Abstract
Wireless communication is inherently vulnerable to passive eavesdropping and active packet injection due to the open nature of the medium. While traditional security relies on upper-layer cryptography, these methods are computationally bounded and do not address physical layer vulnerabilities. This paper presents MIO (Multiple Inter-Symbol Obfuscation), a novel physical layer security (PLS) scheme that injects artificial noise directly into baseband symbols. By combining dynamic key updates with a unique power normalization technique, MIO achieves information-theoretic secrecy and robust integrity checking without requiring complex multi-antenna hardware or static channel conditions.
Introduction
The ubiquity of wireless devices has made securing the physical layer a priority. Traditional methods like AES provide computational secrecy but are susceptible to advances in processing power. Existing PLS techniques often rely on "artificial noise" but typically require multiple-input multiple-output (MIMO) systems or specific channel characteristics to be effective. MIO addresses these limitations by providing a practical, single-antenna solution that "cloaks" data symbols, making them indistinguishable from noise to unauthorized parties.
Methodology: The MIO Architecture
The core of the MIO scheme involves obfuscating baseband data symbols using a "symbols key" through four primary stages:
- Symbol Obfuscation: MIO adds a random complex vector (a key symbol) to a subset of data symbols during modulation.
- Power Normalization: To prevent eavesdroppers from identifying encrypted symbols via power analysis, MIO applies a normalization factor ($\theta$). This scales the obfuscated symbols so their average power matches regular data symbols, effectively hiding them in the constellation map.
- Dynamic Key Update: To ensure long-term security, MIO updates the symbols key for every packet. The next key is generated by hashing a random subset of successfully decrypted symbols from the previous packet. This creates a dependency chain that an attacker cannot easily break.
- Decryption and Integrity Checking: The legitimate receiver uses cross-correlation with its local key to locate and extract the obfuscated symbols. This process inherently validates the packet; if an attacker injects a fake packet, the lack of correlation causes the receiver to drop it immediately.
Experimental Evaluation
The authors evaluated MIO through both a hardware testbed and extensive simulations:
- Hardware Testbed: Five USRP2 nodes running GNURadio on Ubuntu were used to simulate real-world wireless environments.
- Simulations: MATLAB/Simulink was used to test MIO across various modulation schemes, including QPSK, 16-QAM, and 64-QAM.
Key Results
- Eavesdropper Suppression: The Bit Error Rate (BER) for eavesdroppers remained between 0.25 and 0.5 regardless of the Signal-to-Noise Ratio (SNR), rendering intercepted data useless.
- Legitimate Performance: The legitimate receiver experienced a minor power penalty (1.0 to 3.7 dB) but maintained reliable communication.
- Robust Integrity: At SNRs of 11dB or higher, MIO achieved a 0% false positive/negative rate in identifying obfuscated symbols, effectively neutralizing fake packet injection attacks.
Discussion and Insights
MIO’s primary strength lies in its "cloaking" effect. By normalizing power, it removes the statistical signatures that eavesdroppers typically exploit. Furthermore, MIO's security is location-independent; unlike some PLS schemes, an eavesdropper gains no advantage by being physically closer to the transmitter.
However, the system does face challenges. It requires an initial bit-level key exchange (e.g., Diffie-Hellman) to bootstrap the process. Additionally, the dynamic update mechanism is sensitive to packet loss, which can lead to desynchronization between the sender and receiver.
Conclusion
MIO represents a significant step forward in practical physical layer security. By moving obfuscation to the symbol level and implementing dynamic updates, it provides a robust defense against both passive and active attacks on standard, single-antenna wireless hardware.
Summary updated