

Similarly, if $Y=\sum _{j\in \mathrm{TierTwo}\left(i\right)}^{}\mathrm{CaPwr}\left(j\right)·\mathrm{PL}\left(i\right)/\mathrm{PL}\left(j\right),$
for UE at different locations, (which implies different pathloss to BS), Y is a random variable. Define ξ _DL as $\begin{array}{cc}{\xi }_{\mathrm{DL}}=\underset{\min }{\arg }\left\{{\xi }_{\mathrm{DL}}:\mathrm{Pr}\left(Y\le {\xi }_{\mathrm{DL}}·\sum _{j\in \mathrm{TierTwo}\left(i\right)}^{}\mathrm{CaPwr}\left(j\right)\right)>\theta \right\}& \mathrm{Equation}15\end{array}$
Then, Equation 13 can be written as: $\begin{array}{cc}{\mathrm{TX}}_{\mathrm{code}}={\mathrm{SIR}}_t·\left({\alpha }_{\mathrm{DL}}·\mathrm{CaPwr}\left(i\right)+{\omega }_{\mathrm{DL}}·\sum _{j\in \mathrm{Tier}-\mathrm{One}\left(i\right)}^{}\mathrm{CaPwr}\left(j\right)+{\xi }_{\mathrm{DL}}·\sum _{j\in \mathrm{Tier}-\mathrm{Two}\left(j\right)}^{}\mathrm{CaPwr}\left(j\right)\right)& \mathrm{Equation}16\end{array}$

After the new code is added, the sum of code transmit power of existing codes will increase by ΔTX(i). The value of ΔTX(i) is estimated to be: $\begin{array}{cc}\Delta \mathrm{TX}\left(i\right)=\frac{{\alpha }_{\mathrm{DL}}·{\mathrm{TX}}_{\mathrm{code}}}{\begin{array}{c}{\alpha }_{\mathrm{DL}}·\mathrm{CaPwr}\left(i\right)+{\omega }_{\mathrm{DL}}·\sum _{j\in \mathrm{Tier}-\mathrm{One}\left(i\right)}^{}\mathrm{CaPwr}\left(j\right)+\\ {\xi }_{\mathrm{DL}}·\sum _{j\in \mathrm{Tier}-\mathrm{Two}\left(i\right)}^{}\mathrm{CaPwr}\left(j\right)\end{array}}\mathrm{CaPwr}\left(i\right)& \mathrm{Equation}17\end{array}$
Therefore, the carrier power of cell i after admission is estimated to be:
CaPwr ( i )= CaPwr ( i )+ TX _code +ΔTX ( i )+ Margin _{target cell} Equation 18
Where Margin _{target cell} is the margin used for call admission control in the target cell.

The increase to the carrier power of cell j (if jεTier One(i)) after admission, ΔTX(j), is estimated to be: $\begin{array}{cc}\Delta \mathrm{TX}\left(j\right)=\frac{({\omega }_{\mathrm{DL}}·\left({\mathrm{TX}}_{\mathrm{code}}+\Delta \mathrm{TX}\left(i\right)\right)}{\begin{array}{c}{\alpha }_{\mathrm{DL}}·\mathrm{CaPwr}\left(j\right)+{\omega }_{\mathrm{DL}}·\sum _{j\in \mathrm{Tier}-\mathrm{One}\left(j\right)}^{}\mathrm{CaPwr}\left(k\right)+\\ {\xi }_{\mathrm{DL}}·\sum _{k\in \mathrm{Tier}-\mathrm{Two}\left(j\right)}^{}\mathrm{CaPwr}\left(k\right)\end{array}}C\mathrm{aPwr}\left(i\right)& \mathrm{Equation}19\end{array}$

The increase to the carrier power of cell j (if jεTier Two(i)) after admission, ΔTX(j), is estimated to be: $\begin{array}{cc}\Delta \mathrm{TX}\left(j\right)=\frac{{\xi }_{\mathrm{DL}}·\left({\mathrm{TX}}_{\mathrm{code}}+\Delta TX\left(i\right)\right)}{\begin{array}{c}{\alpha }_{\mathrm{DL}}·\mathrm{CaPwr}\left(j\right)+{\omega }_{\mathrm{DL}}·\\ \sum _{k\in \mathrm{Tier}-\mathrm{One}\left(j\right)}^{}\mathrm{CaPwr}\left(k\right)+{\xi }_{\mathrm{DL}}·\\ \sum _{k\in \mathrm{Tier}-\mathrm{Two}\left(j\right)}^{}\mathrm{CaPwr}\left(k\right)\end{array}}\mathrm{CaPwr}\left(j\right)& \mathrm{Equation}20\end{array}$
Therefore, the carrier power of cell j after admission is estimated to be:
CaPwr ( j )′= CaPwr ( j )+Δ TX ( j )+ Margin _{neighbor cell} Equation 21
Where Margin _{neighbor cell} is the margin used for call admission control in neighboring cells.

At call admission control, a code will be admitted into a time slot in cell i only if after admission the following conditions can be satisfied in this time slot:
CaPwr ( i )′≦ CaPwr _maximum Equation 22
and
CaPwr ( j )′≦ CaPwr _maximum , ∀jεTier-One(i)∪Tier-Two(i) Equation 23
Where CaPwr _maximum is the maximum allowed carrier power at Node-B.

A measure of the quality after admission is the average carrier power, {overscore (CaPwr)}, which is defined as the average carrier power of target cell and neighboring cells. It is given by $\begin{array}{cc}\stackrel{\_}{\mathrm{CaPwr}}=\frac{1}{N}\sum _{j=1}^N\mathrm{CaPwr}\left(j\right)& \mathrm{Equation}24\end{array}$

An alternative measure is the weighted average load in the uplink, {overscore (CaPwr _Weighted )}, which is similar to the definition in Equation 24, but gives priority to the target cell's carrier power by using a weight factor w (w>1) for the target cell. It is given by $\begin{array}{cc}\stackrel{\_}{{\mathrm{CaPwr}}_{\mathrm{Weighted}}}=\frac{1}{N}\left(\sum _{j=1,j\ne i}^N\mathrm{CaPwr}\left(j\right)+w·\mathrm{CaPwr}\left(i\right)\right)& \mathrm{Equation}25\end{array}$

Slot Selection Procedure for Call Admission Control in the Downlink

The flowchart of the slot selection procedure in the downlink is the same as in the uplink (shown in FIG. 1), except that call admission control in the downlink tries to minimize the average carrier power instead of average load. Suppose that the new CCTrCH seeking admission has M codes in its code set to be assigned. Since the direction is downlink, the M codes have the same spreading factors 16 or 1. Therefore, the order of assignment for codes does not matter in the downlink. The slot selection follows the procedures below:

• • 1. Start with the first code in the code set (step 12 ).
  • 2. For each downlink time slot, estimate the carrier power of target cell and neighboring cells in this time slot using Equations 16-21 as if this code is assigned into this time slot (step 14 ).
  • 3. For each downlink time slot, judge if this code can be assigned to this time slot by checking if carrier power of target cell and neighboring cells in this time slot after assignment satisfies conditions in Equations 22 and 23 (steps 16 m 17 ).
    • a. If yes, this time slot will be considered as a possible time slot for assignment in step 22 (step 20 ).
    • b. Otherwise, this time slot will not be considered as a possible time slot for assignment in step 22 (step 20 ).
  • 4. Check if there are any possible time slots for assignment (step 22 ).
    • a. If there is at least one possible time slot for assignment, among all possible time slots, select the time slot that yields the lowest average carrier power as defined in Equation 24 or the lowest weighted average carrier power as defined in Equation 25 (step 24 ). This code will be assigned to this selected time slot. Go to step 28 .
    • b. Otherwise, this code cannot be assigned in the target cell (step 26 ). It also means the CCTrCH with this code set cannot be accommodated in the target cell. The slot selection procedure ends (step 33 ).
  • 5. Check if there is any more code to be assigned (step 28 ).
    • a. If yes, select the next code in the code set (step 30 ), and go to step 14 .
    • b. Otherwise, the assignment of the CCTrCH is done (step 32 ). The slot selection procedure ends (step 33 ).

The second embodiment of the present invention is directed to call admission control in the absence of measurements.

Basic Assignment Procedure for Call Admission Control in the Uplink

The load of an uplink time slot in a cell is defined as the load generated by the codes assigned in the same time slot in this cell and in first tier cells and second tier cells (load generated from cells beyond second tier is negligible). Then, the load in a cell k is: $\begin{array}{cc}{\mathrm{Load}}_{\mathrm{UL}}\left(k\right)={\alpha }_{\mathrm{UL}}·\sum _{h\in \Omega \left(k\right)}{\mathrm{SIR}}_h+{\beta }_{\mathrm{UL}}·\sum _{h\in \mathrm{Tier}\mathrm{One}\left(k\right)}{\mathrm{SIR}}_h+{\sigma }_{\mathrm{UL}}·\sum _{h\in \mathrm{Tier}\mathrm{Two}\left(k\right)}{\mathrm{SIR}}_h,k=1,2,\dots ,N& \mathrm{Equation}26\end{array}$
Where α _UL is the average MUD residual factor in the uplink, β _UL is the weight factor for codes in the tier-one cells in the uplink, σ _UL is the weight factor for codes in the tier-two cells in the uplink, Ω(k) is the set of codes assigned in this time slot in the cell k, Tier One(k) is the set of codes assigned in this time slot in tier-one neighboring cells of the cell k, Tier Two(k) is the set of codes assigned in this time slot in tier-two neighboring cells of the cell k.

At call admission control, a code will be admitted into a time slot in cell i only if after admission the following conditions can be satisfied in this time slot:
Load _UL ( i )≦ CAC _— Target _— Load _— Thres _— UL Equation 27
and
Load _UL ( j )≦ CAC _— Neighbor _— Load _— Thres _— UL, ∀jεTier One(i)∪Tier Two(i) Equation 28
Where CAC_Target_Load_Thres_UL is the admission threshold of load in the target cell, and CAC_Neighbor_Load_Thres_UL is the admission threshold of load in the neighboring cells.

A measure of the quality after admission is the average load in the uplink, {overscore (Load _UL )}, which is defined as the average load of the target cell and neighboring cells. It is given by: $\begin{array}{cc}\stackrel{\_}{{\mathrm{Load}}_{\mathrm{UL}}}=\frac{1}{N}\sum _{j=1}^N{\mathrm{Load}}_{\mathrm{UL}}\left(j\right)& \mathrm{Equation}29\end{array}$

An alternative measure is the weighted average load in the uplink, {overscore (Load _{UL — Weighted} )}, which is similar to the definition in Equation 29, but gives priority to the target cell load by using a weight factor w (w>1) for the target cell. It is given by: $\begin{array}{cc}\stackrel{\_}{{\mathrm{Load}}_{\mathrm{UL\_Weighted}}}=\frac{1}{N}\left(\sum _{j=1,j\ne i}^N{\mathrm{Load}}_{\mathrm{UL}}\left(j\right)+w·{\mathrm{Load}}_{\mathrm{UL}}\left(i\right)\right)& \mathrm{Equation}30\end{array}$

Slot Selection Procedure for Call Admission Control in the Uplink

The flowchart of the slot selection procedure in the uplink is shown in FIG. 2. Suppose that the new CCTrCH seeking admission has M codes in its code set to be assigned. These M codes are arranged in the order of increasing spreading factor (decreasing required SIR target). The slot selection follows the procedures below:

• • 1. Start with the first code in the code set (step 12 ′).
  • 2. For each uplink time slot, compute the load of target cell and neighboring cells in this time slot using Equation 26 as if this code is assigned into this time slot (step 14 ′).
  • 3. For each uplink time slot, judge if this code can be assigned to this time slot by checking if load of target cell and neighboring cells in this time slot after assignment satisfies conditions in Equations 27 and 28 (steps 16 ′, 17 ′).
    • c. If yes, this time slot will be considered as possible time slot for assignment in step 22 ′ (step 18 ′).
    • d. Otherwise, this time slot will not be considered as possible time slot for assignment in step 22 ′ (step 20 ′).
  • 4. Check if there are any possible time slots for assignment (step 22 ′).
    • e. If there is at least one possible time slot for assignment, among all possible time slots, select the time slot that yields the lowest average load as defined in Equation 29 or the lowest weighted average load as defined in Equation 30 (step 24 ′). This code will be assigned to this selected time slot. Go to step 28 ′.
    • f. Otherwise, this code cannot be assigned in the target cell (step 26 ′). It also means the CCTrCH with this code set cannot be accommodated in the target cell. The slot selection procedure ends (step 33 ′).
  • 6. Check if there are any more code to be assigned (step 28 ′).
    • a. If yes, select the next code in the code set, and go to step 14 ′ (step 30 ′).
    • b. Otherwise, the assignment of the CCTrCH is done (step 32 ′). The slot selection procedure ends (step 33 ′).

Basic Assignment Procedure for Call Admission Control in the Downlink

The load of a downlink time slot in cell i is defined as the load generated by the codes assigned in the same time slot in this cell and in first tier cells and second tier cells (load generated from cells beyond second tier is negligible). Therefore, the load in the downlink is similar to the load in the uplink. However, there is a difference between them. In the uplink, there is only one receiver, the BS. In the downlink, there are several receivers, UEs, scattered in the cell. To compensate for this difference, a scale factor is added into the load calculation. Then, the load is given by: $\begin{array}{cc}{\mathrm{Load}}_{\mathrm{DL}}\left(k\right)=\mathrm{Scale}·\left({\alpha }_{\mathrm{DL}}·\sum _{h\in \Omega \left(k\right)}{\mathrm{SIR}}_h+{\beta }_{\mathrm{DL}}·\sum _{h\in \mathrm{Tier}\mathrm{One}\left(k\right)}{\mathrm{SIR}}_h+{\sigma }_{\mathrm{DL}}·\sum _{h\in \mathrm{Tier}\mathrm{Two}\left(k\right)}{\mathrm{SIR}}_h\right),k=1,2,\dots ,N& \mathrm{Equation}31\end{array}$
Where α _DL is the average MUD residual factor in the downlink, β _DL is the weight factor for codes in the tier-one cells in the downlink, σ _DL is the weight factor for codes in the tier-two cells in the downlink, Ω(k) is the set of codes assigned in this time slot in the cell k, Tier One(k) is the set of codes assigned in this time slot in tier-one neighboring cells of the cell k, Tier Two(k) is the set of codes assigned in this time slot in tier-two neighboring cells of the cell k.

At call admission control, a code will be admitted into a time slot in cell i only if after admission the following conditions can be satisfied in this time slot:
Load _DL ( i )≦ CAC _— Target _— Load _— Thres _— DL Equation 32
and
Load _DL ( j )≦ CAC _— Neighbor _— Load _— Thres _— DL, ∀jεTier-One(i)∪Tier-Two(i) Equation 33
Where CAC_Target_Load_Thres_DL is the admission threshold of load in the target cell, and CAC_Neighbor_Load_Thres_DL is the admission threshold of load in the neighboring cells.

A measure of the quality after admission is the average load in the downlink, {overscore (Load _DL )}, which is defined as the average load of target cell and neighboring cells. It is given by: $\begin{array}{cc}\stackrel{\_}{{\mathrm{Load}}_{\mathrm{DL}}}=\frac{1}{N}\sum _{j=1}^N{\mathrm{Load}}_{\mathrm{DL}}\left(j\right)& \mathrm{Equation}34\end{array}$

An alternative measure is the weighted average load in the uplink, {overscore (Load _{DL — Weighted} )}, which is similar to the definition in Equation 34, but gives priority to the target cell load by using a weight factor w (w>1) for the target cell. It is given by: $\begin{array}{cc}\stackrel{\_}{{\mathrm{Load}}_{\mathrm{DL\_Weighted}}}=\frac{1}{N}\left(\sum _{j=1,j\ne i}^N{\mathrm{Load}}_{\mathrm{DL}}\left(j\right)+w·{\mathrm{Load}}_{\mathrm{DL}}\left(i\right)\right)& \mathrm{Equation}35\end{array}$

Slot Selection Procedure for Call Admission Control in the Downlink

The flowchart of slot selection procedure is the same as in FIG. 2. Suppose that the new CCTrCH seeking admission has M codes in its code set to be assigned. Since the direction is downlink, the M codes have the same spreading factors 16 or 1. Therefore, the order of assignment for codes does not matter in the downlink. The slot selection follows the procedures below:

• • 1. Start with the first code in the code set (step 12 ′).
  • 2. For each downlink time slot, compute the load of target cell and neighboring cells in this time slot using Equation 31 as if this code is assigned into this time slot (step 14 ′).
  • 3. For each downlink time slot, judge if this code can be assigned to this time slot by checking if load of target cell and neighboring cells in this time slot after assignment satisfies conditions in Equations 32 and 33 (steps 16 ′, 17 ′).
    • a. If yes, this time slot will be considered as possible time slot for assignment in step 22 (step 18 ′).
    • b. Otherwise, this time slot will not be considered as possible time slot for assignment in step 22 (step 20 ′).
  • 4. Check if there are any possible time slots for assignment (step 22 ′).
    • a. If there is at least one possible time slot for assignment, among all possible time slots, select the time slot that yields the lowest average load as defined in Equation 34 or the lowest weighted average load as defined in Equation 35 (step 24 ′). This code will be assigned to this selected time slot. Go to step 28 ′.
    • b. Otherwise, this code cannot be assigned in the target cell (step 26 ′). It also means the CCTrCH with this code set cannot be accommodated in the target cell. The slot selection procedure ends (step 33 ′).
  • 5. Check if there are any more codes to be assigned (step 28 ′).
    • a. If yes, select the next code in the code set, and go to step 14 ′ (step 30 ′).
    • b. Otherwise, the assignment of the CCTrCH is done (step 32 ′). The slot selection procedure ends (step 33 ′).

The third embodiment of the present invention is directed to call admission control based on outage probabilities

Definition of Outage Probability for Call Admission Control in the Uplink

The load of an uplink time slot in a cell is defined as the load generated by the users assigned in the same time slot in this cell and in first tier cells and second tier cells (load generated from cells beyond second tier is negligible). In most technical literature, the load from neighboring cells is assumed to be a fixed ratio of the load from its own cell based on the assumption of homogeneous system. However, in a heterogeneous system, the load cannot be modeled in such a way. We compute the load from neighboring cells based its actual traffic. Then, the load in a cell k is given by: $\begin{array}{cc}{\mathrm{Load}}_{\mathrm{UL}}\left(k\right)={\alpha }_{\mathrm{UL}}·\sum _{h\in \Omega \left(k\right)}{\mathrm{SIR}}_h+{\beta }_{\mathrm{UL}}·\sum _{h\in \mathrm{Tier}\mathrm{One}\left(k\right)}{\mathrm{SIR}}_h+{\sigma }_{\mathrm{UL}}·\sum _{h\in \mathrm{Tier}\mathrm{Two}\left(k\right)}{\mathrm{SIR}}_h,k=1,2,\dots ,N& \mathrm{Equation}36\end{array}$
Where α _UL is the average MUD residual factor in the uplink, β _UL is the weight factor for users in the tier-one cells in the uplink, σ _UL is the weight factor for users in the tier-two cells in the uplink, Ω(k) is the set of users assigned in this time slot in the cell k, Tier One(k) is the set of users assigned in this time slot in tier-one neighboring cells of the cell k, Tier Two(k) is the set of users assigned in this time slot in tier-two neighboring cells of the cell k.

Since the load is defined based on noise rise, we have: $\begin{array}{cc}\mathrm{Noise}\mathrm{rise}=\frac{I_{\mathrm{total}}}{N_0}=\frac{1}{1-{\mathrm{Load}}_{\mathrm{UL}}\left(k\right)}& \mathrm{Equation}37\end{array}$

Because of the dynamic range limitation and for the purpose of power control stability, the noise rise at the BS should be limited a maximum value of NR _max . Then, we have: $\begin{array}{cc}\frac{I_{\mathrm{total}}}{N_0}\le {\mathrm{NR}}_{\max }& \mathrm{Equation}38\end{array}$
Therefore, Equation 38 can written as: $\begin{array}{cc}{\mathrm{Load}}_{\mathrm{UL}}\left(k\right)\le 1-\frac{1}{{\mathrm{NR}}_{\max }}\mathrm{or}& \mathrm{Equation}39\\ {\alpha }_{\mathrm{UL}}·\sum _{h\in \Omega \left(k\right)}{\mathrm{SIR}}_h+{\beta }_{\mathrm{UL}}·\sum _{h\in \mathrm{Tier}\mathrm{One}\left(k\right)}{\mathrm{SIR}}_h+{\sigma }_{\mathrm{UL}}·\sum _{h\in \mathrm{Tier}\mathrm{Two}\left(k\right)}{\mathrm{SIR}}_h\le 1-\frac{1}{{\mathrm{NR}}_{\max }}& \mathrm{Equation}40\end{array}$

The probability of outage in a TDD time slot i, denoted by P _out , is defined as the probability that inequality in Equation 40 does not hold. It is given by $\begin{array}{cc}{P}_{\mathrm{out}}=\mathrm{Pr}\left\{{\alpha }_{\mathrm{UL}}·\sum _{h\in \Omega \left(k\right)}{\mathrm{SIR}}_h+{\beta }_{\mathrm{UL}}·\sum _{h\in \mathrm{Tier}\mathrm{One}\left(k\right)}{\mathrm{SIR}}_h+{\sigma }_{\mathrm{UL}}·\sum _{h\in \mathrm{Tier}\mathrm{Two}\left(k\right)}{\mathrm{SIR}}_h>1-\frac{1}{{\mathrm{NR}}_{\max }}\right\}& \mathrm{Equation}41\end{array}$

Computation of Outage Probability

Because of fading and imperfect power control, the value of SIR _h is a random variable that follows a lognormal distribution. Therefore, SIR _h can be expressed as:
SIR _h =10 ^{N(μ h ,σ h 2} ) Equation 42
Using ψ to replace $1-\frac{1}{{\mathrm{NR}}_{\max }},$
Equation 41 can be written as: $\begin{array}{cc}{P}_{\mathrm{out}}=\mathrm{Pr}\left\{\sum _{h=1}^N{\mathrm{SIR}}_h·A_h>\psi \right\}& \mathrm{Equation}43\end{array}$
Where A _h is given by: $\begin{array}{cc}{A}_h=\{\begin{array}{cc}{\alpha }_{\mathrm{UL}}& h\in \Omega \left(k\right)\\ {\beta }_{\mathrm{UL}}& h\in \mathrm{TierOne}\left(k\right)\\ {\sigma }_{\mathrm{UL}}& h\in \mathrm{TierTwo}\left(k\right)\end{array}& \mathrm{Equation}44\end{array}$
Then, we have:
SIR _h ·A _h =10 ^{(μ h ,σ h 2} )·10 ^{logA h} =10 ^{N(μ h +logA h ,σ h 2} ) Equation 45
Let X _h denote SIR _h ·A _h , then X _h is still a lognormal random variable. Its mean μ _{X h} and variance σ _{X h} ² are given by:
μ _{X h} =10 ^{μ h +logA h} ·10 ^{ln10σ h 2/2} Equation 46
σ _{X h} ² =10 ^{2(μ h +logA h )} ·10 ^{ln10σ h 2} ·(10 ^{ln10σ h 2} −1) Equation 47
Equation 43 becomes: $\begin{array}{cc}{P}_{\mathrm{out}}=\mathrm{Pr}\left\{\sum _{h=1}^N{X}_h>\psi \right\}& \mathrm{Equation}48\end{array}$

Even though the distribution of X _h is known, the computation of P _out in Equation 48 is still very complex, and cannot be done in real time. At moderate or high system load, value of N in Equation 48 is large. Therefore, the Gaussian approximation will have both good approximation result and low computation complexity. Here, we choose the Gaussian approximation approach to allow the Radio Network Controller (RNC) to compute the outage probability and make a decision of resource allocation in real time.

Consider that we have a random variable $Y=\sum _{h=1}^N{X}_h,$
where {X _h } are N independent identical random variables, each with mean μ _{X h} , and variance σ _{X h} ² . Then: $\begin{array}{cc}{\mu }_Y=\sum _{h=1}^N{\mu }_{X_h}& \mathrm{Equation}49\\ {\sigma }_Y^2=\sum _{h=1}^N{\sigma }_{X_h}^2\text{and:}& \mathrm{Equation}50\\ P_{\mathrm{out}}=\mathrm{Pr}\left\{Y>\psi \right\}=Q\left(\frac{\psi -{\mu }_Y}{{\sigma }_Y}\right)& \mathrm{Equation}51\end{array}$
Let P _out (i) denote the outage probability of time slot i. If a user is allocated to use L slots (l=1, 2, . . . , L), the total outage probability of the allocation, denoted by P _{out — total} , is defined as the probability that outage occurs in at least one time slot. It is given by: $\begin{array}{cc}{P}_{\mathrm{out\_total}}=1-\prod _{l=1}^L\left(1-P_{\mathrm{out}}\left(l\right)\right)& \mathrm{Equation}52\end{array}$

Slot Selection Procedure for Call Admission Control in the Uplink

The call admission control function will try to minimize the total outage probability of the CCTrCH while making sure that the outage probability of assigned timeslots in neighboring cells also meets the requirements. The flowchart of the call admission control algorithm is shown in FIG. 3.

Suppose that the new CCTrCH seeking admission in the target cell k has M codes in its code set to be assigned. These M codes are arranged in the order of increasing spreading factor (decreasing required SIR target). The slot selection follows the procedures below:

• • 1. Start with the first code in the code set, (step 36 ).
  • 2. Compute the current outage probability of each time slot in the target cell (step 38 ). Also compute the outage probability of each time slot in the neighboring cells as if this code is assigned into the time slot (step 38 ).
    • a. If the outage probability of all neighboring cells are less than the maximum allowed outage probability, say τr, then this time slot can be considered for assignment.
    • b. Otherwise, this time slot cannot be considered for assignment.
  • 3. Among possible time slots for assignment, start with time slot with the lowest outage probability, say time slot i (step 40 ).
  • 4. Assign the code into the time slot i and compute the updated outage probability of the time slot (step 42 ).
  • 5. Check if there are still more codes not assigned for the user (step 44 ).
    • a. If no, all codes are already assigned. Go to step 46 .
    • b. Otherwise, continue to step 52 to assign the next code in the code set.
  • 6. Branching at step 44 B, compute the outage probability of each time slot in the neighboring cells as if this code is assigned into the time slot (step 52 ). Check if time slot i is still among those possible time slots (step 54 ).
    • a. If no, find the time slot with the lowest outage probability among those possible time slots, say slot j. Set i=j (step 56 ), and go to step 42 .
    • b. Otherwise, check if the outage probability of time slot i is still the lowest among those possible time slots (step 58 ).
      • i. If yes, go to step 42 .
      • ii. Otherwise, compute if it is worthy to assign the next code into the time slot with the lowest outage probability, say slot j (step 60 ). This is done by comparing the contribution to the total outage probability by those codes already assigned to time slot i and this code. The contribution to total outage probability if this code is put into slot j, denoted by P _contribution is given by:
        P _contribution =1−(1 −P _out ( i ))·(1 −P _out ( j ))
      • The contribution to total outage probability if this RU is still assigned into slot i, denoted by P′ _contribution , is same as the outage probability in slot i. That is, P′ _contribution =P _out (i)′. Check if P _contribution ≧P′ _contribution ( 62 ).
        • 1. If no, go to step 40 .
        • 2. Otherwise, set i=j (step 64 ), and go to step 42 .
  • 7. Compute the total outage probability of the allocation (step 46 ), P _{out — total} , as in Equation 52. Check if P _{out — total} ≦θ (step 48 ).
    • a. If yes, the user will be admitted (step 50 ).
    • b. Otherwise, the user will be rejected (step 51 ).

Call Admission Control in the Downlink

The call admission control function in the downlink is similar to uplink. However, there are some differences in load definition and its physical meaning. In the uplink, there is only one receiver, the BS. In the downlink, there are several receivers, UEs, scattered in the cell. To compensate this difference, a scale factor is added into the load calculation. Then, the load is given by: $\begin{array}{cc}{\mathrm{Load}}_{\mathrm{DL}}\left(k\right)=\mathrm{Scale}·\left({\alpha }_{\mathrm{DL}}·\sum _{h\in \Omega \left(k\right)}{\mathrm{SIR}}_h+{\beta }_{\mathrm{DL}}·\sum _{h\in \mathrm{Tier}\mathrm{One}\left(k\right)}{\mathrm{SIR}}_h+{\sigma }_{\mathrm{DL}}·\sum _{h\in \mathrm{Tier}\mathrm{Two}\left(k\right)}{\mathrm{SIR}}_h\right),k=1,2,\dots ,N& \mathrm{Equation}53\end{array}$

In the uplink, the load is defined based on total noise rise at the BS, the common receiver. In the downlink, multiple receivers are scattered in the cell. Therefore, the downlink load is defined based on average downlink noise rise, we have: $\begin{array}{cc}\mathrm{Noise}\mathrm{rise}=\frac{\stackrel{\_}{I_{\mathrm{DL}}}}{N_0}=\frac{1}{1-{\mathrm{Load}}_{\mathrm{DL}}\left(k\right)}& \mathrm{Equation}54\end{array}$

Other than the difference in load definition and physical meaning, outage probability computation and slot selection in the downlink are the same as in the uplink as shown in FIG. 3.