Encrypted Physical Layer Communications Using Synchronized Hyperchaotic Maps

With the proliferation of Internet-of-Things (IoT) devices at public facilities, work places, homes, and beyond, securing data communications is becoming increasingly challenging. This paper describes the analysis and practical implementation of synchronized hyperchaotic maps for securing short-range data links in IoT devices. The data is stream-encrypted in the physical layer using chaotic masking and decrypted using a synchronized chaotic map. Two different energy-efficient chaotic encryption schemes are proposed: 1) direct sampling and masking of analog sensor outputs, thus avoiding the need for analog-to-digital (ADC) conversion; and 2) bit scrambling and masking of digital data. Both schemes were initially investigated and simulated in MATLAB. The effective number of encryption keys when using such hyperchaotic maps was also studied. Simulation results were validated by implementing two maps on circuit boards using high-speed discrete components. Experimental results for digital communication show a bit error rate (BER) of
$\approx 2\times 10^{-6}$
at a bit rate of 10 kbps and a clock frequency of 0.5 MHz, making the approach feasible for high-fidelity real-time speech and image transmission without additional error control coding. An energy-efficient on-chip chaotic encryption system was developed using analog current-mode circuits and has a simulated power consumption of
$< 100~\mu \text{W}$
in the TSMC 180 nm CMOS process, which is sufficiently low for use in many battery-less miniaturized IoT sensor nodes.