Learning CenterDrive Test & Optimization
Network Testing 15 min read

Drive Test & RF Optimization:
A Complete Guide

Drive testing remains the most reliable method for validating radio network performance in real-world conditions. While planning tools predict where coverage should be, and OMC counters report what the network thinks is happening, drive tests reveal what subscribers actually experience — on the specific roads, in the specific environments, and at the specific times that matter to your business.

What Is Drive Testing?

Drive testing is the systematic collection of radio and application performance measurements from a moving vehicle equipped with test UEs (User Equipment) and measurement software. The test vehicle follows a predetermined route — covering major roads, priority coverage areas, and known problem locations — while measurement equipment records RSRP, RSRQ, SINR, throughput, and application-layer KQIs synchronized to GPS coordinates and timestamps.

The output is a georeferenced dataset: every measurement point linked to a precise location, enabling visualization as a color-coded map overlay and statistical analysis by geographic segment (road type, distance from site, indoor/outdoor split). This geospatial dimension is what makes drive test data actionable — it transforms raw numbers into a spatial picture of network performance that engineers can interpret and operators can explain to regulators.

LIVE MEASUREMENTRSRP-72 dBmRSRQ-9 dBSINR14.3 dBDL Tput487 MbpsExcellent > -80 dBmGood -90 to -80Fair -100 to -90Weak < -100 dBm

A drive test route colored by measured RSRP. Green = excellent coverage, yellow/orange = marginal, red = coverage hole requiring attention.

Drive Test Equipment

A typical 5G drive test setup consists of several components working in concert:

Commercial Test UEs

Smartphones running test client software (TEMS Pocket, NEMO Handy, XCAL-Mobile) that log RRC-layer measurements alongside application throughput tests. Multiple UEs per vehicle provide statistical robustness and cover different frequency bands simultaneously.

Dedicated Measurement Receivers

Purpose-built scanners (TEMS Discovery, R&S TSME) decode broadcast channels without connecting as a subscriber, capturing RSRP/RSRQ from all visible cells simultaneously — not just the serving cell. Essential for neighbor cell analysis and interference mapping.

GPS Receiver

A high-accuracy GPS receiver (±2m typical) provides the position lock that makes every measurement geographically meaningful. GNSS-synchronized timestamping ensures accurate correlation between measurements from different devices.

Route Planning Software

Pre-defined routes are loaded into the measurement software so that campaigns can be repeated on the same roads over time — enabling before/after comparisons after optimization or software upgrades.

Test Methodology

Route Design

Effective route design balances coverage breadth against statistical depth. A route that covers every road in the served area produces maximum geographic coverage but minimum samples per location — making it difficult to distinguish genuine coverage holes from measurement noise. Best practice is to define a hierarchical route set: a primary trunk route (major roads, repeated 3+ times per campaign for statistical significance) and supplemental routes for priority areas, coverage complaints, and competitor benchmarking zones.

Test Traffic Generation

Alongside passive radio measurements, active service tests run continuously to measure end-user experience. Standard tests include: continuous FTP download (saturating the downlink to measure achievable throughput), periodic HTTP downloads (measuring TCP session establishment and content download time), VoNR call sequences (capturing MOS and packet loss), and ping sequences (measuring round-trip latency). The active tests must be coordinated to avoid mutual interference — running a 1 Gbps FTP download simultaneously with a VoNR call test will make the voice test look worse than it really is.

Sampling Rate and Speed

At 50 km/h driving speed with 1-second measurement intervals, samples are spaced approximately 14 meters apart. For macro cell coverage testing this is adequate — coverage changes slowly with distance from a macro tower. For dense urban small cell testing or handover analysis, finer spatial resolution (sub-second sampling) is needed. Some measurement systems support 100ms or 200ms sampling rates, but this increases file size and post-processing time significantly.

Key Metrics and Analysis

Beyond individual sample values, the most useful drive test analysis looks at statistical distributions of each metric across geographic zones. Standard reports include:

CDF Analysis (Cumulative Distribution Function)

The CDF of RSRP shows what percentage of measurement locations fall above each RSRP threshold. The standard coverage KPI — "percentage of route samples above −100 dBm RSRP" — comes directly from the CDF. Comparing CDFs between campaigns reveals whether a software update or parameter change improved coverage across the distribution, not just at specific locations.

Serving Cell Analysis

Which cells are serving which parts of the route, and how frequently does the UE switch serving cells? Excessive serving cell changes (pilot pollution) indicate that multiple cells are at similar RSRP and the UE cannot settle — a classic cause of throughput degradation near cell boundaries.

Handover Analysis

Successful handover rate, handover ping-pong ratio, and time-on-target cell after handover all reveal the health of the mobility configuration. A handover ping-pong (A→B→A within seconds) is almost always a misconfigured handover offset or TTT (Time-to-Trigger) value.

Throughput vs SINR Correlation

Plotting measured throughput against measured SINR for each sample reveals whether the actual MCS selection is following the expected spectral efficiency curve. Anomalously low throughput at high SINR is a strong indicator of scheduler bugs, configuration errors, or front-haul bottlenecks.

The RF Optimization Cycle

Drive testing is only valuable when it feeds directly into an optimization action. The cycle moves through four stages, repeated until KPI targets are met:

1

Identify

Plot drive test measurements on the map and compare against planning tool predictions. Identify locations where measured RSRP or throughput falls below target, and correlate with serving cell, antenna azimuth, and terrain.

2

Diagnose

Determine the root cause: Is it a coverage gap (no cell close enough)? Pilot pollution (too many cells at similar strength)? Antenna tilt too steep (coverage not reaching far enough)? Handover failure (UE stuck on a distant cell)?

3

Optimize

Apply the corrective action: mechanical or electrical tilt adjustment, power level change, azimuth rotation, neighbor list update, handover parameter modification, or site addition. Changes are tracked in the network parameter management system.

4

Verify

Repeat the drive test on the same route after the change. Compare the new measurements against the pre-change baseline. Confirm that the KPI improved and that no neighboring area was inadvertently degraded.

80991181371561750km1km2km3km4km5kmPath Loss (dB)Free SpaceOkumura-HataCOST 231

Path loss model curves are calibrated against drive test measurements. A well-calibrated model reduces prediction error standard deviation below 8 dB.

Advanced Techniques

Walk Testing and Indoor Coverage

Walk testing extends the drive test methodology to pedestrian environments. Testers carry measurement equipment on foot through shopping centres, transport hubs, office buildings, and dense urban areas where vehicles cannot operate. Floor-by-floor indoor walk tests characterize the transition between outdoor macro coverage penetrating through building facades and indoor small cells or DAS providing dedicated indoor coverage. Indoor walk test data is georeferenced using BLE beacons, Wi-Fi fingerprinting, or CAD floor plans rather than GPS (which is unavailable indoors).

Drone Testing for Remote and Elevated Environments

Drones equipped with measurement UEs can test coverage in environments impossible for ground vehicles: highway interchanges, industrial sites, coastal cliffs, construction sites, and rural areas with no road access. At altitude, the drone also captures the three-dimensional coverage pattern of antenna beams — particularly useful for validating massive MIMO elevation beamforming, where coverage extends both horizontally and vertically above the antenna.

Post-Processing and Reporting

Raw drive test log files (TEMS .log, NEMO .nmf, XCAL .xcal) must be processed before analysis. The standard processing pipeline: parse vendor-specific binary format → extract georeferenced samples → bin samples by geographic cell (100m × 100m) → compute per-bin statistics (mean, median, 5th/95th percentile) → compare against planning tool prediction → generate KPI report.

Automated report generation converts the processed dataset into regulator-ready coverage maps (PDF with color scale, legend, and statistics tables) and engineering reports (anomaly list, optimization priority ranking, before/after comparison). Regulators in most markets require annual or quarterly drive test evidence of coverage claim compliance.

NEXT GIS Drive Test Integration

NEXT GIS imports drive test data in standard formats (CSV with lat/lon/RSRP columns, GeoPackage, GeoJSON) and automatically renders it as a georeferenced point layer on the map canvas. The planning tool's coverage prediction is displayed simultaneously on the same map as a raster or polygon layer. The delta between predicted and measured RSRP is computed per sample and visualized as a separate layer — immediately surfacing the cells and locations where calibration is most needed.

Multi-format Import

CSV, GeoPackage, GeoJSON, and TEMS native log formats supported.

Prediction vs Measured

Delta layer computed automatically — filter to samples where error exceeds ±8 dB.

Regulator Export

PDF coverage maps with statistics tables, ready for regulatory submission.

Import drive test data