Wind Profile Tracking in MST Radar Using Viterbi Data Association

Abstract

The mesosphere–stratosphere–troposphere (MST) radar is a very high frequency (VHF) pulsed coherent Doppler radar used for ground-based remote sensing. In an MST radar, the backscattered pulses due to fluctuations in atmospheric refractive index are observed at the receiver. Echoes from higher altitudes are typically weaker, resulting in a low signal-to-noise ratio (SNR). They could also be contaminated with clutter and interference, making conventional spectral estimation techniques in the Fourier domain unreliable. In this article, we propose an algorithm to estimate the atmospheric wind profile using a dynamic programming approach employing the Viterbi data association (VDA) algorithm. Progressively for each range bin, the proposed algorithm chooses a set of probable wind speeds as nodes in a trellis. A branch metric incorporating the kinematic parameters inclusive of the differential wind shear is then constructed for state transitions within the trellis, from nodes in one range bin to the subsequent ones. A dynamic programming approach then chooses the optimal wind profile across altitudes. The proposed algorithm is experimented on simulated as well as real MST radar data sets to verify its robustness. Our experiments show that the suggested method has improved performance when compared to previously established techniques.