5G Network Performance Comparison between SDR-Based vs. AW2S Base Station Unit for Underground Infrastructure
|
Configuration
|
Avg Latency (ms)
|
Max Latency (ms)
|
Stability
|
|---|---|---|---|
|
Leaky RX–TX
|
28.93
|
38059
|
Excellent
|
|
Leaky TX–RX
|
72.08
|
237.02
|
Unacceptable
|
|
Standard Duplexer
|
29.83
|
45.71
|
Excellent
|
|
Amplified Duplexer
|
33.77
|
61.80
|
Moderate
|
Abstract
This comparative analysis evaluates the performance differences between two distinct base station configurations developed by the University of Patras and deployed in underground tunnel environments: an SDR-based (Software-Defined Radio) system installed in a corridor environment and a commercial AW2S Radio Unit deployed in a laboratory space. The study examines bandwidth performance, latency characteristics, coverage, and system reliability using a leaky cable-based 5G distribution system. In the experiment, five 5G modems acting as sensor nodes were installed directly along the leaky feeder cable, enabling controlled and repeatable measurements as a function of distance along the cable.
The findings reveal significant performance differences between the two deployment types. The commercial AW2S Radio Unit consistently demonstrates superior performance across most metrics, delivering average downlink speeds of 300.64 Mbps compared to the SDR system’s 65.80 Mbps, representing a 357% performance advantage. The commercial unit achieves complete coverage success at all five sensor positions. The SDR system experiences connectivity issues at sensor 5417 due to technical issues with SIM card validation, preventing detection by the base station. Latency characteristics show the AW2S unit maintaining 27.30 ms average compared to the SDR’s 28.52 ms average.
1. Experimental Setup and Leaky Cable Deployment
1.1 Base Station Development and Installation
Both base station systems evaluated in this study were developed by the University of Patras as part of research into 5G infrastructure for underground environments. The SDR-based system was installed in a corridor environment to provide optimal signal propagation conditions, while the commercial AW2S Radio Unit was deployed in a laboratory space. This installation configuration allows for direct comparison of the two systems under different environmental conditions representative of realistic underground deployment scenarios.
1.2 Leaky Cable Architecture
All measurements in this study were conducted using a leaky feeder cable deployed along the tunnel/corridor environment to emulate realistic underground 5G coverage conditions. Instead of discrete antennas at each measurement point, radio frequency energy was radiated continuously along the length of the leaky cable, providing distributed coverage and minimizing sharp coverage transitions typical of point-antenna deployments.
The 5G base station (either the commercial AW2S Radio Unit or the SDR-based system) was connected to the leaky cable, and user equipment was represented by multiple 5G modems installed directly along the cable. This approach ensures that variations in performance are primarily attributable to distance along the leaky cable and base station characteristics, rather than antenna placement variability.
1.3 5G Modem Placement and Sensor Identification
A total of five 5G modems were installed along the leaky cable, acting as fixed sensor nodes. The modems were mounted at intervals along the cable to ensure consistent spatial sampling. Each modem is uniquely identified by a sensor ID and its corresponding distance from the reference point at the initiation of the leaky cable, as delineated below:
|
Sensor ID
|
Distance from Reference Point
|
|---|---|
|
5417
|
150 cm
|
|
5381
|
230 cm
|
|
5395
|
310 cm
|
|
5469
|
390 cm
|
|
5188
|
460 cm
|
<Table 1. Sensor ID per Distance>
This configuration enables direct correlation between measured performance metrics (throughput, latency, and reliability) and physical distance along the leaky cable. The spacing ensures sufficient resolution to observe signal degradation, modulation and coding scheme (MCS) adaptation, and coverage characteristics of both deployment configurations.
1.4 Measurement Methodology
At each modem location, multiple downlink and uplink throughput tests were performed along with round-trip latency measurements. Tests were repeated to capture short-term variability and to evaluate measurement consistency. All measurements were conducted under controlled conditions to ensure valid comparison between the commercial radio unit and the SDR-based system.
1.5 Technical Note on Sensor 5417
During testing with the SDR-based system, sensor 5417 (150 cm position) experienced technical issues related to SIM card validation, preventing the modem from being detected by the base station. As a result, no performance data was collected for this sensor position with the SDR system. All measurements at this position show null values in the dataset. The AW2S Radio Unit successfully connected to sensor 5417 without issues, providing complete performance data across all five sensor positions.
The remainder of this report analyzes the measured results in detail, focusing on bandwidth performance, latency behavior, coverage completeness, and overall system reliability as a function of base station configuration and leaky cable propagation characteristics.
2. System Architecture Overview
2.1 Commercial Radio Unit Configuration (AW2S RU)
The commercial AW2S Radio Unit represents a purpose-built 5G infrastructure solution deployed in corridor space for optimal coverage throughout the tunnel environment. The system operates at 3750MHz center frequency with 100MHz bandwidth allocation, enabling high-capacity data transmission through efficient spectrum utilization. The 30kHz subcarrier spacing numerology provides balanced performance for underground propagation characteristics, while the TDD configuration allocating 7 downlink slots and 2 uplink slots optimizes for typical asymmetric traffic patterns. The single transmit and receive antenna configuration (1T1R) simplifies deployment while providing adequate coverage for tunnel geometries.
The most significant architectural advantage of the commercial unit lies in its +33dBm RF antenna port output power, delivering 2000 times the transmit power of the SDR system (33dB difference). This substantial power advantage enables robust signal penetration throughout the tunnel environment, overcoming propagation losses and maintaining reliable connectivity along the leaky cables. The commercial hardware implementation includes optimized RF front-end components, efficient power amplification, and sophisticated signal processing that together enable the superior performance characteristics observed in testing.
Dynamic modulation and coding schemes with maximum 256-QAM capability allow the system to adapt transmission parameters based on channel conditions, maximizing throughput when signal quality permits while maintaining connectivity under challenging conditions. The default network slice configuration with SST:1 and 5QI:9 provides non-guaranteed bit rate service appropriate for general data applications, while maintaining quality-of-service parameters suitable for real-time communications.
2.2 SDR-Based Configuration
The SDR-based system represents a flexible software-defined radio implementation installed in laboratory space, offering configurability and experimental capabilities at the cost of reduced performance compared to commercial solutions. Operating at the same 3750MHz center frequency as the commercial unit, the SDR system utilizes 50MHz bandwidth allocation, exactly half the spectrum resources of the commercial deployment. This bandwidth limitation fundamentally constrains maximum achievable throughput regardless of other system optimizations.
The critical limitation of the SDR configuration manifests in its 0dBm RF antenna port output power, delivering only 0.05% of the commercial unit’s transmit power. This 33dB power deficit severely limits coverage range, signal penetration capability, and resilience to propagation impairments. The low output power combined with laboratory installation in enclosed space rather than optimal corridor placement creates coverage gaps that manifest as complete connectivity failures at certain sensor positions.
Despite using identical numerology (30kHz SCS), TDD configuration (7 DL / 2 UL slots), antenna configuration (1T1R), and dynamic MCS with 256-QAM maximum, the fundamental limitations of 50MHz bandwidth and 0dBm output power prevent the SDR system from achieving performance comparable to the commercial unit. The software-defined architecture provides valuable flexibility for research and development applications but cannot overcome the hardware limitations that constrain operational deployment viability.
3. Bandwidth Performance Analysis
3.1 Commercial Radio Unit Performance (AW2S RU)
The commercial AW2S Radio Unit demonstrates exceptional bandwidth performance across all sensor positions, achieving an average downlink throughput of 300.64 Mbps. The system maintains remarkably consistent performance with minimal variation between positions, indicating robust signal strength and reliable high-order modulation throughout the deployment area despite being installed in a laboratory environment.
|
Position
|
Downlink (Mbps)
|
Uplink (Mbps)
|
|---|---|---|
|
150cm
|
311
|
35.6
|
|
230cm
|
310
|
32.4
|
|
310cm
|
313
|
34
|
|
390cm
|
299
|
27.1
|
|
470cm
|
256.83
|
36.2
|
<Table 2. Average Throughput per Distance for AW2S RU>
The commercial unit’s ability to maintain high throughput across all positions demonstrates effective coverage throughout the tunnel environment. The minimal performance variation between the first four positions (299 to 313 Mbps) indicates remarkably stable channel conditions. The modest reduction at position 5188 (256.83 Mbps average) represents only a 14.6% decrease from peak performance, demonstrating excellent signal propagation characteristics along the leaky cable.
Uplink performance shows consistent results with an average of 34.24 Mbps, ranging between 27.1 and 38.8 Mbps across positions. The stable uplink performance indicates balanced bidirectional communication capability suitable for interactive applications requiring significant reverse-channel capacity.
3.2 SDR-Based System Performance
The SDR-based system deployed in corridor space demonstrates reduced bandwidth performance compared to the commercial unit, achieving 65.80 Mbps average downlink throughput across the four positions where connectivity was established. The system exhibits substantial performance variability across different positions and between repeated measurements at the same location.
|
Position
|
Downlink (Mbps)
|
Uplink (Mbps)
|
|---|---|---|
|
150cm
|
FAILED
|
FAILED
|
|
230cm
|
92.67
|
25.33
|
|
310cm
|
68
|
28.17
|
|
390cm
|
45.57
|
28.77
|
|
470cm
|
65.62
|
25.97
|
<Table 3. Average Throughput per Distance for SDRs>
The performance pattern shows interesting characteristics, with position 5381 at 230 cm achieving the highest throughput (92.67 Mbps average), followed by declining performance at greater distances. The furthest position (5188 at 460 cm) shows significant variability, with measurements spanning from 36.3 to 82.2 Mbps, indicating inconsistent channel conditions or frequent modulation scheme adaptations.
Uplink performance averages 27.02 Mbps across measured positions, ranging from 21.0 to 29.8 Mbps. While this represents adequate capacity for many applications, it demonstrates lower reverse-channel throughput compared to the commercial unit’s 34.24 Mbps average.
3.3 Comparative Bandwidth Analysis
The performance differential between configurations proves substantial, with the commercial AW2S unit delivering 357% more downlink bandwidth on average (300.64 Mbps versus 65.80 Mbps). This significant advantage demonstrates the superior performance capabilities of the commercial system despite its laboratory installation environment compared to the SDR system’s corridor placement.
Peak throughput comparison shows even greater disparity, with the commercial unit achieving 313 Mbps maximum (at position 5395) versus the SDR’s 97.7 Mbps maximum (at position 5381), representing a 220% advantage at best-case performance points. The commercial unit’s minimum single measurement of 245 Mbps exceeds the SDR system’s best measurement of 97.7 Mbps by 151%, demonstrating that commercial worst-case performance substantially surpasses SDR best-case performance.
The uplink performance differential shows the commercial unit delivering 27% more reverse-channel capacity (34.24 Mbps versus 27.02 Mbps). While this advantage is less pronounced than the downlink differential, it still represents meaningful additional capacity for applications requiring substantial uplink bandwidth.
Regarding distance performance characteristics, the SDR system shows its best performance at position 5381 (230 cm) with 92.67 Mbps average, while the commercial unit maintains peak performance across the range from 150 cm to 310 cm (299-313 Mbps) before showing modest degradation at 460 cm. This suggests the commercial unit maintains more consistent performance across distance variations.
4. Latency Performance Analysis
4.1 Commercial AW2S Radio Unit Latency
The commercial AW2S Radio Unit maintains excellent latency characteristics with 27.30 ms average round-trip time across all positions where ping data was collected. The system demonstrates consistent low-latency operation suitable for real-time applications including voice communications, video conferencing, and interactive control systems.
The commercial AW2S Radio Unit maintains excellent latency characteristics with 27.30 ms average round-trip time across all positions where ping data was collected. The system demonstrates consistent low-latency operation suitable for real-time applications including voice communications, video conferencing, and interactive control systems.
|
Position
|
Latency (ms)
|
|---|---|
|
150cm
|
23.52
|
|
230cm
|
24.63
|
|
310cm
|
29.45
|
|
390cm
|
29.45
|
|
470cm
|
29.44
|
<Table 4. Average Latency per Distance for AW2S RU>
The latency range shows excellent consistency for the first four positions, with average latency spanning only 5.93 ms between the lowest (23.52 ms at position 5417) and highest values (29.45 ms at positions 5395 and 5469). The furthest position (5188) shows slightly elevated maximum latency at 55.23 ms but maintains similar average latency, indicating that even at maximum tested distance, the system delivers predictable responsiveness.
The minimum latency values (13.81 to 18.76 ms) demonstrate the system’s capability for low-latency transmission when channel conditions are optimal, while maximum values remaining below 56 ms even at the furthest position confirm reliable performance for time-sensitive applications.
4.2 SDR System Latency
The SDR-based system achieves 28.52 ms average latency where connectivity exists and ping data was collected, representing acceptable performance comparable to the commercial unit’s 27.30 ms average. However, the system exhibits one extreme latency spike that raises concerns about performance stability.
|
Position
|
Latency (ms)
|
|---|---|
|
150cm
|
FAILED
|
|
230cm
|
27.14
|
|
310cm
|
28.34
|
|
390cm
|
31.35
|
|
470cm
|
26.21
|
<Table 5. Average Latency per Distance for SDRs>
While typical average latency values remain reasonable and competitive with the commercial unit, position 5381 exhibits an extreme maximum latency spike of 335.99 ms. This single anomalous event dramatically exceeds normal latency ranges and indicates potential system instability or processing delays under certain conditions. Excluding this outlier, the remaining maximum latency values (39.19 to 51.48 ms) remain within acceptable bounds for most applications.
The minimum latency values for the SDR system (8.84 to 18.50 ms) actually demonstrate lower latency potential than the commercial unit under optimal conditions, suggesting that the SDR system’s processing and transmission path can achieve excellent responsiveness when channel conditions are favorable and system load is minimal.
4.3 Latency Comparison
Average latency performance shows minimal difference between the two systems, with the commercial unit achieving 27.30 ms versus the SDR’s 28.52 ms, representing only a 4.5% advantage. This similarity in typical latency performance indicates that both systems deliver comparable responsiveness under normal operating conditions, despite the substantial difference in throughput capabilities.
The critical difference emerges in maximum latency behavior. The commercial unit demonstrates controlled maximum latency with 55.23 ms worst-case across all measurements, while the SDR system experiences a 335.99 ms extreme spike. This 509% worse-case differential indicates significantly greater susceptibility to severe performance degradation events in the SDR system. For applications requiring consistent responsiveness such as remote equipment control or voice communications, this stability difference proves critical even though average performance remains similar.
Minimum latency comparison favors the SDR system slightly, with 8.84 ms best-case versus the commercial unit’s 13.81 ms, suggesting the SDR system can achieve lower latency under optimal conditions. However, this 5 ms advantage in best-case scenarios proves less operationally significant than the commercial unit’s superior worst-case stability.
5. Coverage Analysis and System Reliability
5.1 Coverage Completeness
The commercial AW2S Radio Unit achieves 100% coverage success across all five sensor positions despite being deployed in a laboratory environment. This complete coverage indicates adequate signal strength at all locations, enabling consistent service availability essential for operational deployments where coverage gaps create safety risks or operational limitations.
The universal connectivity success demonstrates that the commercial unit successfully overcomes propagation losses inherent in underground environments, maintaining link budgets sufficient for reliable communication at all tested distances. The leaky cable distribution system combined with the commercial radio unit provides effective coverage throughout the tunnel geometry without dead zones or unreliable fringe areas.
5.2 SDR System Coverage Status
The SDR-based system achieves connectivity at four out of five sensor positions (80% success rate). Sensor 5417 at the 150 cm position shows no measurements due to technical issues with SIM card validation that prevented the modem from being detected by the base station. This represents a technical limitation rather than a signal coverage issue, as evidenced by the successful measurements at all other positions including those at greater distances from the base station.
The 80% coverage success rate, while impacted by the specific technical issue at sensor 5417, demonstrates that the SDR system deployed in the corridor environment can establish and maintain connectivity across the tested deployment area where technical conditions permit proper modem registration.
5.3 Measurement Consistency
The commercial unit exhibits excellent measurement consistency with multiple readings at position 5188 showing controlled variability. The three downlink measurements at this position (245, 256.5, 269 Mbps) demonstrate a standard deviation of approximately 12.3 Mbps, representing only 4.8% coefficient of variation. This low variability indicates stable channel conditions and consistent base station performance over time.
The SDR system shows greater measurement variability, particularly evident in the six measurements at position 5188 where downlink speeds range from 36.3 to 82.2 Mbps, a 45.9 Mbps range representing significant performance variation. The standard deviation of approximately 18.7 Mbps yields a 28.5% coefficient of variation, indicating less stable performance than the commercial unit. Similar but less pronounced variability appears in other position measurements, suggesting that the SDR system’s performance is more sensitive to changing channel conditions or system state.
At position 5469, the SDR system shows six measurements ranging from 42.7 to 46.9 Mbps, demonstrating more consistent performance (standard deviation ~1.5 Mbps, coefficient of variation 3.3%) at this specific location. This indicates that measurement consistency varies by position, with some locations showing stable performance while others exhibit high variability.
At position 5381, the SDR system achieves its highest throughput with three measurements ranging from 89.7 to 97.7 Mbps (standard deviation ~4.3 Mbps, coefficient of variation 4.6%), showing good consistency at this near-optimal position.
5.4 Performance Pattern Analysis
The commercial AW2S unit demonstrates a remarkably flat performance profile across distance, maintaining 299-313 Mbps throughput for the first four positions (150-390 cm) before showing only modest degradation to 256.83 Mbps average at 460 cm. This pattern indicates excellent signal propagation and robust link budget that maintains high-order modulation across the tested range.
The SDR system shows a more complex performance pattern. Performance peaks at position 5381 (230 cm) with 92.67 Mbps average, then gradually declines at positions 5395 (67.97 Mbps), 5469 (45.23 Mbps), and 5188 (65.62 Mbps). The non-monotonic pattern with position 5188 showing higher throughput than position 5469 suggests that distance is not the sole determining factor, and local propagation characteristics or multipath effects along the leaky cable may influence performance.
5.5 Reliability Scoring
Applying a comprehensive reliability scoring methodology that weights coverage completeness (40%), measurement consistency (30%), latency stability (20%), and maximum performance (10%), the configurations show different profiles reflecting their respective strengths and limitations.
The commercial AW2S Radio Unit achieves a reliability score of 94 out of 100, benefiting from 100% position success rate (40 points), excellent measurement consistency with 4.8% coefficient of variation (28 points), controlled maximum latency of 55.23 ms (18 points), and exceptional throughput performance of 300.64 Mbps average (10 points). The only modest deduction comes from the slightly elevated maximum latency at the furthest position.
The SDR system receives a reliability score of 70 out of 100, reflecting its mixed performance characteristics. The 80% coverage success rate (due to the SIM card validation issue at sensor 5417) yields 32 points, acknowledging the technical rather than coverage-related nature of the connectivity failure. The higher measurement variability of 28.5% coefficient of variation at the most variable position reduces the consistency score to 18 points. The extreme 335.99 ms latency spike decreases the stability rating to 12 points despite otherwise acceptable latency performance. The lower absolute performance of 65.80 Mbps average yields 5 points for maximum performance capability.
This 24-point reliability gap quantifies the difference between the commercial radio unit deployment and SDR system implementation. The scoring methodology reflects operational deployment priorities where coverage completeness ensures universal access, performance consistency enables capacity planning, latency stability supports real-time applications, and maximum throughput determines application capabilities.
6. Conclusions
6.1 Primary Findings
The comprehensive comparative analysis of two base station configurations developed by the University of Patras establishes several critical findings regarding their suitability for underground 5G deployment.
First, the commercial AW2S Radio Unit, despite being deployed in a laboratory environment, delivers 357% more downlink bandwidth than the SDR system installed in a corridor, achieving 300.64 Mbps average versus 65.80 Mbps. This substantial performance advantage demonstrates the superior capabilities of the commercial unit’s design and implementation.
Second, the commercial unit achieves 100% coverage success across all five tested sensor positions while the SDR system successfully connected to 80% of positions, with the connectivity issue at sensor 5417 attributed to technical problems with SIM card validation rather than signal coverage limitations.
Third, the commercial unit demonstrates superior performance consistency with 4.8% coefficient of variation at the most variable measurement location compared to the SDR’s 28.5%, indicating more stable operation across varying conditions. However, at certain positions the SDR system shows good consistency (3.3% coefficient of variation at position 5469), suggesting location-dependent performance characteristics.
Fourth, average latency performance proves comparable between systems, with the commercial unit at 27.30 ms and the SDR at 28.52 ms representing only a 4.5% difference. However, the SDR system exhibits an extreme maximum latency spike of 335.99 ms at position 5381, compared to the commercial unit’s controlled 55.23 ms maximum, indicating potential stability concerns under certain conditions.
Fifth, the commercial unit achieves a reliability score of 94 compared to the SDR’s 70, confirming systematic performance advantages in consistency, stability, and throughput capacity, despite the comparable average latency performance.
6.2 Performance Characteristics by Position
Analysis of position-specific performance reveals distinct patterns for each system. The commercial AW2S unit maintains remarkably consistent throughput across the range, with only 5% variation between positions 5417 through 5469 (299-313 Mbps), and a 14.6% reduction at the furthest position 5188 (256.83 Mbps average). This flat performance profile indicates robust link budget and effective signal propagation along the leaky cable.
The SDR system shows more complex spatial performance characteristics, with peak throughput at position 5381 (92.67 Mbps average), declining to 67.97 Mbps at position 5395, further declining to 45.23 Mbps at position 5469, before partially recovering to 65.62 Mbps at position 5188. This non-monotonic pattern suggests that factors beyond simple distance attenuation influence performance, potentially including local propagation characteristics, leaky cable coupling variations, or multipath effects at different positions.
Latency characteristics show less positional variation for both systems, with the commercial unit maintaining 23.52 to 29.45 ms average across positions and the SDR system showing 26.21 to 31.35 ms average (excluding the outlier spike at position 5381). This relative stability in latency across distance contrasts with the more variable throughput performance, suggesting that latency is less sensitive to signal strength variations within the range of received signal levels achieved in this deployment.
6.3 System Comparison Summary
The comparative analysis reveals that while both systems successfully provide 5G connectivity in underground environments, they demonstrate significantly different performance profiles. The commercial AW2S Radio Unit excels in throughput capacity, consistency, and stability, making it well-suited for applications requiring high bandwidth, predictable performance, and robust connectivity. The 300.64 Mbps average throughput enables demanding applications such as high-definition video transmission, multi-sensor data aggregation, and dense user scenarios.
The SDR-based system provides functional 5G connectivity with adequate performance for moderate-bandwidth applications. Its 65.80 Mbps average throughput suffices for standard communications, sensor networks with modest data rates, and single-user scenarios. The comparable average latency to the commercial unit (28.52 ms versus 27.30 ms) indicates that the SDR system can support real-time applications requiring low latency, though the extreme latency spike observed raises concerns about worst-case behavior.
Both systems, developed by the University of Patras, demonstrate the feasibility of providing 5G coverage in underground environments using leaky cable distribution. The commercial AW2S unit’s laboratory deployment achieving superior performance compared to the SDR system’s corridor installation highlights that system design and implementation quality significantly impact performance outcomes, potentially outweighing environmental installation advantages.
The technical issue preventing sensor 5417 detection by the SDR system due to SIM card validation problems represents an operational consideration for deployment planning, emphasizing the importance of thorough modem registration testing and network authentication procedures in production deployments.
This project has received partial funding from the Horizon Europe programme of the European Union under HORIZON-JU-SNS-2022 FIDAL program, grant agreement No. 101096146
English