TL;DR
This paper proposes a novel selection policy for a surveillance UAV to eavesdrop on suspicious UAV communications while minimizing its own power consumption. By intelligently choosing when and how to jam, the surveillance UAV can maximize the data it intercepts, even over challenging fading channels, making it a more efficient and stealthy monitoring tool.
What Problem Does This Paper Address?
Unmanned Aerial Vehicles (UAVs) are increasingly used for various applications, but this also means malicious actors can leverage them for criminal activities, such as illicit communication. Traditional security focuses on preventing such attacks. However, there's a growing need for legitimate surveillance of these suspicious UAV communications. This paper tackles the challenge of how a surveillance UAV can effectively monitor and intercept these communications while operating under strict power constraints, which is crucial for extending its operational time. The problem is further complicated by dynamic UAV movement and the unpredictable nature of wireless communication channels (fading).
Methods
The core of this paper is a proposed "selection policy" for a legitimate surveillance UAV (UAV₁) to eavesdrop on suspicious UAV communications (UAVST to UAVSR). The policy aims to optimize two objectives: minimizing UAV₁'s power consumption and maximizing the amount of intercepted data.
Here's a breakdown of the approach:
- Four Surveillance Scenarios: The paper defines four distinct scenarios for how UAV₁ can eavesdrop and jam, based on which suspicious UAV (transmitter or receiver) it targets for jamming and which it targets for eavesdropping.
- Power Consumption Model: A model is developed to analyze the power consumption of UAV₁ for both eavesdropping and jamming activities. This is critical because UAVs have limited battery life.
- Optimization Problem Formulation: An optimization problem is formulated to maximize the eavesdropped data packets (bytes) under several constraints:
- Maintaining a minimum Signal-to-Interference-plus-Noise Ratio (SINR) for the suspicious link to ensure it's not immediately detected.
- A total power budget for UAV₁.
- A maximum power limit per time slot for UAV₁.
- Selection Policy Algorithm: A real-time selection policy is proposed. At each time slot, UAV₁ assesses the channel conditions and its position to select the most efficient eavesdropping/jamming strategy from the four scenarios. This involves dynamically allocating jamming power. The policy aims to find the optimal jamming power that maximizes eavesdropped data while staying within power constraints.
- Fading Channel Analysis: The performance of the selection policy is evaluated across four common fading channel models: Rayleigh, Ricean, Weibull, and Nakagami, to understand its robustness in different wireless environments.
Datasets
Not discussed in the paper. The paper relies on simulated data generated based on the defined system model and parameters.
Key Insights and Ideas
- Proactive Eavesdropping is Key: Instead of just passively listening, the paper advocates for an active approach where jamming is used strategically to facilitate eavesdropping by degrading the suspicious link's quality.
- Dynamic Power Allocation is Crucial: A static jamming strategy is inefficient. The proposed policy dynamically adjusts jamming power based on real-time conditions, leading to significant power savings and improved eavesdropping rates.
- Location Matters: The relative positions of the UAVs significantly impact channel gains and, consequently, the effectiveness of eavesdropping and jamming. The policy accounts for this dynamic mobility.
- Fading Channels Impact Performance: While the selection policy generally performs well, the type of fading channel can influence the amount of data intercepted. Weibull channels, often associated with dominant line-of-sight propagation, appear to be more favorable for this eavesdropping strategy.
Key Results
- Power Savings: The proposed selection policy demonstrates significant power savings compared to constant jamming or no-jamming strategies. For instance, at a distance of 500m, it saves up to 65.79% more power than constant jamming and 74.73% more than no jamming.
- Improved Eavesdropping Rate: The policy effectively increases the number of eavesdropped packets. At 500m, it outperforms other methods by nearly 1.2 times.
- Performance Across Fading Models: The policy shows consistent performance across different fading models, with a slight edge in Weibull fading channels. Nakagami fading channels resulted in fewer eavesdropped packets, likely due to their more complex noise characteristics.
- Time Complexity: The selection policy's time complexity is analyzed as O(n²m + nm), where 'n' is the number of cases (4 in this study) and 'm' is the number of time slots. This is considered acceptable for real-time application.
Limitations and Future Work
- Limited Simulation Time Slots: Due to laboratory equipment limitations, simulations were conducted with only six time slots. The authors suggest future work to research algorithms for determining the optimal number of time slots for accurate problem resolution.
- Two-Dimensional Model: The current distance model is two-dimensional. Future work could extend this to three dimensions to incorporate more realistic UAV movement.
- Suspicious UAV Groups: The paper focuses on a single suspicious communication link. Future research plans to address jamming and eavesdropping towards suspicious UAV groups, which presents a more complex challenge for policy selection.
- Practical Implementation: While simulated, the authors note that real-world experiments would require actual equipped UAVs.
Summary updated