A Soft⁻Hard Combination Decision Fusion Scheme for a Clustered Distributed Detection System with Multiple Sensors.

Area of Science:

• Signal Processing
• Distributed Systems
• Information Fusion

Background:

• Distributed detection systems rely on sensors transmitting decisions to a fusion center (FC).
• Decisions are typically one-bit hard or multiple-bit soft, with trade-offs in performance and communication cost.
• Non-ideal communication channels complicate decision fusion.

Purpose of the Study:

• To propose a soft-hard combination decision fusion scheme for clustered distributed detection systems.
• To achieve a balance between communication costs and detection performance.
• To address challenges posed by non-ideal communication channels.

Main Methods:

• A clustered distributed detection system configured using fuzzy logic and fuzzy c-means clustering.
• Local sensors transmit soft decisions to cluster heads (CHs) over non-ideal channels.
• A fusion center combines CHs' hard decisions using an optimal fusion rule.

Main Results:

• The proposed scheme achieves detection performance close to centralized systems with significantly lower energy consumption.
• The system outperforms conventional distributed detection systems using only hard decisions.
• Increased bits in soft decisions improve detection performance.

Conclusions:

• The soft-hard combination fusion scheme offers an effective trade-off for clustered distributed detection.
• The approach enhances performance and energy efficiency compared to existing methods.
• Further improvements in detection are possible with higher-bit soft decisions.