1
\$\begingroup\$

I am trying to get my head around the effects of interference on a wireless signal and its signal strength. I do not have an engineering background, but I do have a background in wireless sensor networking (from the MAC layer up.)

Imagine following scenarios:

  • (a) A transmitter TX that transmits packets to a receiver RX in a room. Let's assume a free-space, line-of-sight transmissions from TX to RX. The received signal strength (RSSI) of the packets received at the receiver is -100 dBm (I took a random strength here.)
  • (b) The same scenario as (a), but now there are interfering transmissions in the room, on the same frequency channel.

Let's say that the packet reception ratio (PRR) is 100% in scenario (a), as the receiver sensitivity is lower (i.e., -120 dBm in my imagined scenario) than the -100 dBm.

In scenario (b) I assume the PRR will be < 100%, let's say 70%. I am trying to fully understand what the effect is of interference on the packets at RX:

  • For packets not received at RX anymore (i.e., the 30% of the transmitted packets,) did the interference lower the RSSI too low < -120 dBm, or how does interference "screw" up the signal so it can't be demodulated anymore?
  • For the packets that are still correctly received (the 70% of the transmitted packets,) does interference lower the RSSI of these packets received at RX, will the average RSSI be lowered to a RSSI between -120 dBm and -100 dBm (so still above the sensitivity threshold,) or does interference not decrease the signal strength of our transmission?

When I search for research about the effects of interference on a transmission (e.g. among others [1]), the authors rather talk about how the SIR (signal-to-interference ratio) affects the PRR, but never mention how it affects the RSSI. I am wondering if the RSSI of a tranmission in the presence of interference iss affected, that is, is lower or not?

[1] https://www.mdpi.com/2079-9292/8/9/1045/pdf

\$\endgroup\$
1
  • \$\begingroup\$ "RSSI" is not actually defined. Some standards define such a measure involving interference powers, others one not, most simply don't define such a handwave-y thing, and then hardware vendors invent "magical formulas". Take 4G/5G for example: "receive signal strength and quality" is a whole set of different, well defined measures. None of them is called RSSI. \$\endgroup\$ Mar 7, 2022 at 12:27

3 Answers 3

0
\$\begingroup\$

The 'RSSI' relates to the power of the overall signal at the receiver. However, there are always other contributing factors such as thermal noise, which makes it difficult for the receiver to decode the desired signal (by eliminating the effects of noise). Mathematically, we model the thermal noise effects as additive white gaussian noise with a certain variance. The signal-to-noise ratio (SNR) determines the bit-error-rate (BER) and consequently the packet error rate (PER). The notion of the receiver sensitivity comes into play here. For a given BER or PER requirement, and a thermal noise at the receiver (which is chipset dependent), the receiver sensitivity simply imposes a minimum RSSI level that meets this aforementioned BER/PER requirement.

When interference is present, it adds to the overall effect of the 'noise' i.e. undesired part of the received signal (thermal noise + interference). This is when we consider the notion of the signal-to-interference-plus-noise-ratio (SINR). In most scenarios but not all, when interference is present, the thermal noise doesn't matter much, since the thermal noise power may be a lot lower than the interference power.

To answer your primary question, the interference itself does not affect the RSSI of the desired signal. However, it would certainly affect the RSSI of the overall received signal (desired+interference). The interfering signal merely makes it more difficult to decode the desired signal from the overall signal. That is why when we are talking about the PER or conversely the 'PRR' that you refer to, we should not examine the RSSI, but instead the SINR. Once again, the SINR directly relates to the BER and PER performance. For instance, if we are transmitting a BPSK signal, the BER is analytically given by the expression derived here.

\$\endgroup\$
0
\$\begingroup\$

So I am wondering if the RSSI of a transmission in the presence of interference gets affected, i.e. is lower or not?

RSSI isn't a magic bullet that tells you the signal strength of the radio signal you want to receive; it tells you generally how much signal strength there is in the spectral area of your wanted signal. This means that when interference comes along in the presence of your wanted signal, the RSSI will tend to rise thus indicating there is either a bigger wanted signal or, there is the same signal but more interference.

\$\endgroup\$
2
  • \$\begingroup\$ Thanks for your fast answer. It is helpful. An additional question: if one mentions the signal-to-interference (SIR) ratio, should this be interpreted as the RSSI of the signal over the RSSI of the interferer's signal, and both RSSI values measured separately? Because, in the light of what you say, how can this SIR otherwise be measured? \$\endgroup\$
    – rf_coding
    Mar 7, 2022 at 11:21
  • \$\begingroup\$ The bottom line is that RSSI indicates the band-limited combined signal and interference level. If you had separate measurements of RSSI and you are able to turn those readings into received dBm levels then yes, you could approximate signal-to-interference as a power ratio. However, there is no defined way in which RSSI is measured by a device so this is open to much speculation without details of the specific receiver device. \$\endgroup\$
    – Andy aka
    Mar 7, 2022 at 11:29
0
\$\begingroup\$

The reasons for packet loss have more to do with signal do noise ratio SNR and the RSSI is more of an indication of the receiver signal strength. You could imagine a scenario where even with a fairly high RSSI you could have an interfering signal of high strength causing packet losses. This would essentially be what jamming signal would aim to do, on purpose or by chance. Otherwise more natural environmental sources of noise, say from anything such as atmospheric electromagnetic noise appearing random but actually could be from other radio transmitters (like reflections), electric motors, or even solar flares etc will result in packet loss. This occurs at the receiver level where the actual detection is “misinterpreted” by the reception apparatus for the wrong “information”. That is, in digital terms, a 1 is misinterpreted as a 0 or vice versa – purely because the decoding mechanism in the receiver no longer retains enough information to discern what is noise and what is wanted signal. In analogue terms there is no misinterpreted per se, but rather an increasing “blending” of the wanted signal into the interfering noise as the SNR reduces, up to the point where the two are indistinguishable and all is just noise. Then there is the noise that is inherent in the actual electronic or optical devices that are of a quantum nature like shot noise, or from thermal noise etc. but the SNR implications remain the same for the packet losses experienced.

The RSSI is an indication (mostly an average) of the detected level of the wanted signal, but it will also include the level of noise in that quantity (that part of noise that enters the receiver channel), up to the point where signal and noise are essentially indistinguishable.

\$\endgroup\$
2
  • \$\begingroup\$ Thanks for your answer. I would like to jump to another question, which makes it extra confusing for me. Instead of the scenario with interference, assume a scenario with obstacles etc., so multi-path fading comes into play, leading to a lower PRR. In contrast to interference, I do find references online stating that multi-path fading does decrease signal strength (ie, out-of-phase waves of multi-path signals reduce the signal strength perceived at the receiver). So, multiple signals from multi-path propagation arriving at a receiver, is not the same as interference, nor has the same effect? \$\endgroup\$
    – rf_coding
    Mar 7, 2022 at 16:44
  • \$\begingroup\$ Actually it could be taken as intereference but it also depends on the receiver architecture and how it handles multipath interference. This question is however not related to the original question and would be best answered in another thread. have you checked if this has already been asked at another time ? \$\endgroup\$
    – citizen
    Mar 8, 2022 at 8:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.