4.4 DCC Design
4.4.5 Call Admission Controller
In FQ-DCC scheme, the output action represents the estimated level of the pilot power PbbI. Since the maximum link power is the function of the pilot power in (4.38), as soon as the optimal pilot power has been calculated, the corresponding maximum link power bepb can be updated by
be
pb[dBm] = bPbI[dBm] + HR(r∗)[dB] − HRI[dB], (4.39) where bHRI is the estimated receiver sensitivity of the pilot signal received at cell boundary with referenced service rate r∗. Moreover, from (4.36) and (4.39), estimated call admission threshold Λb of signal-to-interference-ratio (SIR) for cell b becomes
Λb[dB] = bHRI[dB] − (ΩI+ ηo)[dBm],
= bPbI[dBm] − bepb[dBm] + HR(r∗)[dBm] − (ΩI+ ηo)[dBm]. (4.40)
• New Call Admission Control Algorithm: A new call user m measures and re-ports the received signal Eb/Io, γb,m(r), to base station b to request traffic channel for transmitting service rate r. The received SIR of the mobile station, bHS[dB], can thus be estimated as: γb,m(r)[dB]−GP(r)[dB]. The base station will accept the request with HbS[dB] being larger than Λb[dB], otherwise the new call will be blocked.
• Handoff Call Admission Control Algorithm: The soft handoff algorithm is im-plemented based on [51]. The MRC method (4.4) is used to combine received Eb/Io, γh(r), from all serving base stations in the active set Dh. Thus, the mobile station h’s received SIR, bHS, can thus be estimated as: γh(r)[dB]−GP(r)[dB]. The handoff request will be issued to base station b whenever an add event occurs. The base sta-tion will accept a user with bHS[dB] being larger than Λb[dB], otherwise the handoff call request will be blocked. The admitted handoff call will add new member of the base station into active set Dh. On the other hand, if the blocked handoff call doesn’t exceed handoff delay time, it can make handoff request again as long as its Ec/Io does not fall off the Ec/Io requirement, Υ[dB].
4.5 Simulation Results and Discussions
4.5.1 Simulation Model
Consider a hexagon cellular system with 19 wrap-around cells, in which the central cell is a hotspot cell with high traffic load. Define the load ratio ρ as the ratio between the arrival rate of the hotspot cell and that of each other cell. Geographically, the cellular deployment is homogenous, and the default cell radii can be determined by the link budget design in subsection 4.3.3. Assume mobile stations are uniformly distributed in each cell, and the initials speeds of mobile stations are uniformly distributed. The maximum speeds for users in the hotspot cell, 1st-tier cells, and 2nd-tier cells are assumed to be 30, 60, and 60 km/hr, respectively. Whenever one mobile station moves into different cell tiers, the speed will be changed. Moreover, the probability of moving direction change for mobile stations is 0.2 and the direction update is among ±45 degrees [51]. During the mobility, the correlated shadowing effect is based on Gudmundson model [51], [52], in which the decorrelation length is 20 meters in a vehicular environment. Also, for the channel model [51], the pathloss is obtained by
40 × (1 − 0.004hb) × log10(d) − 18 × log10(hb) + 21 × log10(fd) + 80, (4.41)
where d is the distance between a base station and a mobile station; hb and fdare the antenna height of the base station and the downlink frequency, respectively. In our simulations, the downlink frequency is 2.4 Ghz and the antenna height is 20 meters. The link budget design of this simulation is provided in Table 4.1 [4], [53].
Table 4.1: Link budget in the multimedia WCDMA system System Parameters
Chip rate [Mchips] 3.84
Orthogonality Factor, fα 0.5
Referenced service rate r = r∗ (kbps) 144
Processing Gain of service rate r, GP(r) [dB] 14.26 Required Eb/No of service rate r, γ∗(r) [dB] 3 Required SIR of service rate r, HS(r) [dB] -11.26 Receiver Sensitivity of service rate r, HR(r) [dBm] -105.35 Transmitter (Base station)
Maximum transmission power of base station b, ePb [dBm] 43.01 Maximum link power constraint, epb [dBm] 30
Cable loss of the base station, Gc [dB] 3
Antenna gain of the base station, GB [dBi] 2
EIRP, EP [dBm] 29
Receiver (Mobile station)
Antenna gain of the mobile station, GM [dBi] 2.0 Body loss of the mobile station, LD [dB] 3.0
Soft handoff Gain, GS [dB] 3.0
Thermal noise density [dBm/Hz] -173.93
Noise figure [dB] 9.0
Noise power [dBm] -99.09
Margin of interference [dB] 5.0
Margin of log-normal fade [dB] 5.0
Total EIRP, ET [dBm] 31.0
Maximum allowable pathloss of service rate r, P L(r) [dB] 131.35 Pilot channel
Processing Gain of the pilot signal, GIP [dB] 0.28
Required Ec/Io [dB] -20
Receiver Sensitivity of the pilot signal, HRI [dBm] -114.37 Maximum allowable pathloss of pilot signal, P L(r) [dB] 144.35
Minimum allowable Ec/Io, Υ [dB] -19.28
We assume that the call arrival process is Poisson. There are three service classes in the system, real-time voice, data, and non-real-time data services. The call holding times of voice and data traffic are exponentially distributed with means 60 and 30 seconds, respectively.
For the real-time services, a two-level Markov modulated Poisson process (MMPP) is used to model voice traffic while a 5-level MMPP is used to model the data traffic. The mean duration of each state in the 5-level MMPP is 1 second. The transmission rate (required bit-energy-to-noise ratio) of the voice traffic is 12.2 kbps (5 dB), and the service rates (required
bit-energy-to-noise ratio) of the data traffic are 16, 32, 64, and 144 kbps (5, 4, 3, and 2 dB). Note that adaptive rate transmission is applied whenever the power resources are not enough to support the existing users. For the non-real-time services, the variable length data bursts are assumed to be geometrically distributed with mean data burst size of 200 frames.
Moreover, there are 6 different service rates (required bit-energy-to-noise ratio), 16, 32, 64, 144, 384, and 512 kbps (5, 4, 3, 2, 1.5, and 1 dB). The transmission is on burst-by-burst basis. That is, each burst should request for a traffic channel and release the channel as soon as it completes the burst transmission. In simulations, the traffic percentages of voice, real-time data, and non-real-time data are 60%, 35%, and 5%, respectively.
Because of different processing gain and required signal quality, different service classes have different service coverage. Apparently, higher rate services can be supported in smaller cell coverage, as shown in Fig. 4.5. In simulations, we consider seven service rates such that r0 < r1 < · · · < r6. Based on the link budget, the cell radius is calculated in terms of different referenced service rates, represented as ref = {0, 1, · · · , 6}. Consider the fixed pilot power design, 1 watt power is allocated to the pilot channel. Fig. 4.8 shows the capacity results by applying SSDT and LPPA schemes in terms of different referenced service coverage under uniform (ρ = 1) and non-uniform (ρ = 4) cell load cases. Because of less propagation loss of small cell coverage, smaller cell can provide better signal quality for mobile stations.
In all cases, the system capacity can be increased for smaller cell coverage. We can also see that the slope of throughput increments to the difference of referenced service coverage becomes flatter when reference service rate is larger than 4. This means the system reaches the capacity-limited situation. Hence, in this chapter, cell radius is referred to service rate 144 kbps, which is the maximum service rate of real-time data service.