Analysis of Blocking Probability for different Coverage Conditions

  • In this notebook, we compute the blocking probability as a function of AL probability distribution of UEs.

  • One can find how ALs are distributed in a CORESET by performing PDCCH link level performance with different ALs or system level simulations to determine SINR distribution of UEs.

  • In fact, suitable ALs are used for UEs to meet suitable PDCCH performance requirements. ____________________________________________________________________________________________________________________

  • We consider 3 scenarios corresponding to Good, Medium and Poor coverage for evaluation.

    • Good Coverage : Most of the UEs are in good coverage and require low ALs (i.e., ALs 1 and 2), with AL distribution [0.5, 0.4, 0.07, 0.02, 0.01]

    • Medium Coverage : Most of the UEs are in medium coverage and require medium ALs (i.e., AL 4), with AL distribution [0.05, 0.2, 0.5, 0.2, 0.05]

    • Extreme Coverage : Most of the UEs are in poor coverage and require high ALs (i.e., ALs 8 and 16), with AL distribution [0.01, 0.02, 0.07, 0.4, 0.5]

  • The CORESET size Nccep is 54 CCEs for this simulation.

Import Python Libraries

[1]:

import os
os.environ["CUDA_VISIBLE_DEVICES"] = "-1"
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'

# %matplotlib widget
import matplotlib.pyplot  as plt
import matplotlib.patches as mpatches
import matplotlib as mpl

import numpy      as np

Import 5G Toolkit Libraries

[2]:

import sys
sys.path.append("../../")
from toolkit5G.Scheduler import PDCCHScheduler

Simulation Parameters

The following parameters are used for this simulation: - coresetID denotes the coreset ID. - slotNumber denotes the slot-number carrying the PDCCH. - searchSpaceType denotes the search space type. UE specific search space (USS) or Common search space (CSS). - nci denotes the variable corresponding to carrier aggregation. Current simulation does not assume carrier aggregation.

[3]:

mu                = np.random.randint(4)    # numerlogy for sub-carrier spacing
numSlotsPerFrame  = 2**mu * 10              # number of slots per radio frame
coresetID         = 1                       # coreset ID
slotNumber        = 0
searchSpaceType   = "USS"                   # search space type. UE specific search space
nci               = 0                       # variable corresponding to carrier aggregation

PDCCH Scheduling Parameters

Following parameters are crucial for PDCCH scheduling performance: - Nccep denotes coreset size or number of CCEs available for scheduling UEs. - strategy denotes the scheduling strategy. - numCandidates denotes number of PDCCH candidates per each Aggregation Level.

[4]:

Nccep             = 54
maxNumUEs         = 40
numUEs            = np.arange(0,maxNumUEs+1,4)
numUEs[0]         = 1
numIterations     = 1000
numCandidates     = np.array([6,6,4,2,1], dtype=int)
strategy          = "Conservative"
pdcchSchedulerObj = PDCCHScheduler(mu,slotNumber, coresetID, nci)

PDCCH Scheduling for Good Coverage Scenarios

  • The characteristics of good coverage is parameterized by aggLevelProbDistributionForGoodCoverage.

  • Simulation outcome (blockage probability) parameterized by probOfBlockingForGoodCoverage.

  • Scenario where the SNR/SINR is good enough, so that a lower AL (i.e., AL of 1 and 2) is choosen by the scheduler.

[5]:

################
# Good Coverage
################
aggLevelProbDistributionForGoodCoverage = np.array([0.5, 0.4, 0.07, 0.02, 0.01])
probOfBlockingForGoodCoverage = np.zeros(numUEs.shape)

for n in range(numUEs.size):
    print("Simulating (n,numUEs) : "+str(n)+", "+str(numUEs[n]))
    prob      = 0
    for i in range(numIterations):
        ueALdistribution  = np.random.multinomial(numUEs[n], aggLevelProbDistributionForGoodCoverage)
        rnti              = np.random.choice( np.arange(1,65519+1), size = (numUEs[n],), replace=False)
        count             = pdcchSchedulerObj(Nccep,searchSpaceType,ueALdistribution,numCandidates,rnti,strategy)[0]
        numBlockedUEs     = np.sum(count)
        prob              = prob + numBlockedUEs/numUEs[n]
    probOfBlockingForGoodCoverage[n] = prob/numIterations

Simulating (n,numUEs) : 0, 1
Simulating (n,numUEs) : 1, 4
Simulating (n,numUEs) : 2, 8
Simulating (n,numUEs) : 3, 12
Simulating (n,numUEs) : 4, 16
Simulating (n,numUEs) : 5, 20
Simulating (n,numUEs) : 6, 24
Simulating (n,numUEs) : 7, 28
Simulating (n,numUEs) : 8, 32
Simulating (n,numUEs) : 9, 36
Simulating (n,numUEs) : 10, 40

PDCCH Scheduling for Medium Coverage Scenarios

  • The characteristics of medium coverage is parameterized by aggLevelProbDistributionForMediumCoverage.

  • Simulation outcome (blockage probability) parameterized by probOfBlockingForMediumCoverage.

  • Scenario where the SNR/SINR is not good enough or medium, so that a medium AL (i.e., AL of 4) is choosen by the scheduler.

[6]:

##################
# Medium Coverage
##################
aggLevelProbDistributionForMediumCoverage = np.array([0.05, 0.2, 0.5, 0.2, 0.05])
probOfBlockingForMediumCoverage = np.zeros(numUEs.shape)

for n in range(numUEs.size):
    print("Simulating (n,numUEs) : "+str(n)+", "+str(numUEs[n]))
    prob      = 0
    for i in range(numIterations):
        ueALdistribution  = np.random.multinomial(numUEs[n], aggLevelProbDistributionForMediumCoverage)
        rnti              = np.random.choice( np.arange(1,65519+1), size = (numUEs[n],), replace=False)
        count             = pdcchSchedulerObj(Nccep,searchSpaceType,ueALdistribution,numCandidates,rnti,strategy)[0]
        numBlockedUEs     = np.sum(count)
        prob              = prob + numBlockedUEs/numUEs[n]
    probOfBlockingForMediumCoverage[n] = prob/numIterations

Simulating (n,numUEs) : 0, 1
Simulating (n,numUEs) : 1, 4
Simulating (n,numUEs) : 2, 8
Simulating (n,numUEs) : 3, 12
Simulating (n,numUEs) : 4, 16
Simulating (n,numUEs) : 5, 20
Simulating (n,numUEs) : 6, 24
Simulating (n,numUEs) : 7, 28
Simulating (n,numUEs) : 8, 32
Simulating (n,numUEs) : 9, 36
Simulating (n,numUEs) : 10, 40

PDCCH Scheduling for Extreme Coverage Scenarios

  • The characteristics of extreme coverage is parameterized by aggLevelProbDistributionForExtremeCoverage.

  • Simulation outcome (blockage probability) parameterized by probOfBlockingForExtremeCoverage.

  • Scenario where the SNR/SINR is extreme or poor, so that a higher AL (i.e., AL of 8 and 16) is choosen by the scheduler.

[7]:

################
# Extreme Coverage
################
aggLevelProbDistributionForExtremeCoverage = np.array([0.01, 0.02, 0.07, 0.4, 0.5])
probOfBlockingForExtremeCoverage = np.zeros(numUEs.shape)

for n in range(numUEs.size):
    print("Simulating (n,numUEs) : "+str(n)+", "+str(numUEs[n]))
    prob      = 0
    for i in range(numIterations):
        ueALdistribution  = np.random.multinomial(numUEs[n], aggLevelProbDistributionForExtremeCoverage)
        rnti              = np.random.choice( np.arange(1,65519+1), size = (numUEs[n],), replace=False)
        count             = pdcchSchedulerObj(Nccep,searchSpaceType,ueALdistribution,numCandidates,rnti,strategy)[0]
        numBlockedUEs     = np.sum(count)
        prob              = prob + numBlockedUEs/numUEs[n]
    probOfBlockingForExtremeCoverage[n] = prob/numIterations

Simulating (n,numUEs) : 0, 1
Simulating (n,numUEs) : 1, 4
Simulating (n,numUEs) : 2, 8
Simulating (n,numUEs) : 3, 12
Simulating (n,numUEs) : 4, 16
Simulating (n,numUEs) : 5, 20
Simulating (n,numUEs) : 6, 24
Simulating (n,numUEs) : 7, 28
Simulating (n,numUEs) : 8, 32
Simulating (n,numUEs) : 9, 36
Simulating (n,numUEs) : 10, 40

Plotting the results

The following results clearly demonstrates the improvement in blocking probability with improvement in coverage. - For 10 UEs, Medium coverage and Good coverage experiences ~25%, 70% improvement in performance in comparison to Extreme coverage. - For 20 UEs, Medium coverage and Good coverage experiences ~25%, 60% improvement in performance in comparison to Extreme coverage. - For 40 UEs, Medium coverage and Good coverage experiences ~40%, 45% improvement in performance in comparison to Extreme coverage.

Its the recreation of Fig. 8. Blocking probability for different AL distributions. from the paper referenced below.

[8]:

fig, ax = plt.subplots()
ax.plot(numUEs, probOfBlockingForGoodCoverage,    marker  = "*", linestyle = "solid", ms = 6, c = 'b', label = "Good Coverage")
ax.plot(numUEs, probOfBlockingForMediumCoverage,  marker  = "o", linestyle = "solid", ms = 6, c = 'y', label = "Medium Coverage")
ax.plot(numUEs, probOfBlockingForExtremeCoverage, marker  = "+", linestyle = "solid", ms = 6, c = 'k', label = "Extreme Coverage")

ax.legend()
ax.set_xlabel('Total Number of Scheduled UEs')
ax.set_ylabel('Blocking Probability')
ax.set_title('Blocking Probability for different AL distributions', fontsize=12)
ax.set_xticks(numUEs)
ax.grid()
plt.show()
../../../_images/api_Projects_Project2_Blocking_Probability_for_Different_AL_distributions_16_0.png

References

[1] Blocking Probability Analysis for 5G New Radio (NR) Physical Downlink Control Channel. Mohammad Mozaffari, Y.-P. Eric Wang, and Kittipong Kittichokechai

[ ]: