Source-linked AI summary
A Novel 6G Dynamic Channel Map Based on a Hybrid Channel Model
Tianrun Qi, Cheng-Xiang Wang, Chen Huang, Jiayue Shi, Junling Li, Shuaifei Chen, El-Hadi M. Aggoune
TL;DR
The paper addresses the difficulty of updating accurate channel maps when 6G physical environments change. It proposes an RT-GSHCM-based dynamic channel map that combines offline RT modeling with online stochastic updates, and reports accurate channel information with substantially faster updates than conventional channel maps. The study also derives statistical channel properties and examines model scope and assumptions.
Problem
Conventional channel maps focus on static environments and cannot timely reflect physical-environment changes, while 6G scenarios require accurate channel information for system design and optimization.
Method
The paper constructs a dynamic channel map by modeling static interaction objects with RT and dynamic interaction objects with 6GPCM in the proposed RT-GSHCM.
Results
The evaluation reports accurate DCM channel information and much lower update time than conventional channel maps, based on comparisons with measurements, RT, and 6GPCM.
Takeaways & Limitations
The RT-GSHCM provides a channel-map construction approach that combines time-varying updates with the accuracy requirements evaluated for 6G communication environments.
Abstract
from arXiv · showhide
In the sixth generation (6G) wireless communication networks, the device density, antenna number, and the complexity of communication scenarios will significantly increase, which brings great challenges for system design and network optimization. By obtaining channel information in advance, channel map has become a promising solution to these challenges in 6G era. However, conventional channel maps cannot be updated in time as physical environment changes. To solve the problem, a novel dynamic channel map (DCM) is proposed in this work. For DCM construction, we further present a ray tracing (RT) and geometric stochastic hybrid channel model (RT-GSHCM), which pre-constructs the DCM offline by RT and updates it online by geometry-based stochastic channel model (GBSM). By this way, the DCM can provide time-varying channel information and channel properties while matintaining accuracy. Next, a channel measurement campaign is conducted, and the measurement results are compared with the RT-GSHCM, RT, and GBSM. The comparison results validate the accuracy of DCM. Meanwhile, the time cost on DCM update is compared with that of conventional channel maps, illustrating the time-efficiency of DCM. Finally, important statistical channel properties of RT-GSHCM are further derived, analyzed, and compared under different configurations of interaction objects in physical environment.
I. INTRODUCTION
6G channel maps can provide advance channel information for system design and network optimization, but conventional environment-reconstructed maps struggle to balance accuracy, update speed, and changing physical environments. The paper proposes a dynamic channel map built offline with RT and updated online with 6GPCM, then evaluates its accuracy, update efficiency, and statistical properties.
- Motivation: 6G communication scenarios combine dense nodes, high channel dimensions, and dynamic environments, increasing the need for accurate channel modeling.Advance channel information can support network design, spectrum management, and system optimization.
- Limitations of Existing Approaches: Conventional channel maps focus on static environments, while RT is computationally intensive and measurement-based maps are costly to update.These limitations make it difficult to provide both accuracy and time-efficient updates as physical conditions change.
- Proposed Approach: The proposed DCM is constructed offline with RT for static interaction objects and updated online with 6GPCM as the physical environment changes.RT-GSHCM models static and dynamic interaction objects with RT and 6GPCM, respectively.
- Evaluation: The paper evaluates DCM using an urban channel measurement campaign by comparing its channel properties and update time with RT, 6GPCM, and conventional channel maps.The evaluation includes measured results and time-cost comparisons.
- Statistical Analysis: RT-GSHCM statistical analysis covers frequency, delay, angular, temporal, and level-crossing properties under different interaction-object configurations.Reported properties include FCF, delay PSD, angular PSD, RMS delay, angular and Doppler spreads, and LCR.
B. Static CIR
The static CIR models line-of-sight and static-scattering propagation generated from the environment reconstructed by RT. Its parameters are expressed as location-dependent powers, delays, angles, and polarization-related quantities supplied or calculated for static multipath components.
- Static CIR Construction: RT generates the line-of-sight path and multipath components associated with static scatterers such as buildings, trees, and ground.The resulting electromagnetic propagation parameters are used to calculate the static CIR.
- Static CIR Construction: The static CIR separates the line-of-sight component from the non-line-of-sight component of static propagation.The two components are calculated using the static channel formulation.
- Static CIR Parameters: The static-channel formulation also includes polarization quantities, cross-polarization power ratio, co-polar imbalance, initial phases, and carrier frequency.The initial phases are uniformly distributed between [0, 2π].
- Static CIR Parameters: For static multipath components, RT provides or determines power, propagation delay, and elevation or azimuth departure and arrival angles.These parameters are indexed by transmitter and receiver antenna locations.
C. Dynamic CIR
The dynamic component uses 6GPCM to model time-varying non-line-of-sight scattering and update channel parameters online, supporting accurate, low-latency mapping as scatterers move.
- Dynamic component: 6GPCM models the dynamic component of RT-GSHCM and provides pervasiveness, applicability, and time-efficiency for DCM updates.Only the NLoS path is considered in the dynamic scattering channel.
- Dynamic CIR formulation: The dynamic CIR includes power, propagation delay, and virtual-link delay terms for time-varying multipath components.The virtual-link delay is determined from the virtual-link distance and a randomly generated exponential variable.
- Parameter optimization: Dynamic CIR parameters are optimized for environment-specific dynamic interaction objects, unlike globally optimized 6GPCM parameters.
- Online updating: RT-GSHCM updates cluster positions and velocities in real time using sensor information when vehicles or UAVs act as moving scatterers.
- Outcome: The update mechanism enables accurate and low-latency channel mapping in highly dynamic environments.
D. Channel Transfer Function (CTF)
The channel transfer function is obtained from the Fourier transform of the hybrid channel response, combining deterministic static components with stochastic dynamic scattering.
- CTF construction: The RT-GSHCM channel transfer function is obtained by Fourier transforming the time-domain channel impulse response.
- Component decomposition: The CTF comprises dynamic-scattering, line-of-sight, and static-scattering non-line-of-sight components.Antenna polarization is omitted from the CTF expressions for computational convenience.
- Static component: RT-generated static multipath components have fixed amplitudes and phases at a given transmitter-receiver location and form one equivalent deterministic component.Their vector sum is treated as composite rather than as a single physical path.
- Dynamic component: The GBSM-generated dynamic component contains many random-amplitude and random-phase paths that collectively approximate a zero-mean complex Gaussian process.The K-factor is incorporated into the deterministic and dynamic components.
- Envelope distribution: The total channel envelope follows a Rician distribution when the deterministic component is nonzero and a Rayleigh distribution when it is zero.
E. Model Assumptions and Limitations
The RT-GSHCM depends on measurement calibration and RT propagation assumptions, limiting direct generalization across frequencies and substantially different environments.
- Scope: The model’s scope requires explicit consideration of its key assumptions and current limitations when assessing applicability and planning refinement.
- Measurement dependence: Channel parameters and stochastic-component calibration depend on measurement data from limited frequency bands and specific scenarios.Without corresponding measurements, accuracy may degrade at other frequencies or in substantially different propagation environments.
- High-frequency applicability: RT relies on GO and GTD assumptions that are accurate for electrically smooth surfaces but may fail when roughness and diffuse scattering dominate above 24 GHz.Higher-order rough-surface scattering modules or additional calibration may be required.
III. DCM CONSTRUCTION
The DCM construction combines calibrated channel measurements with RT-based static scattering and 6GPCM-based dynamic scattering to build a map for the measurement scenario.
- DCM construction: The DCM construction first models static scattering with RT, estimates model parameters by comparing RT and measured multipath components, then models dynamic scattering with 6GPCM.The complete construction procedure is illustrated in Fig. 3.
- Measurement campaign: The campaign used a Keysight time-domain channel sounder with distinct transmitter and receiver setups for channel measurements.The transmitter and receiver supported 64-channel sequential measurements.
- Measurement campaign: Measurements covered LoS and NLoS routes at ChinaNetwork Valley, with building obstructions mainly causing NLoS propagation.The transmitter was fixed on the eighth floor of building B1, while the receiver array moved along routes.
- Measurement Data Process: The channel sounder was calibrated through a direct back-to-back connection before measurements to remove the setup’s inherent system response.The calibration and subsequent processing produced channel impulse responses from the measured signals.
- Measurement Data Process: The SAGE algorithm extracted multipath components from delay power spectra averaged over measurement snapshots.A threshold 6 dB above the noise floor was used to distinguish multipath components from background noise.
B. Offline Pre-construction by RT
Offline pre-construction reconstructs the measurement environment with RT and derives static-channel information for the hybrid model and later statistical analysis.
- Environment reconstruction: The ChinaNetwork Valley scenario was reconstructed in Wireless Insite using measurement-matched configurations and LoS and reflected propagation mechanisms.RT analysis limited reflections to sixth order to control computational load.
- Environment reconstruction: The RT model represented the intended 8×8 dual-polarized cylindrical array with 32 dual-polarized antenna pairs using co-located orthogonally polarized elements.Wireless Insite did not directly support setting the radiation pattern for the entire MIMO array.
- MPC identification: Joint AAoA-delay comparisons show that each RT multipath component matched at least one measured component with nearly the same power, while unmatched measured components formed dynamic clusters.These unmatched components may arise from diffraction and diffuse scattering by dynamic interaction objects omitted from RT.
- Hybrid-model parameters: The values of Kvu(ℓ), Nvu(t, ℓ), KS, and KD were obtained by identifying measured multipath components associated with different physical interaction objects.These values were used in the hybrid-model simulation settings summarized in Table III.
- Statistical analysis: The RT-GSHCM study evaluates spatial, temporal, frequency, angular, Doppler, and delay statistical properties to verify model accuracy and correctness.The analysis includes correlation functions, power spectral densities, RMS spreads, and level crossing rate.
- Statistical analysis: The dynamic-scattering space-time-frequency correlation function is further decomposed into antenna-array spatial correlation, channel-map geographical correlation, temporal autocorrelation, and frequency correlation.The decomposition is obtained by setting the relevant time, frequency, spatial, and geographic offsets to zero.
B. RMS Angular Spread and LCR
The section defines RMS angular spread from angular power distributions and describes level crossing rate as an envelope-threshold crossing measure.
- RMS Angular Spread: RMS angular spread at the receiver is the square root of the second-order central moment of the angular power spectral density.The mean angle of arrival is used in the central-moment definition.
- RMS Angular Spread: The angular power spectral density is obtained from the spatial-Doppler power spectral density through the receiver-side spatial correlation function.The spatial-Doppler frequency variable is related to the cosine of the angle between the arriving wave and receiver-array orientation.
- LCR: Level crossing rate is expressed by counting how often the channel envelope crosses a specified threshold level R.The formulation includes the error function.
C. RMS Doppler Spread
The section characterizes Doppler dispersion through RMS Doppler spread and relates frequency correlation, delay power spectrum, and RMS delay spread through Fourier-transform relationships.
- RMS Doppler Spread: RMS Doppler spread measures channel dispersion in the Doppler-frequency domain caused by motion of dynamic scatterers.It is defined using the Doppler power spectral density and average Doppler shift.
- RMS Doppler Spread: Doppler power spectral density is the Fourier transform of the temporal autocorrelation function with respect to the time interval.The Doppler frequency is denoted by ξ.
- Frequency and delay statistics: The frequency correlation function measures channel frequency correlation, while delay power spectral density is obtained by inverse Fourier transforming it with respect to frequency separation.The delay power spectral density is then used to define RMS delay spread.
V. EVALUATION AND ANALYSIS
The evaluation compares RT-GSHCM and related channel-map approaches against measurements and assesses update cost. RT-GSHCM agrees well with measured channel statistics, while DCM updates are substantially faster than conventional map reconstruction.
- Accuracy Evaluation: RT-GSHCM delay PSDs accord better with measurements than 6GPCM in Route LoS Tx1.
- Accuracy Evaluation: RT-GSHCM captures scenario-dependent channel behavior across LoS and NLoS routes and generalizes to varied propagation environments.
- Accuracy Evaluation: RMS angular spread and RMS delay spread from RT-GSHCM, RT, and 6GPCM all accord well with measurement results.
- Time-Efficiency: 874 seconds versus 0.4 seconds: CKM reconstruction takes far longer to update than DCM online updating.The comparison tracks and calculates 120 rays for CKM, whereas DCM updates channel information online with 6GPCM.
C. Channel Properties and Analysis
The RT-GSHCM analysis examines how dynamic-cluster configurations affect angular, level-crossing, Doppler, and frequency-correlation properties. Increasing cluster activity changes channel dispersion and LCR in ways consistent with the modeled physical environment and analytical results.
- RMS Angular Spread: More dynamic clusters increase large-angle MPC power, while static scattering determines the unchanged angular-spread bounds.Dynamic interaction objects provide large-angle MPCs, but the simulated dynamic-scattering range remains a subset of the static-scattering range.
- LCR: LCR changes little below the curve peak but increases with more dynamic clusters above the peak threshold.The maximum LCR also increases because more dynamic interaction objects produce more intense physical-environment changes.
- LCR: Simulated LCR values are slightly higher than analytical results, particularly near the peak region.The reported bias is associated with finite dynamic-cluster simulation and discrete sampling relative to the analytical assumptions.
- RMS Doppler Spread: RMS Doppler spread increases with dynamic-cluster speed and with the number of dynamic clusters.The paper attributes these changes to higher mobility, more frequent cluster motion, and greater Doppler-domain dispersion.
- RMS Doppler Spread: RMS Doppler spread increases with decreasing carrier frequency, and analytical results closely match simulation curves.
- FCF: FCF results accord well with analysis, supporting the correctness of the RT-GSHCM.
- Overall Analysis: The study derives and analyzes RMS delay spread, RMS angular spread, delay PSD, LCR, RMS Doppler spread, and FCF under different configurations.