Link-Level Simulation

The primary metric of link-level simulation is Block Error Rate (BLER)—the probability that a transport block is decoded incorrectly after all physical layer processing. The simulator models the complete chain: CRC attachment, channel coding (LDPC or Polar codes in 5G), rate matching, scrambling, modulation mapping, MIMO precoding, resource element mapping, and OFDM waveform generation. At the receiver, the inverse operations are performed, including channel estimation, equalization, LLR computation, and iterative decoding. A transport block is declared in error if any bit remains incorrect after decoding, making BLER the ground truth for link adaptation decisions.

Link-level simulators implement bit-exact models of modern channel codes. For 5G NR, this includes Low-Density Parity-Check (LDPC) codes for data channels and Polar codes for control channels. The simulator models the belief propagation decoder or successive cancellation list decoder with configurable iteration limits. Modulation orders from QPSK to 256-QAM are simulated with exact constellation mapping. Mutual information-based abstract models are often used to accelerate simulations by predicting decoder performance without running the full iterative decoding chain, trading a small accuracy loss for orders of magnitude speedup.

Link-level simulation captures the spatial dimension of modern wireless systems through detailed Multiple-Input Multiple-Output (MIMO) modeling. This includes:

• Precoding matrix selection from standardized codebooks (Type I and Type II CSI)
• Spatial multiplexing with up to 8 layers in 5G NR
• Transmit diversity schemes like Space-Frequency Block Coding (SFBC)
• Beam management procedures including initial acquisition, beam refinement, and beam failure recovery
• Antenna array modeling with configurable element spacing, polarization, and element radiation patterns The simulator computes the post-processing SINR per layer, which directly determines the achievable spectral efficiency.

Realistic receiver performance depends critically on channel estimation accuracy. Link-level simulators model the insertion of demodulation reference signals (DMRS) at specific resource element positions and implement estimation algorithms such as Least Squares (LS), Minimum Mean Square Error (MMSE), or Deep Neural Network-based estimators. Time and frequency synchronization errors—including carrier frequency offset, sampling clock offset, and symbol timing offset—are explicitly injected to evaluate receiver robustness. The simulator also models phase noise from local oscillators and IQ imbalance in the RF front-end, capturing hardware impairments that degrade real-world performance.

Link-level simulation is driven by a channel model that generates the complex impulse response between transmitter and receiver. Standardized models include Tapped Delay Line (TDL) and Clustered Delay Line (CDL) profiles from 3GPP TR 38.901, which define power delay profiles, angular spreads, and Doppler spectra for various scenarios (Urban Micro, Urban Macro, Indoor). The simulator convolves the transmitted waveform with the time-varying channel impulse response, adding Additive White Gaussian Noise (AWGN) at the configured SNR. Fast fading is generated using the sum-of-sinusoids or filtered Gaussian noise methods to produce statistically correct temporal correlation.

A critical function tested in link-level simulation is the closed-loop link adaptation process. The UE measures the downlink channel quality and reports a Channel Quality Indicator (CQI), Rank Indicator (RI), and Precoding Matrix Indicator (PMI). The simulator models the measurement gap, reporting delay, and quantization error inherent in this feedback. The gNB scheduler uses these reports to select the Modulation and Coding Scheme (MCS) for the next transmission. The simulator evaluates whether the selected MCS achieves the target BLER, quantifying the efficiency loss from outdated or quantized feedback.

A higher-abstraction simulation methodology that models a multi-cell network with numerous user equipment (UEs) to evaluate resource management, scheduling, and overall network performance metrics. Unlike link-level simulation, which focuses on a single point-to-point link, system-level simulators abstract the physical layer to manage computational complexity. They are essential for testing MAC schedulers, handover algorithms, and inter-cell interference coordination across hundreds of cells.

A deterministic propagation modeling technique that simulates the paths of individual radio waves by calculating reflections, diffractions, and scattering based on a precise 3D geometric environment. Ray tracing generates highly accurate, site-specific channel parameters—including path loss, angle of arrival, and delay spread—that can be fed into a link-level simulator for realistic performance evaluation in a specific urban or indoor deployment scenario.

A simulation technique where a physical hardware component, such as a gNB modem or UE chipset, is integrated into a real-time virtual simulation environment. The link-level simulator generates the digital baseband signal, which is then converted to analog RF and fed into the device under test. HIL bridges the gap between pure software simulation and over-the-air testing, enabling validation of real silicon against complex, repeatable channel scenarios.

A simulation-based methodology that replaces physical drive tests by emulating network conditions and user mobility in a lab. It combines a propagation model, a user mobility model, and link-level or system-level simulation to recreate a virtual route. This allows engineers to validate handover performance, throughput, and call drop rates across a city-wide network without deploying a fleet of measurement vehicles.