Page 75 - Kaleidoscope Academic Conference Proceedings 2021

Connecting physical and virtual worlds

Table 2 – Simulation parameters

Parameter Value
Scenario Indoor HotSpot (12 Node 50m × 120m)
(HP)Traffic 1: Bit rate = 1.8Mbps, packet delay budget = 5ms
Traffic (LP)Traffic 2: Bit rate = 30Mbps, packet delay budget = 15ms
Option1: SU-MIMO priority-based adaptive preemptive scheduler
Scheduler
Option2: SU-MIMO proportion fair scheduler
Inter-BS distance 20m
Simulation bandwidth 100MHz
Subcarrier spacing 30kHz
TDD pattern DDDSU
32Tx: (M, N, P, M g, N g; M p, N p) = (4, 4, 2, 1, 1; 4, 4)
BS antenna configuration The antenna tilt is 90 degrees.
Channel estimation realist
Target BLER 10% for first transmission
PDCCH overhead 2 symbols per 14 symbols
DMRS overhead 1 symbol per 14 symbol
BS receiver MMSE-IRC

indication of DCI can be useful in the following calculated as
aspects. In case the generation/detection is done
j
after the Physical Downlink Shared Channel (PDSCH) T j = r ∗ N RB ∗ v, (7)
i,ins i
reception, this indication can be useful in the hybrid
j
automatic repeat request combination of the follow-up where r denotes the Mutual Information Rate (MIR)
i
retransmission. In case the generation/detection is done of the i-th user in the j-th sub-band, N RB denotes the
prior to the PDSCH reception, this information can be number of resource blocks in a given sub-band, and v
j
directly used to aid its decoding. denotes the number of layers. r can be obtained using
i
j
Following this principle, a priority-based adaptive reported SINR s via a mapping function [15, 16]. In
i
preemptive scheduler is proposed to guarantee the all, the MIR of i-th user across the total N SB sub-bands
j
performance of multi-streams XR or XR in concurrence can be denoted as r i , {r }, j = 1, 2, 3, · · · , N SB and
i
with uRLLC or other services. Depending on the reported CQI of the i-th user can be expressed as s i ,
j
service required, the user in a given network can be {s }, j = 1, 2, 3, · · · , N SB .
i
differentiated as a High Priority (HP) or Low Priority The metric for determining which UE a sub-band should
(LP) UE. A conventional Proportional Fairness (PF) get assigned to is proposed as follows,
scheduler is transparent to this priority information in
j
compiling the multi-users resource allocation results. j T i,ins
The frequency domain resource allocation is done at the M = T j ∗ F(p(i)), (8)
i
sub-band level taking into account the Channel Quality i,avr
Indicator (CQI) as well as the Transport Block Size
where the function F(x) is
(TBS) at the scheduling instant. Consider the case
where priority of services can be known in advance, (
10, x = 1
the frequency domain resource allocation can take into F(x) = (9)
account the relative priority of different packets in the 1, x = 0
resource allocation by assigning a PF priority coefficient
in the sub-band allocation phase. Specifically, let p ∈ The scaling operation e.g. 10 times herein, to the
L×1 metric can be fine-tuned according to the buffer
N denotes the priority of different users. And the
status ratio of high priority and low priority UE. The
priority vector p can be expressed as j
average throughput T can be calculated considering
i,aver
different weights for the long-term average data rate and
(
0, the i-th UE has low priority
p(i) = (6) instantaneous data rate as follows:
1, the i-th UE has high priority
j
j
T i,aver = αT i,ins + (1 − α) ∗ T last , (10)
where i = 1, 2, 3, . . . , L and L denotes the total number
of users in the system. Like a proportional fairness where T last is the average throughput till the latest
scheduler, the priority-based adaptive preemptive scheduled instant and α is originally set related to the
scheduler requires the knowledge of instantaneous filtering duration of the average data rate and can be
throughput T j , average throughput T j to derive
i,ins i,aver adjusted according to the proportion of the high priority
j j
metric M for user i at sub-band j. T can be within the system. In our simulation, α is set sufficiently
i i,ins
– 13 –