Condition Number

The dominant factor degrading the condition number. When antenna elements are spaced too closely or the scattering environment is poor, the fading paths become statistically dependent.

• Mechanism: Correlation reduces the linear independence of the channel matrix rows/columns, pushing singular values toward zero.
• Impact: High correlation directly increases the condition number, reducing the number of viable spatial streams.
• Mitigation: Requires antenna spacing of at least λ/2 and a rich multipath environment.

A strong, unobstructed direct path creates a Rician fading channel with a high K-factor. While good for signal power, a dominant LOS path can be detrimental to spatial multiplexing.

• Mechanism: The LOS component adds a deterministic, non-fading element that reduces the rank of the channel matrix if the antenna array geometry is not carefully designed.
• Key Insight: A pure LOS environment can cause the channel matrix to become rank-deficient (condition number → ∞), collapsing MIMO capacity to that of a SISO link.

The physical arrangement and orientation of antenna elements directly influence the orthogonality of spatial signatures.

• Uniform Linear Arrays (ULA): Susceptible to high condition numbers when the angle of arrival is near end-fire.
• Uniform Planar Arrays (UPA): Provide better conditioning by exploiting both azimuth and elevation dimensions.
• Polarization Diversity: Using orthogonal polarizations can create well-conditioned sub-channels even with compact antenna spacing.

A dense scattering environment generates numerous independent propagation paths, which is the ideal scenario for MIMO.

• Mechanism: Rich multipath creates a channel matrix with statistically independent, identically distributed Rayleigh fading entries.
• Result: The singular values are more evenly distributed, yielding a condition number close to 1 (well-conditioned).
• Contrast: Sparse scattering environments (e.g., rural areas) lead to a few dominant paths and a high condition number.

A physical phenomenon where radio waves must pass through a narrow aperture (like a tunnel or a small gap in a building), causing all spatial sub-channels to experience the same fading.

• Mechanism: The channel matrix is modeled as a product of a transmit steering vector and a receive steering vector, resulting in a unit-rank matrix.
• Consequence: Even with rich scattering on both sides of the keyhole, the condition number becomes extremely high, and spatial multiplexing gain drops to zero.

Electromagnetic interaction between closely spaced antenna elements alters the impedance and radiation patterns of individual elements.

• Mechanism: Mutual coupling modifies the channel matrix by introducing a deterministic coupling matrix at the transmitter and receiver.
• Dual Impact: Can either improve or degrade the condition number depending on the specific array geometry and matching network design. Often intentionally exploited in compact arrays to decorrelate signals.

The mathematical factorization that decomposes a MIMO channel matrix into parallel, non-interfering eigenmodes. The condition number is defined as the ratio of the largest to the smallest singular value obtained from SVD. A high ratio indicates that some spatial subchannels are significantly weaker than others, limiting the effectiveness of spatial multiplexing. SVD-based precoding uses these singular values to optimally allocate power across eigenmodes.

The statistical dependence between antenna elements caused by insufficient spacing or a sparse scattering environment. High spatial correlation directly degrades the condition number by reducing the rank of the channel matrix. When antennas are too closely spaced or the angular spread is narrow, the singular values become highly skewed, making the matrix ill-conditioned and reducing the achievable spatial multiplexing gain.

The known channel properties—including scattering, fading, and power decay—used by a transmitter to adapt its signal. Accurate CSI is essential for computing the condition number at the transmitter. With perfect CSI, precoding algorithms can mitigate the effects of a high condition number through water-filling power allocation. Imperfect or outdated CSI leads to erroneous condition number estimates and suboptimal spatial stream allocation.

A linear MIMO detection algorithm that applies the pseudo-inverse of the channel matrix to completely eliminate inter-stream interference. The performance of ZF receivers is highly sensitive to the condition number. When the channel is ill-conditioned, the pseudo-inverse amplifies noise dramatically, a phenomenon known as noise enhancement. This makes ZF impractical for channels with high condition numbers, where MMSE or non-linear detectors are preferred.

A UE feedback parameter in MIMO systems that indicates the number of usable independent spatial layers under current channel conditions. The RI is directly influenced by the condition number. A well-conditioned channel with uniformly distributed singular values supports a high RI, enabling maximum spatial multiplexing. As the condition number increases, the effective rank decreases, and the UE reports a lower RI to avoid decoding errors on weak eigenmodes.

A UE feedback index that recommends a specific precoding matrix from a predefined codebook. The PMI selection algorithm considers the condition number to choose a precoder that best matches the channel's eigenstructure. For ill-conditioned channels, the PMI may recommend precoders that effectively concentrate power on the strongest eigenmodes, sacrificing spatial multiplexing gain for link reliability.