Interference Helper gives negative trace value for Interference Power

122 views
Skip to first unread message

Yusuya

unread,
Oct 5, 2017, 1:23:54 AM10/5/17
to ns-3-users
Dear all,

I am printing out my interference helper outcome using

LogComponentEnable ("InterferenceHelper", LOG_LEVEL_ALL)  together with LOgCOmponentEnable("YansWifiPhy ", LOG_LEVEL_ALL)

Sometimes I see no interference at all denoted by "interference (W)= 0" , sometimes I see positive value such as interference(W) = 6.17e-13 .

However, what appears weird for me is that sometimes the interference(W) value is negative such as the following (look at the third interference(W) value is negative :
==============================output of interest======
receiving plcp payload
YansWifiPhy:StartReceivePacket(0x263cee0, 0x25d9480, OfdmRate6MbpsBW10MHz, 0, 0)
bandwidth(MHz)=10, signal(W)= 4.30941e-11, noise(W)=4.00287e-14, interference(W)=0, snr(linear)=1076.58
InterferenceHelper:CalculatePlcpHeaderPer(0x263cf00)
previous= +147683923.0ns, current=+147685258.0ns
previous= +147685258.0ns, current=+148315923.0ns
bandwidth(MHz)=10, signal(W)= 4.30941e-11, noise(W)=4.00287e-14, interference(W)=1.14475e-12, snr(linear)=36.373
Case 5a - previous is in the preamble and current is after payload start: mode=OfdmRate3MbpsBW10MHz, psr=1
snr(dB)=30.3205, per=0
receiving plcp payload
YansWifiPhy:StartReceivePacket(0x24fd8d0, 0x2938d80, OfdmRate6MbpsBW10MHz, 0, 0)
bandwidth(MHz)=10, signal(W)= 1.59013e-11, noise(W)=4.00287e-14, interference(W)=-3.23117e-27, snr(linear)=397.248
InterferenceHelper:CalculatePlcpHeaderPer(0x24fd8f0)
previous= +147684223.0ns, current=+147685558.0ns
previous= +147685558.0ns, current=+148316223.0ns
bandwidth(MHz)=10, signal(W)= 1.59013e-11, noise(W)=4.00287e-14, interference(W)=6.1731e-13, snr(linear)=24.1904
Case 5a - previous is in the preamble and current is after payload start: mode=OfdmRate3MbpsBW10MHz, psr=1
snr(dB)=25.9906, per=0
receiving plcp payload

=============================end of output =====================

What causes such negative value ?
When I check in the code  interference(W) value is from th variable "noiseinterferenceW" and in the code there is this expression
noiseInterferenceW += (*j).GetDelta ();

WHAt does this expression do ? Is this one cause the negative value ? Why is the interference power negative ?

Thank you.

































Rediet

unread,
Oct 9, 2017, 3:51:06 AM10/9/17
to ns-3-users
Hi,

It's because the InterferenceHelper module works by cumulating the received intereference levels; adding the packet's power and then scheduling a subtraction of the same amount at the end of the packet. See manual of Wifi module for further detail.

Rediet

Rediet

unread,
Oct 9, 2017, 5:52:08 AM10/9/17
to ns-3-users
Please ignore my previous message. Tom has logged an bug (#2791) to investigate on the issue.

Sebastien Deronne

unread,
Oct 17, 2017, 3:07:10 PM10/17/17
to ns-3-users
Can I get a script to reproduce?

pdbarnes

unread,
Oct 18, 2017, 11:07:10 PM10/18/17
to ns-3-users
Is this just roundoff from adding then subtracting each interference source?

Peter

Sebastien Deronne

unread,
Dec 23, 2017, 10:41:42 AM12/23/17
to ns-3-users
Yusuya, if I do not get a script to reproduce, I cannot help on this.

saumil shah

unread,
Jan 5, 2018, 9:31:44 AM1/5/18
to ns-3-users
Dear Sebastien,

I am also getting negative value for interference power and more of that i am getting negative value for SNR (linear).

Kindly check the below thread in which i have posted detail about the error i am facing and how to reproduce the error with all necessary information.

i am using ns3-bt module which i got from http://code.nsnam.org/laa/ns-3-lbt.

Thread : https://groups.google.com/forum/#!topic/ns-3-users/t6L35V_eZ8M

Best Regards
Saumil Shah

Sebastien Deronne

unread,
Jan 6, 2018, 10:40:57 AM1/6/18
to ns-3-users
Lbt is not an official repository, you need to check whether you can reproduce it as well in ns-3-dev, otherwise I suggest to close the bug. 

pdbarnes

unread,
Jan 6, 2018, 10:01:23 PM1/6/18
to ns-3-users
It’s a real bug; please don’t close the report.

There was a recent post or email that localized it pretty well; I don’t have access to it ATM but I’ll add it to this thread.

I’m fairly sure the issue is adding and subtracting reals, leading to a small negative value. More details and suggested patch when get back to a real computer.

Peter

pdbarnes

unread,
Jan 8, 2018, 7:26:07 PM1/8/18
to ns-3-users

pdbarnes

unread,
Feb 23, 2018, 5:38:17 PM2/23/18
to ns-3-users
Hello Yusuya and Rediet,

Sebastien recently pushed a fix to ns-3-dev.  Could you please try your problem again and report if it still persists?

We are tracking this now on this bug report:

Thanks,
Peter
Reply all
Reply to author
Forward
0 new messages