Distributed Hybrid Two-Stage Multi-Sensor Fusion for Cooperative Modulation Classification in Large-Scale Wireless...

Area of Science:

• Electrical Engineering
• Signal Processing
• Wireless Communications

Background:

• Cooperative modulation classification (MC) using multiple sensors improves performance over single sensors.
• Centralized fusion of features or decisions in cooperative MC can lead to information loss due to unreliable quality measures.
• Existing cooperative MC methods suffer performance degradation because of the non-cooperative nature and data fusion challenges.

Purpose of the Study:

• To propose a distributed two-stage fusion concept for cooperative modulation classification.
• To improve the performance of cooperative MC by preserving information during the fusion process.
• To offer a flexible and scalable solution for large-scale wireless networks.

Main Methods:

• A distributed two-stage fusion approach combining feature (cumulant) fusion and decision fusion.
• Utilizing a clustered architecture to restrict the influence of mismatched references.
• Employing multiple uncorrelated signal observations for enhanced information gathering.

Main Results:

• The proposed distributed fusion significantly improves cooperative MC performance.
• Information preservation during fusion is facilitated by the distributed approach.
• The clustered architecture effectively manages data fusion within intra-cluster groups.

Conclusions:

• The proposed distributed two-stage fusion is a superior method for cooperative modulation classification.
• This approach mitigates information loss inherent in centralized fusion techniques.
• The flexible and scalable design is well-suited for large-scale network implementations.