Sub-meter wind detection with pulsed coherent Doppler lidar

High-resolution wind detection plays a crucial role in aviation safety and aerodynamic research. A pulsed coherent Doppler wind lidar (CDWL) with sub-meter/sub-second resolution is demonstrated. The pseudorandom modulation (PRM) method is used to break the link between laser pulse duration and spatial resolution. Benefiting from the flexible pulse duration, the detection accuracy degradation related to the short pulse is largely mitigated. The backscattered spectra and wind profiles measured by the proposed lidar are compared with those obtained by a conventional pulsed CDWL, validating its capacity of radial wind detection within 700 m at 0.9 m/0.5 s resolution. With the help of the proposed high-resolution lidar, the meter-scale perturbation on the wind field from an electric fan is detected in a field experiment.

Figure 1

System layout of the high-resolution lidar. CW, continuous-wave laser; BS, beam splitter; Cir, circulator; and PC, personal computer.

Figure 2

(a) Eye diagram and intensity envelope of the modulated lightwave in the PRM method. (b) Enlarged eye diagram of the modulated lightwave. (c) Noise spectrum of the balanced photodetector.

Figure 3

(a)–(h) Continuous measurement results in four seconds. Left axis: radial wind velocity profiles of the two lidars; right axis: spectrum width of the CP lidar.

Figure 4

Backscattered spectra at 350 m and 480 m measured by the (a), (b) PRM and (c), (d) CP lidars continuously.

Figure 5

Illustration of the experiment location and instruments layout.

Figure 6

(a) Continuous radial wind profiles in 4 s with spatial/temporal resolution of 0.9 m/0.5 s. (b) Radial shear intensity of the wind field. (c)–(j) Zoomed wind profiles between 315 m and 345 m, where the perturbation from the fan occurs around 329 m.

Figure 7

(a)–(h) Backscattered spectra around the fan continuously measured in four seconds.

Figure 8

(a) Narrowband CNR and corresponding CRLB. (b) Spectrum of the observed wind field, where the noise floor is around $3.9\times {10}^{-3}{\left(\mathrm{m}{\mathrm{s}}^{-1}\right)}^2/\left(\mathrm{rad}{\mathrm{m}}^{-1}\right)$.