Project
Wireless Site Survey and Interference Mitigation Project
Wireless RF site survey project for spectrum occupancy, EIRP, receiver margin, channel planning, interference mitigation, traffic validation, thresholds, and release criteria.
This project builds a wireless site-survey and interference-mitigation package for a shared RF environment. The goal is not only to make a radio link work during a walk test. The goal is to produce engineering evidence that the selected channel, antenna placement, receiver margin, coexistence controls, service tests, monitoring thresholds, and release decision can be defended after the site changes.
The project is written for an industrial wireless network, but the workflow also applies to campus telemetry, warehouse automation, utility monitoring, test facilities, temporary construction networks, maritime terminals, and other installations where radios share spectrum with other equipment.
Project Objective
Design and validate a wireless survey package for a production hall that uses wireless sensors and handheld maintenance terminals. The final deliverable should answer:
- Which service boundary and RF environment are being accepted?
- Which channels are usable based on measured spectrum occupancy, not only a vendor default?
- Does the selected installation satisfy EIRP, receiver sensitivity, interference margin, and measurement uncertainty checks?
- Which mitigation actions are required before release?
- Does traffic performance meet packet loss, latency, jitter, and update-rate requirements under realistic load?
- Which monitoring thresholds and change-control rules should operations inherit?
The deliverable is a site-survey report and release package. It should be detailed enough for a field engineer, controls engineer, network engineer, or maintenance engineer to repeat the survey and understand why the chosen configuration was accepted.
Baseline Scenario
Use the following scenario or replace it with site data.
| Parameter | Value |
|---|---|
| Site | industrial production hall |
| RF service | wireless sensor and maintenance-terminal network |
| Candidate band | 2.4\ \text{GHz} ISM band |
| Channel bandwidth | 2.0\ \text{MHz} equivalent receiver noise bandwidth |
| Required update interval | less than 1.0\ \text{s} for sensor values |
| Packet loss acceptance limit | less than 0.5\% during survey load |
| 95th-percentile one-way latency limit | less than 50\ \text{ms} |
| Peak-to-peak jitter limit | less than 20\ \text{ms} |
| Required guarded RF margin above sensitivity | at least 12\ \text{dB} |
| Transmitter output power | 18\ \text{dBm} |
| Feeder and connector loss | 1.0\ \text{dB} |
| Antenna gain | 6\ \text{dBi} |
| EIRP limit for the reviewed configuration | 30\ \text{dBm} |
| Receiver noise figure | 7.0\ \text{dB} |
| Required detector SNR | 14\ \text{dB} |
| Implementation allowance | 2.0\ \text{dB} |
| Combined measurement and installation uncertainty | u_c=1.5\ \text{dB} |
| Coverage factor for release guard | k=2 |
These values are simplified. A real survey should also state regulatory domain, antenna model, polarization, final mounting height, cable part numbers, radio firmware, cybersecurity boundary, traffic classes, spectrum-analyzer settings, calibration status, nearby transmitters, metallic obstructions, production schedule, and expected future site changes.
Step 1: Define the Survey Boundary
The survey boundary should state exactly what is being accepted. For this project, accept the radio path from the installed access radio through the antenna, local RF environment, wireless endpoints, upstream switch port, monitoring system, and application traffic test.
Do not accept only received signal strength. A wireless system can show strong RSSI and still fail because of co-channel interference, adjacent-channel leakage, receiver blocking, multipath, retry storms, queueing delay, poor roaming behavior, or a missing operational baseline.
The survey record should include:
- floor or site plan with measurement grid points;
- antenna location, height, polarization, azimuth, and mounting notes;
- radio model, firmware, channel, bandwidth, transmit power, and security profile;
- measurement equipment, calibration date, resolution bandwidth, detector, averaging, and antenna setup;
- production state during survey, including moving equipment, vehicles, doors, cranes, and temporary transmitters;
- pass criteria for RF margin, packet loss, latency, jitter, and monitoring thresholds;
- exception list for areas not surveyed or conditions not reproduced.
Step 2: Check EIRP and Configuration Legality
Effective isotropic radiated power is:
where P_t is transmitter output power, L_f is feeder and connector loss, and G_t is antenna gain.
For the baseline:
Regulatory headroom is:
Engineering Comment
The proposed configuration is below the stated EIRP limit. That does not mean the site should increase transmit power. Extra power may worsen coexistence, raise adjacent-channel leakage, create hidden-terminal behavior, or overload nearby receivers. In a shared industrial space, antenna placement and channel planning are often better mitigation tools than simply increasing power.
Step 3: Estimate Receiver Sensitivity
Receiver thermal-noise power is:
For:
the bandwidth term is:
With:
the receiver noise floor is:
Required received signal is:
Use:
Engineering Comment
This sensitivity estimate is a design screen, not a substitute for modem qualification. The actual receiver limit depends on waveform, coding, packet length, interference, adjacent-channel selectivity, implementation loss, antenna mismatch, fading statistics, and target error rate. The estimate is still useful because it gives the survey a numeric release boundary.
Step 4: Build the Measurement Grid
Define a grid that represents the service area. In this example, the hall is divided into six representative zones.
| Zone | Location | Initial measured RSSI | Notes |
|---|---|---|---|
| A | control-room doorway | -57\ \text{dBm} | clear path |
| B | packaging line | -63\ \text{dBm} | moving metal frames |
| C | compressor skid | -72\ \text{dBm} | partial obstruction |
| D | maintenance bay | -68\ \text{dBm} | nearby handheld radios |
| E | loading door | -75\ \text{dBm} | door open/closed changes path |
| F | mezzanine cabinet | -69\ \text{dBm} | cable trays and switchgear |
The weakest initial point is Zone E:
Nominal margin above sensitivity is:
The nominal margin exceeds the required 12\ \text{dB} by only 1\ \text{dB} before uncertainty is applied. That is not enough for release.
Step 5: Apply Measurement-Uncertainty Guarding
The guarded release margin is:
where:
and:
For the initial survey:
The release criterion is:
The initial installation fails by:
Engineering Comment
The unguarded link appears acceptable, but the guarded result does not. This is exactly why a site survey should not use a single RSSI threshold with no uncertainty allowance. A small margin can disappear through cable replacement, antenna movement, endpoint orientation, seasonal equipment placement, or measurement setup variation.
Step 6: Compare Candidate Channels
Run a spectrum survey at representative locations using documented resolution bandwidth, detector, averaging, span, antenna orientation, and measurement time. For this example, three candidate channels are compared.
| Candidate channel | Median noise floor | Worst in-channel interferer | Occupancy above -85\ \text{dBm} | Engineering note |
|---|---|---|---|---|
| A | -91\ \text{dBm} | -78\ \text{dBm} | 38\% | busy handheld traffic |
| B | -96\ \text{dBm} | -89\ \text{dBm} | 6\% | lowest occupancy |
| C | -94\ \text{dBm} | -73\ \text{dBm} | 4\% | rare high-power pulse |
Assume the measured desired signal after antenna adjustment is expected to be:
Screen carrier-to-interference ratio as:
For Channel A:
For Channel B:
For Channel C:
If the survey requires at least 15\ \text{dB} carrier-to-interference margin under credible observed occupancy, Channel B passes by:
Channels A and C fail.
Engineering Comment
Channel C has lower occupancy percentage than Channel A, but the strongest observed pulse is too close to the desired signal level. Occupancy percentage alone is not enough. A site-survey decision should consider occupancy, interferer amplitude, duty cycle, waveform type, receiver selectivity, service consequence, and whether the interferer is seasonal or operationally controlled.
Step 7: Select Mitigation Actions
The survey team evaluates four mitigation actions.
| Action | Expected benefit | Cost or risk | Decision |
|---|---|---|---|
| Move antenna from 2.5\ \text{m} to 4.0\ \text{m} height | improves obstruction margin in Zones C and E | requires safer cable route and bracket review | accept |
| Use Channel B | avoids busy handheld channel and high-power pulse | must update channel plan and labels | accept |
| Add receiver preselector | improves blocker tolerance | adds 1.5\ \text{dB} desired-signal loss | conditional |
| Increase transmit power by 4\ \text{dB} | improves RSSI | worsens coexistence and is unnecessary after relocation | reject |
After antenna relocation and Channel B selection, repeat the grid survey.
| Zone | Corrected measured RSSI |
|---|---|
| A | -56\ \text{dBm} |
| B | -61\ \text{dBm} |
| C | -67\ \text{dBm} |
| D | -66\ \text{dBm} |
| E | -72\ \text{dBm} |
| F | -67\ \text{dBm} |
The corrected weakest point is:
Nominal margin:
The corrected installation passes the 12\ \text{dB} guarded-margin requirement by:
Engineering Comment
The pass is real but thin. The release package should not hide that fact. A 1\ \text{dB} guarded surplus means change control matters: moving the antenna, changing cable type, adding nearby transmitters, or changing the channel width can invalidate the survey.
Step 8: Check Receiver Blocking Tradeoff
If strong nearby transmitters are credible, evaluate whether a receiver preselector is needed. Suppose two nearby maintenance transmitters can appear at the receiver input at:
and:
The receiver input third-order intercept is:
For a two-tone screen:
Before filtering:
Using the corrected weakest desired signal:
carrier-to-intermodulation ratio is:
If the requirement is 15\ \text{dB}, the unfiltered receiver fails by:
Now evaluate a preselector that attenuates each blocker by 15\ \text{dB} but adds 1.5\ \text{dB} desired-signal loss.
Filtered blocker levels:
Filtered desired signal at the weakest point:
New intermodulation estimate:
New carrier-to-intermodulation ratio:
The preselector strongly resolves the intermodulation screen, but it reduces RF margin. Recompute guarded margin with the filter:
The filtered configuration would miss the 12\ \text{dB} guarded weak-signal criterion by:
Engineering Comment
The filter is excellent for intermodulation but marginal for weakest-point sensitivity. The correct decision is conditional: install the preselector only if the antenna relocation, cable-loss reduction, or endpoint antenna orientation recovers at least 0.5\ \text{dB} plus practical margin. This is a real engineering tradeoff, not a one-line accessory choice.
Step 9: Validate Packet Service Under Load
RF margin is not the final acceptance test. Run traffic validation with representative packet size, update rate, retry configuration, security overhead, and background network load.
For the corrected Channel B configuration without preselector, the survey records:
| Metric | Requirement | Measured result | Decision |
|---|---|---|---|
| packet loss during survey load | less than 0.5\% | 0.18\% | pass |
| 95th-percentile one-way latency | less than 50\ \text{ms} | 38\ \text{ms} | pass |
| peak-to-peak jitter | less than 20\ \text{ms} | 16\ \text{ms} | pass |
| sensor update interval | less than 1.0\ \text{s} | 0.62\ \text{s} | pass |
| weakest guarded RF margin | at least 12\ \text{dB} | 13\ \text{dB} | pass |
Calculate the packet-loss count for traceability. If 20{,}000 packets were sent:
The maximum allowed loss count is:
The traffic test passes with:
of loss-count margin during the reviewed condition.
Engineering Comment
The traffic test should be tied to the RF state. If the site later changes channel, transmit power, antenna location, endpoint firmware, retry policy, traffic class, or nearby transmitters, the traffic evidence is no longer fully valid. The release package should make those dependencies visible.
Step 10: Set Monitoring Thresholds
Monitoring thresholds should be tied to the accepted survey baseline.
The minimum received level required for guarded release is:
The corrected weakest measured value is:
The accepted installation is therefore only:
above the release threshold.
Recommended monitoring:
| Signal | Warning threshold | Critical threshold | Response |
|---|---|---|---|
| 5th-percentile RSSI at weakest endpoints | below -73\ \text{dBm} | below -76\ \text{dBm} | inspect antenna, cable, endpoint orientation, obstructions |
| packet loss over survey-like load | above 0.3\% | above 0.5\% | check channel occupancy and retries |
| 95th-percentile latency | above 40\ \text{ms} | above 50\ \text{ms} | inspect queueing, retries, and traffic class |
| peak-to-peak jitter | above 16\ \text{ms} | above 20\ \text{ms} | inspect scheduling, congestion, and interference |
| occupancy above -85\ \text{dBm} on Channel B | above 10\% | above 20\% | rerun spectrum survey and channel-plan review |
Engineering Comment
The warning thresholds are intentionally close to the release evidence because the accepted surplus is small. A loose default alarm would allow the site to drift below the reviewed condition before anyone investigates.
Deliverable Structure
The final site-survey and interference-mitigation package should contain:
- service boundary, required performance, and acceptance criteria;
- site plan with measurement grid and excluded areas;
- radio configuration, firmware, security profile, channel, bandwidth, antenna, cable, and EIRP calculation;
- survey equipment settings, calibration status, date, time, production state, and measurement uncertainty;
- spectrum occupancy plots or tables for candidate channels;
- coverage measurements before and after mitigation;
- receiver sensitivity, carrier-to-interference, and guarded-margin calculations;
- mitigation decision table with accepted and rejected actions;
- traffic validation results with packet count, packet loss, latency, jitter, and update interval;
- operational monitoring thresholds and change-control triggers;
- release decision, residual risks, and re-survey conditions.
Release Decision
For the baseline project, release the corrected Channel B configuration with the relocated antenna and documented monitoring thresholds. Do not install the preselector unless a follow-up field adjustment recovers enough weak-signal margin to preserve the 12\ \text{dB} guarded criterion after filter insertion.
The release is conditional on change control. Re-survey is required if any of the following change:
- antenna location, height, polarization, gain, or cable loss;
- channel, bandwidth, firmware, security mode, transmit power, or retry policy;
- nearby transmitters, rotating equipment, switchgear, radar, temporary gateways, or handheld fleet;
- production layout, doors, racks, cable trays, cranes, large metal parts, or access restrictions;
- packet load, traffic class, endpoint density, update interval, or service consequence.
Common Mistakes
A site survey is weak if it reports only colored coverage points without the RF and service basis. Common mistakes include:
- accepting strong RSSI while ignoring interference and packet retries;
- using occupancy percentage without checking interferer amplitude;
- increasing transmit power when channel planning or antenna placement is the real issue;
- omitting measurement uncertainty from the release decision;
- testing traffic in a quiet maintenance window and then accepting production operation;
- failing to record analyzer settings, antenna orientation, firmware, channel width, and security overhead;
- leaving operations with no monitoring thresholds tied to the accepted baseline.
The strongest wireless survey is not the one with the prettiest coverage map. It is the one that leaves a repeatable engineering record: what was measured, what was rejected, why the selected channel was accepted, how much margin remains, and which future changes invalidate the evidence.