A Real-Time Phase Interrogator Design of the Ф-OTDR System

Phase-sensitive optical time-domain reflectometry (Ф-OTDR) has made groundbreaking development in recent years and shows significant application potentials across various fields such as oceanographic exploration, perimeter security, and geological surveying. Contemporary Ф-OTDR systems are capable of supporting detection range up to more than one hundred kilometers, with sampling interval typically on the order of meters. This implies that the number of sensing channels of the Ф-OTDR system has reached tens of thousands to hundreds of thousands. For each sensing channel, the task of high-precision phase demodulation must be accomplished. This is a very tight computation-demanding requirement, and it is quite challenging to realize the real-time demodulation of the Ф-OTDR system. This paper proposes the first multi-channel, high-throughput real-time data processing architecture based on field programmable gate array (FPGA) platform. The designed architecture mainly includes the arctan calculation through the CORDIC algorithm, and the phase unwrapping through the unscented Kalman filtering. To support the high throughput Kalman filtering module, an improved division method based on Newton-Raphson iteration is introduced and implemented in the pipelined fashion. Besides that, a multi-channel scheduling control module is designed to involve 18 Kalman filtering modules working in parallel to swallow all raw data of the Ф-OTDR system in real time. Compared to floating-point computation results, the precision error of the calculation on FPGA is verified to be less than 0.03%. The entire design was implemented on the Xilinx VCU128 FPGA platform. Experimental results demonstrate that the proposed method can process in real time over 2500 channels (2 m sampling interval) of optical intensity data with the data width of 16 bits under the pulse repetition rate of 20 KHz (corresponding to the processing rate of 50 million per second). The proposed phase interrogator can achieve the signal-to-noise ratio (SNR) of up to 45.9 dB.