Channel Access Simulation Study for Bluetooth-Wi-Fi Coexistence in IEEE 802.11-23

September 2023
Slide 1
Bluetooth Wi-Fi Coexistence: Channel
Access Simulation Study
Date:
 2023-09-13
Ratnesh Kumbhkar, Intel
Authors:
Simulation setup
Bluetooth:
Low Energy (LE) 2M PHY
Gaming audio: 2 earbuds
1 earbud contains mic
1 earbud has inertial measurement
unit (IMU) sensor
Tx power: 14 dBm
ISO interval = 7.5ms
Max latency = 15ms (2×ISO)
Number of subevents, NSE = 2
1 main TX + 3 retry opportunities
Total 4 attempts
Number of channels = 80 (each channel
2MHz wide)
Total hopping BW = 160MHz
Bi-directional traffic
2
Bluetooth
Wi-Fi:
1 AP, 1STA
Tx power: 20 dBm at AP/STA
MCS11
Single stream
One directional traffic
Traffic load: Full buffer or 60% constant
bit rate, or, beacon-only
TXOP duration ≈ 5ms
Bandwidth = 80MHz
Transmission starts at time=3s
Channel model:
AWGN channel with a breakpoint at 5 m
pathloss = 40.05 + 20log10(f/2.4) + 20log10(min(d, b)) + (d > b) *
(35log10(d/b))
where f = 5.18 GHz, d = distance, b = breakpoint = 5 m
September 2023
Ratnesh Kumbhkar, Intel
Wi-Fi + Multiple Bluetooth
3
Wi-Fi
AP
Wi-Fi STA
Bluetooth
peripheral with
2 earbuds
Bluetooth
central
meters
meters
September 2023
Ratnesh Kumbhkar, Intel
Listen-Before-Talk (LBT) with CCA
trigger
4
September 2023
Ratnesh Kumbhkar, Intel
LBT parameters:
Trigger threshold = 3
When the threshold of 3 is
reached, 3 more is added to
CCA-busy for additional
hysteresis.
Trigger count is capped at 24
The above set of parameters
is referred to as 3+3/24 [1]
LBT with CCA trigger
5
Full buffer Wi-Fi
Observations
Considerably adverse impact on Bluetooth
latency with 50% spectrum overlap. Higher
overlap would further degrade the
performance
Wi-Fi transmission is not impacted, as
Bluetooth defers the transmission in Wi-Fi
occupied channels
Wi-Fi starts
at ~3s
Some packets
did not get
transmitted at
all resulting in
glitches
Significant impact
on Bluetooth
latency
Bluetooth transmissions
start with ~0.1s offset
Wi-Fi
throughput is
not impacted
Bluetooth
Wi-Fi
Wi-Fi latency
is low
September 2023
Ratnesh Kumbhkar, Intel
LBT with CCA trigger
6
Wi-Fi : 
60% duty-cycle, constant bit rate
Observations:
With low-duty-cycle Wi-Fi, Bluetooth cannot
always detect, and block Wi-Fi occupied
channels
CCA-busy count keeps oscillating  due to gaps in
Wi-Fi transmission
Bluetooth keeps capturing channels during gaps
in Wi-Fi transmission, resulting in higher latency
for Wi-Fi
Continuous
impact on Wi-Fi
throughput
Bluetooth
Wi-Fi
Increased
Wi-Fi
latency
Still resulting
in some
glitches
September 2023
Ratnesh Kumbhkar, Intel
Enhanced Detect and Avoid (eDAA)
7
eDAA parameters:
A separate 20MHz wideband
scanner is used for signal
detection
4 sweeps of Bluetooth
hopping BW (in this case
160MHz) in 500ms
Vacate a 20MHz band as
soon as the band is found
busy
September 2023
Ratnesh Kumbhkar, Intel
eDAA
8
Full buffer Wi-Fi
Observations
All Bluetooth links can detect and vacate
Wi-Fi occupied channels
Wi-Fi throughput is good
Bluetooth latency is not impacted
Bluetooth
performance
is good
Bluetooth
Wi-Fi
Wi-Fi throughput
is not significantly
impacted
September 2023
Ratnesh Kumbhkar, Intel
eDAA
9
Wi-Fi : 
60% duty-cycle, constant bit rate
Observations
Frequent channel restoration caused by
unreliable Wi-Fi signal detection in low-duty
cycle Wi-Fi scenarios
Results in collision with Wi-Fi and increase in the size
of contention window
Continuous interference from multiple Bluetooth
links is detrimental for Wi-Fi
Bluetooth
Wi-Fi
Continuous
interference to
Wi-Fi
Significant
increase in
Wi-Fi
latency
Bluetooth
performance
is good
Frequent
channel
restoration
September 2023
Ratnesh Kumbhkar, Intel
eDAA: An example with modified
restoration procedure
10
 
Bluetooth
Wi-Fi
Wi-Fi : 
60% duty-cycle, constant bit rate
Restoration of a 20MHz band is based on last
10 sweeps
This example is one of the potential solutions for
a more reliable restoration procedure
Observation: the restoration needs to be more
rigorous than what is currently proposed in
eDAA.
Bluetooth
performance
is good
Wi-Fi throughput
is not significantly
impacted
September 2023
Ratnesh Kumbhkar, Intel
Coexistence with Wi-Fi beacon-only traffic
Wi-Fi beacons have very low duty cycle with ~100ms
period in primary channel.
Parameters
Wi-Fi beacon-only traffic (300 us, 20 MHz wide) with 102.4ms
period
Bluetooth traffic – same as previous slides
Slide 11
Ratnesh Kumbhkar, Intel
September 2023
While eDAA results in higher number of collisions with beacons, there were no
occurrences where two or more consecutive beacons were not received by the
Wi-Fi (non-AP) STA.
Wi-Fi
AP
Wi-Fi
STA
Observations
LBT with CCA trigger
Unable to meet the Bluetooth latency requirement in the case of 50% spectrum
overlap. With growing overlap, the latency will likely become worse.
Still seen with newly suggested LBT  3+3/24
While the performance for full-buffer Wi-Fi is good, Wi-Fi latency is impacted for
low duty-cycle Wi-Fi
No impact on beacon-only transmission
eDAA
Better for Bluetooth performance
Significant throughput impact on Wi-Fi specifically when detection is not reliable
e.g., low duty-cycle Wi-Fi
Coexistence with Wi-Fi can be improved by a more robust restoration procedure.
This results in better Wi-Fi throughput and latency.
Slightly higher impact on beacon-only transmission, but may not be significant
enough causing loss of connection
12
September 2023
Ratnesh Kumbhkar, Intel
Next steps
Agreement on realistic simulation scenarios
Wi-Fi traffic load: Low duty cycle (5-10%), medium duty cycle (50-60%) and full
buffer
Bluetooth low energy traffic pattern: Based on realistic cases such as gaming audio
Network topologies: Capturing mutual impact
Agreement on performance metrics for Bluetooth and Wi-Fi
Throughput
Latency
Detection accuracy and speed
Energy efficiency
Others
Rethink Frequency hopping kernel
Impact of the latency associated with channel map update
Slide 13
Ratnesh Kumbhkar, Intel
September 2023
September 2023
Slide 14
References
[1]  Bluetooth SIG bt_sig_nb_with_lbt_qualcomm_september_2023_budapest_r1
[2]  IEEE 11-23-1279-coex-00, IEEE Berlin, July 2023
[3]  BRAN(21)109h004r2_EN_303_687_NB_Proposals_for_DAA_Optimisation
[4]  BRAN(21)109h007_NB_coexistence_with_WB_in_6_GHz
Ratnesh Kumbhkar, Intel
Slide Note

doc.: IEEE 802.11-11/1503r0

September 2023

Ratnesh Kumbhkar, INTEL

Page

Embed
Share

A detailed simulation study was conducted to analyze the coexistence of Bluetooth and Wi-Fi in the IEEE 802.11-23 standard. The study covered aspects such as channel access, transmission setup, Bluetooth-Wi-Fi interactions, impact on latency, throughput, and observations on performance under various scenarios like spectrum overlap and duty cycles. Key findings included insights on latency impact, channel interference, and throughput variations between the two technologies.


Uploaded on Sep 22, 2024 | 0 Views


Download Presentation

Please find below an Image/Link to download the presentation.

The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. Download presentation by click this link. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.

E N D

Presentation Transcript


  1. September 2023 doc.: IEEE 802.11-23/1503r0 Bluetooth Wi-Fi Coexistence: Channel Access Simulation Study Date: 2023-09-13 Authors: Name Ratnesh Kumbhkar Affiliations Intel Address Santa Clara, USA email ratnesh.kumbhkar@intel.com Submission Slide 1 Ratnesh Kumbhkar, Intel

  2. September 2023 doc.: IEEE 802.11-23/1503r0 Simulation setup Wi-Fi: Bluetooth: Low Energy (LE) 2M PHY Gaming audio: 2 earbuds 1 earbud contains mic 1 earbud has inertial measurement unit (IMU) sensor Tx power: 14 dBm ISO interval = 7.5ms Max latency = 15ms (2 ISO) Number of subevents, NSE = 2 1 main TX + 3 retry opportunities Total 4 attempts Number of channels = 80 (each channel 2MHz wide) Total hopping BW = 160MHz Bi-directional traffic 80MHz Wi-Fi 1 AP, 1STA Tx power: 20 dBm at AP/STA MCS11 Single stream One directional traffic Traffic load: Full buffer or 60% constant bit rate, or, beacon-only TXOP duration 5ms Bandwidth = 80MHz Transmission starts at time=3s Channel model: AWGN channel with a breakpoint at 5 m pathloss = 40.05 + 20log10(f/2.4) + 20log10(min(d, b)) + (d > b) * (35log10(d/b)) where f = 5.18 GHz, d = distance, b = breakpoint = 5 m Overlap Bluetooth 160MHz O int r a 5ms s s s s ntra 5 s P s s P s s tr in transmission u nt or P u nt or P onn t d iso ronous rou nt 55 ms Submission 2 Ratnesh Kumbhkar, Intel

  3. September 2023 doc.: IEEE 802.11-23/1503r0 Wi-Fi + Multiple Bluetooth Bluetooth central Wi-Fi AP meters Bluetooth peripheral with 2 earbuds Wi-Fi STA meters Submission 3 Ratnesh Kumbhkar, Intel

  4. September 2023 doc.: IEEE 802.11-23/1503r0 Listen-Before-Talk (LBT) with CCA trigger LBT parameters: Trigger threshold = 3 When the threshold of 3 is reached, 3 more is added to CCA-busy for additional hysteresis. Trigger count is capped at 24 The above set of parameters is referred to as 3+3/24 [1] Submission 4 Ratnesh Kumbhkar, Intel

  5. September 2023 doc.: IEEE 802.11-23/1503r0 LBT with CCA trigger i i Full buffer Wi-Fi O r a Observations Considerably adverse impact on Bluetooth latency with 50% spectrum overlap. Higher overlap would further degrade the performance Wi-Fi transmission is not impacted, as Bluetooth defers the transmission in Wi-Fi occupied channels Some packets did not get transmitted at all resulting in glitches Bluetooth Significant impact on Bluetooth latency Bluetooth transmissions start with ~0.1s offset Wi-Fi latency is low Wi-Fi throughput is not impacted Wi-Fi Wi-Fi starts at ~3s Submission 5 Ratnesh Kumbhkar, Intel

  6. September 2023 doc.: IEEE 802.11-23/1503r0 LBT with CCA trigger i i Wi-Fi : 60% duty-cycle, constant bit rate O r a Observations: With low-duty-cycle Wi-Fi, Bluetooth cannot always detect, and block Wi-Fi occupied channels CCA-busy count keeps oscillating due to gaps in Wi-Fi transmission Bluetooth keeps capturing channels during gaps in Wi-Fi transmission, resulting in higher latency for Wi-Fi Still resulting in some glitches Bluetooth Wi-Fi Continuous impact on Wi-Fi throughput Increased Wi-Fi latency Submission 6 Ratnesh Kumbhkar, Intel

  7. September 2023 doc.: IEEE 802.11-23/1503r0 Enhanced Detect and Avoid (eDAA) eDAA parameters: A separate 20MHz wideband scanner is used for signal detection 4 sweeps of Bluetooth hopping BW (in this case 160MHz) in 500ms Vacate a 20MHz band as soon as the band is found busy i i O r a Submission 7 Ratnesh Kumbhkar, Intel

  8. September 2023 doc.: IEEE 802.11-23/1503r0 eDAA i i Full buffer Wi-Fi O r a Observations All Bluetooth links can detect and vacate Wi-Fi occupied channels Wi-Fi throughput is good Bluetooth latency is not impacted Bluetooth Bluetooth performance is good Wi-Fi throughput is not significantly impacted Wi-Fi Submission 8 Ratnesh Kumbhkar, Intel

  9. September 2023 doc.: IEEE 802.11-23/1503r0 eDAA i i Wi-Fi : 60% duty-cycle, constant bit rate O r a Observations Frequent channel restoration caused by unreliable Wi-Fi signal detection in low-duty cycle Wi-Fi scenarios Bluetooth Results in collision with Wi-Fi and increase in the size of contention window Continuous interference from multiple Bluetooth links is detrimental for Wi-Fi Bluetooth performance is good Wi-Fi Continuous interference to Wi-Fi Significant increase in Wi-Fi latency Frequent channel restoration Submission 9 Ratnesh Kumbhkar, Intel

  10. September 2023 doc.: IEEE 802.11-23/1503r0 eDAA: An example with modified restoration procedure Wi-Fi : 60% duty-cycle, constant bit rate i i O r a Restoration of a 20MHz band is based on last 10 sweeps This example is one of the potential solutions for a more reliable restoration procedure Bluetooth Observation: the restoration needs to be more rigorous than what is currently proposed in eDAA. Bluetooth performance is good Wi-Fi Wi-Fi throughput is not significantly impacted Submission 10 Ratnesh Kumbhkar, Intel

  11. September 2023 doc.: IEEE 802.11-23/1503r0 Coexistence with Wi-Fi beacon-only traffic Wi-Fi AP Wi-Fi beacons have very low duty cycle with ~100ms period in primary channel. Parameters Wi-Fi beacon-only traffic (300 us, 20 MHz wide) with 102.4ms period Bluetooth traffic same as previous slides Wi-Fi STA 20MHz eDAA LBT with CCA trigger Wi-Fi Overlap % of beacons received 91.5% 99% Bluetooth 160MHz While eDAA results in higher number of collisions with beacons, there were no occurrences where two or more consecutive beacons were not received by the Wi-Fi (non-AP) STA. Submission Slide 11 Ratnesh Kumbhkar, Intel

  12. September 2023 doc.: IEEE 802.11-23/1503r0 Observations LBT with CCA trigger Unable to meet the Bluetooth latency requirement in the case of 50% spectrum overlap. With growing overlap, the latency will likely become worse. Still seen with newly suggested LBT 3+3/24 While the performance for full-buffer Wi-Fi is good, Wi-Fi latency is impacted for low duty-cycle Wi-Fi No impact on beacon-only transmission eDAA Better for Bluetooth performance Significant throughput impact on Wi-Fi specifically when detection is not reliable e.g., low duty-cycle Wi-Fi Coexistence with Wi-Fi can be improved by a more robust restoration procedure. This results in better Wi-Fi throughput and latency. Slightly higher impact on beacon-only transmission, but may not be significant enough causing loss of connection Submission 12 Ratnesh Kumbhkar, Intel

  13. September 2023 doc.: IEEE 802.11-23/1503r0 Next steps Agreement on realistic simulation scenarios Wi-Fi traffic load: Low duty cycle (5-10%), medium duty cycle (50-60%) and full buffer Bluetooth low energy traffic pattern: Based on realistic cases such as gaming audio Network topologies: Capturing mutual impact Agreement on performance metrics for Bluetooth and Wi-Fi Throughput Latency Detection accuracy and speed Energy efficiency Others Rethink Frequency hopping kernel Impact of the latency associated with channel map update Submission Slide 13 Ratnesh Kumbhkar, Intel

  14. September 2023 doc.: IEEE 802.11-23/1503r0 References [1] Bluetooth SIG bt_sig_nb_with_lbt_qualcomm_september_2023_budapest_r1 [2] IEEE 11-23-1279-coex-00, IEEE Berlin, July 2023 [3] BRAN(21)109h004r2_EN_303_687_NB_Proposals_for_DAA_Optimisation [4] BRAN(21)109h007_NB_coexistence_with_WB_in_6_GHz Submission Slide 14 Ratnesh Kumbhkar, Intel

Related


More Related Content

giItT1WQy@!-/#giItT1WQy@!-/#giItT1WQy@!-/#