A GPU-based phase tracking method for planetary radio science applications

This paper introduces a phase tracking method with a computational algorithm for planetary radio science implemented on NVIDIA GPUs. In contrast to the phase-locked loop phase counting method used in traditional Doppler data processing, this method fits the tracking data with optimal parameters into the shape expressed by Taylor expansion. The differential evolution (DE) algorithm is employed for polynomial fitting. In order to cope with the high computational intensity of the proposed phase tracking method, graphics processing units (GPUs) are employed. The method estimates the instantaneous phase, the frequency, the derivative of frequency (line-of-sight acceleration) and the total count phase for different integration scales. These observables can be further used in planetary radio sciences. The new method was tested on the Mars Express (MEX, ESA) and Chang'e 4 relay satellite (China) tracking data. In a real experiment with 400 K data block size and  ∼80 000 DE solver objective function evaluations, we were able to achieve the target convergence threshold in 6.5 s, executing real-time processing on NVIDIA GTX580 and NVIDIA K80 GPUs, respectively. The instantaneous Doppler precision for occultation research of this method is higher than the result of a traditional Doppler method. The precision of the integral Doppler (60 s integration) was 2 mrad s⁻¹ and 4 mrad s⁻¹ for MEX(3-way) and the Chang'e 4 relay satellite(3-way) respectively, which is equal to traditional Doppler results. (paper)