Md Tarek Hassan | PhD Researcher in 6G Wireless, RIS and ISAC

Define the Question Before the Metric

Start with one question: is the feed matched, is power radiated in the right direction, or does the deployed link actually work?

Port test: measures what reflects at the antenna feed.
Radiation test: measures efficiency, pattern, gain, and polarization.
Field test: measures received power and service under real paths and interference.
Fair comparison: record frequency, bandwidth, polarization, port, angle, distance, environment, and measurement method.

References for this section3

Parameter	Physical meaning	Use it when	Do not infer
$S_{11}$ / impedance bandwidth	Reflection at a specified feed reference	Checking usable matching across frequency	Good radiation or coverage from matching alone
Radiation efficiency	Fraction of accepted power actually radiated	Comparing loss in antenna designs	Beam direction from efficiency alone
Realized gain and pattern	Useful radiation in each direction including mismatch	Link-budget and coverage evaluation	Whole-area quality from only peak gain
HPBW and sidelobe level	Main-lobe width and unintended directional peaks	Planning pointing and interference	A full 3D pattern from a single 2D cut
Cross-polarization and port isolation	Unwanted orthogonal field or coupling between ports	Dual-polarized antennas and arrays	Independent MIMO streams without checking the channel

Directivity: directional concentration without loss.
Gain: includes radiation efficiency.
Realized gain: also includes feed mismatch. Compare values only with the same definition and reference.

References for this section3

Feed port: calibrate the vector network analyser and measure $S_{11}$ across the band.
Interpretation: $S_{11}=-10$ dB corresponds to roughly 10% reflected power; it does not measure radiation efficiency.
Decision: choose a matching threshold from the link requirement, not from a universal pass/fail rule.
Radiation stage: measure co- and cross-polar patterns, realized gain, HPBW, sidelobes, and efficiency.
Array scan: repeat away from boresight; steered gain can fall while sidelobes grow.
Comparison: use identical angular cuts and power normalization.
Multiport isolation: tells how much one port couples into another.
MIMO check: also test channel correlation, receive SNR, and independent spatial paths. Isolation alone does not prove two usable streams.

References for this section3

Parameter	Physical meaning	Use it when
EIRP and received power	Radiated budget in a direction and power at the receiver	Checking coverage and regulatory or hardware constraints
Path loss and shadowing margin	Average attenuation and slower obstruction variation	Siting and coverage prediction
PDP and RMS delay spread	Relative strength and time separation of received paths	Assessing frequency selectivity and equalization needs
Doppler spread	Speed of channel variation	Choosing channel-tracking and pilot update rates
SINR / outage probability	Quality relative to interference and noise; fraction of cases below a criterion	Evaluating robust service, not just average coverage

The basic power budget is

P_{\mathrm{r,dBm}}= P_{\mathrm{t,dBm}}+G_{\mathrm{t,dBi}}+G_{\mathrm{r,dBi}} -L_{\mathrm{feed}}-L_{\mathrm{path}}-L_{\mathrm{other}}

$P_{\mathrm{r,dBm}}$ , $P_{\mathrm{t,dBm}}$ : received and transmit power in dBm.
$G_{\mathrm{t,dBi}}$ , $G_{\mathrm{r,dBi}}$ : transmit and receive directional gain in dBi.
$L_{\mathrm{feed}}$ , $L_{\mathrm{path}}$ , $L_{\mathrm{other}}$ : feed, propagation, and remaining losses in dB.
Avoid double-counting: EIRP already includes transmit-side gain and relevant preceding losses. State whether receive gain is peak, realized, or measured toward the user.

References for this section3

Take an illustrative 30 dBm EIRP link with 14 dBi receive gain, 110 dB path loss, 3 dB other loss, and $-80$ dBm receiver sensitivity:

Clear path: $P_r=30+14-110-3=-69$ dBm; nominal margin is 11 dB.
Obstructed path: another 15 dB loss gives $-84$ dBm; margin becomes $-4$ dB.
Evaluation: report the fraction of user locations below the service threshold, not only the healthy-looking average.

Sensitivity and loss values here are example inputs, not universal requirements.

References for this section3

Compare Design A (narrow, high peak gain) with Design B (broad, lower peak gain):

Keep transmit power, frequency, mounting height, and receiver positions the same.
Measure received power and SINR on the same spatial grid.
Report a coverage percentile or outage fraction, not only peak gain.
Include beam-tracking overhead for moving users. A may win point-to-point while B covers more people.

RIS experiment: record the no-RIS baseline, tested states, positions, and channel realizations.
Fairness: a peak selected from many RIS states is not directly comparable with one baseline measurement.
Report: selection procedure and distribution of gains.

References for this section3

Fix the scenario: operating band, bandwidth, link distance, orientation, traffic or interference, and target service.
Check the antenna port and radiation properties separately: matching, efficiency, pattern, and polarization.
Build a link budget with explicit gains and losses; note whether gain already includes mismatch.
Test representative LOS, blocked, and mobile channel conditions rather than reporting only one favourable point.
Report distributions or worst-case margins alongside averages, and state the threshold used for outage.

Cellular measurements: RSRP, RSSI, SNR, SINR, and CQI.
Array design: geometry and phased arrays.
Decision rule: no universal optimal number; compare against a stated service target and test condition.

References for this section3