LAUSR

The saturated waveform is a main limiting factor of dynamic detection range of light detection and ranging (LiDAR), and thus, extracting the range from saturated waveform accurately can extend the detection range of full-waveform LiDAR. A Gauss–Newton online ranging method based on saturated waveform compensation (GNC) was proposed, including rough estimation and optimized estimation of waveform parameters. To realize the rough estimation, the saturated echo waveform needs to be compensated using the Gaussian mixture model by scaling in amplitude and delaying in time of emitted waveform. A scale factor was estimated by piecewise fitting of the corrected saturation length, and a delay factor can be quickly estimated by inexact line search. To extract the optimized parameters, a Gauss–Newton method was used to optimize rough parameters based on the compensated waveform. To measure the distance of the target in real time, the GNC method was implemented on field-programmable gate array (FPGA) (ZYNQ-7000) using a pipeline structure, taking the execution time of $7.87 \mu \text{s}$ for one saturated echo. A series of ranging experiments was conducted to evaluate the ranging performance of GNC for saturated waveforms. The experimental results showed that the ranging standard deviation (RStD) of GNC implemented on FPGA was less than 2.5 cm, and the absolute of mean ranging error (MRE) was less than 1.9 cm for 0%–165% saturation. The proposed method provides an effective way to enhance the dynamic detection range with high precision.