[Paper Review] Towards a disaster response system based on cognitive radio ad hoc networks

Noman Islam, Ghazala Shafi Shaikh|arXiv (Cornell University)|Oct 3, 2017

Mobile Ad Hoc Networks13 references3 citations

TL;DR

This paper proposes a cognitive radio ad hoc network (CRAHN)-based disaster response system integrating artificial neural network (ANN)-driven disaster detection, ANN-based spectrum sensing, and a service discovery mechanism for emergency coordination. Evaluated in NS-2, the system achieves a low false negative alarm rate, fast spectrum switching, and reduced service discovery latency, demonstrating feasibility for resilient disaster management communications.

ABSTRACT

This paper presents an approach towards disaster management based on cognitive radio ad hoc network. Despite the growing interests on cognitive radio ad hoc networks, not much work has been reported on using them for disaster management. This paper discusses opportunities for disaster management based on cognitive radio ad hoc networks. In this direction, the paper presents a novel technique for disaster detection based on Artificial Neural Network (ANN). The ANN is trained using backward propagation algorithm. An ANN-based spectrum sensing scheme is also presented. Finally, a service discovery scheme is presented for coordination during the time of disaster. The simulation of proposed approach has been performed in NS-2 simulator. The proposed approach shows very low false negative alarm rate using the proposed disaster detection system. The spectrum switching time of spectrum sensing scheme is also analyzed along with an analysis of latency of proposed service discovery scheme

Motivation & Objective

• To address the lack of integrated, resilient communication systems in disaster scenarios.
• To enable rapid, adaptive communication in post-disaster environments where infrastructure is damaged.
• To develop a cognitive radio-based ad hoc network that supports real-time disaster detection and resource coordination.
• To reduce false alarms in disaster detection using machine learning.
• To optimize spectrum access and service discovery for emergency response efficiency.

Proposed method

• An artificial neural network (ANN) with backpropagation training is used for disaster detection using environmental sensor data.
• A separate ANN-based spectrum sensing scheme enables dynamic spectrum access by detecting available channels.
• A service discovery protocol is designed to locate and coordinate emergency services and resources in the ad hoc network.
• The system is simulated in NS-2 to evaluate performance under disaster conditions.
• Spectrum switching time and service discovery latency are measured as key performance indicators.
• The ANN models are trained on historical disaster and spectrum data to improve detection and sensing accuracy.

Experimental results

Research questions

• RQ1 How can cognitive radio ad hoc networks be effectively leveraged for disaster response communication?
• RQ2 Can an ANN-based approach achieve low false negative rates in disaster detection?
• RQ3 What is the performance of ANN-based spectrum sensing in terms of switching time and accuracy?
• RQ4 How can emergency services be efficiently discovered and coordinated in a decentralized, infrastructure-less network?
• RQ5 What are the latency and reliability characteristics of the proposed service discovery mechanism?

Key findings

• The proposed disaster detection system achieved a very low false negative alarm rate, indicating high reliability in identifying disaster events.
• The spectrum sensing scheme demonstrated fast switching times, enabling rapid adaptation to available spectrum bands.
• The service discovery scheme exhibited low latency, supporting timely coordination of emergency resources.
• Simulation results in NS-2 confirmed the feasibility and robustness of the integrated CRAHN-based disaster response framework.
• The ANN-based spectrum sensing method effectively identified spectrum opportunities, enhancing spectral efficiency.
• The overall system performance supports deployment in real-world disaster scenarios requiring resilient, adaptive communication.