Spatial Diversity

The fundamental principle enabling spatial diversity is that multipath fading is a localized phenomenon. When sensing nodes are separated by a sufficient distance—typically greater than the coherence distance of the channel—they experience statistically independent signal fades.

• A deep fade at one node is unlikely to occur simultaneously at another
• The probability of all nodes experiencing a fade simultaneously decays exponentially with the number of nodes
• This transforms the fading channel from a detection bottleneck into a diversity advantage

Spatial diversity directly addresses the hidden node problem, where a single cognitive radio may be shadowed by a building or terrain feature and fail to detect a transmitting primary user.

• A receiver in a deep shadow relative to the primary transmitter will report a false 'spectrum vacant' decision
• Geographically distributed nodes provide multiple observation perspectives around obstacles
• The fusion center combines these diverse views to reconstruct an accurate picture of spectrum occupancy
• This is the primary motivation for cooperative sensing over standalone detection

The diversity gain quantifies the improvement in detection reliability as the number of independent sensing branches increases. In a cooperative sensing network with N spatially diverse nodes:

• The probability of missed detection decreases exponentially with N under soft decision fusion
• Hard decision fusion with a K-out-of-N rule provides a flexible tradeoff between sensitivity and false alarm rate
• Even with only 2-3 well-separated nodes, the diversity gain can overcome noise uncertainty limitations that cripple single-node energy detection
• The gain is maximized when node placements are optimized to minimize correlated shadowing

The diversity benefit degrades when nodes experience correlated shadowing—similar large-scale fading due to common obstructions. This occurs when nodes are clustered too closely together.

• The decorrelation distance for shadowing in urban environments is typically 10-50 meters
• Nodes must be separated beyond this distance to achieve independent fading realizations
• Cluster-based CSS architectures intentionally distribute cluster heads to maximize inter-cluster diversity
• Correlated shadowing can reduce the effective diversity order from N to a much smaller number, undermining the cooperative gain

Spatial diversity is one of several diversity techniques, each exploiting a different domain to create independent signal paths:

• Temporal diversity: Uses interleaving and coding across time slots separated by the channel coherence time
• Frequency diversity: Spreads the signal across multiple frequency bands exceeding the coherence bandwidth
• Polarization diversity: Exploits orthogonal polarization states that fade independently
• Spatial diversity is uniquely suited to cooperative sensing because it requires no additional spectrum or time resources—only the geographic distribution of existing sensing nodes

At the physical layer, spatial diversity is closely related to Multiple-Input Multiple-Output (MIMO) antenna systems. Each sensing node may itself employ multiple antennas to provide an additional layer of diversity.

• Antenna spacing of at least λ/2 (half-wavelength) is required for uncorrelated signals at each element
• Combining node-level spatial diversity with antenna-level diversity creates a two-tier diversity architecture
• The fusion center can exploit both the inter-node and intra-node diversity for maximum detection reliability
• This approach is particularly powerful in wideband spectrum sensing where frequency-flat fading assumptions break down

A propagation phenomenon where sensing nodes in close physical proximity experience similar large-scale signal fading. This correlation degrades the spatial diversity gain expected from cooperative sensing, as nodes no longer provide independent observations. The spatial correlation coefficient, often modeled exponentially with distance, dictates the minimum separation required for effective diversity. Ignoring this effect leads to overly optimistic performance predictions.

A central processing node that collects local observations or decisions from geographically distributed sensing nodes and applies a fusion rule to make a global decision about spectrum occupancy. The fusion center is the architectural lynchpin that transforms spatial diversity into detection gain. It must handle imperfect reporting channels and may employ weighted combining strategies to prioritize nodes with higher instantaneous signal-to-noise ratios.

A bandwidth-efficient fusion strategy where sensing nodes transmit a binary local decision ('1' for occupied, '0' for vacant) to the fusion center. The center then applies a voting rule like the K-out-of-N rule to declare a global decision. While simple and low-overhead, this method discards the confidence level of each local observation, sacrificing some of the potential diversity gain for communication efficiency.

A high-fidelity fusion strategy where sensing nodes transmit raw or quantized test statistics (e.g., energy levels) to the fusion center. This preserves the full richness of the spatially diverse observations, allowing the center to apply optimal combining algorithms like Weighted Gain Combining. Soft fusion achieves superior detection sensitivity at the cost of higher reporting channel bandwidth, representing the upper bound of cooperative diversity performance.

The fundamental challenge that spatial diversity solves. A single cognitive radio may be shadowed by a building or terrain, rendering it unable to detect a primary user transmitter. This creates a false spectrum hole and leads to harmful interference. By deploying multiple spatially separated sensing nodes, the probability that all nodes are simultaneously shadowed drops exponentially, dramatically improving primary user protection.

A fully decentralized cooperative sensing approach that eliminates the single point of failure of a fusion center. Nodes iteratively exchange information only with their neighbors and run a consensus algorithm to converge on a common global decision. This architecture leverages spatial diversity while providing robustness against node failure and is well-suited for ad-hoc cognitive radio networks without fixed infrastructure.