Small solar system bodies are remnants of planetary formation and collisional history, preserving conditions in unique, mostly pristine settings [1
]. The origins, evolution, and present dynamical behavior of small bodies are closely tied to their present geophysical properties including shape, structure, and coherence, spin state, and active surface processes. In addition, near-Earth objects (NEOs) may pose a threat of Earth impact, and a better understanding of their interior structure and dynamics will aid in developing mitigation strategies [4
]. Recent space missions to small bodies have carried one or more lidars to measure their shape, perform gravity investigations, and to aid in navigation for precision surface sampling. The Near-Earth Asteroid Rendezvous (NEAR) mission carried a laser rangefinder to obtain high-resolution topographic measurements aimed at constraining the shape and internal structure of the asteroid 433 Eros [5
]. The Hayabusa mission to asteroid 25143 Itokawa [6
] and the Haysabusa2 mission to asteroid 162173 Ryugu [7
] both carried short-range laser altimeters to performance guidance and navigation functions and aid in landing. Recently, the OSIRIS-REx mission to asteroid 101955 Bennu carried two lidars: a scanning lidar for mapping and shape determination [8
] and a flash lidar for navigation and landing [9
]. A single lidar used for these missions must operate over a wide dynamic range, providing long-distance Reconnaissance as well as near-range support for landing and surface sampling. We have developed a new type of planetary lidar, namely the small all-range lidar (SALi), to support all mission phases [10
]. SALi is a swath-mapping orbital laser altimeter that uses a pseudo-noise (PN) code modulated fiber laser to produce a laser pulse train at 1550 nm that is transmitted toward the target surface [11
]. Laser photons scattered from the target are collected by a 6.4-cm off-axis parabolic telescope. The received light passes through a bandpass filter and is focused onto a 2 × 8-pixel HgCdTe avalanche photodiode (APD) array detector. Both the laser power and the detector gain are adjustable, which enables ranging over five orders of magnitude in range (i.e., meters to 100s of km). To our knowledge, no other lidar has been designed to operate over such a wide range of target distances. The SALi instrument requirements as well as detailed instrument and subsystem descriptions are presented by Sun et al. [11
In the SALi design, the fiber laser is modulated with a pseudo noise (PN) code. PN codes are repeating binary patterns widely used in spread spectrum communication and ranging [12
]. PN codes have the property that the circular autocorrelation of the sequence results in a Kronecker delta function when the codes are perfectly aligned and a small constant value elsewhere [14
]. PN code laser ranging has been studied and implemented on the laboratory scale for applications including laser altimetry [15
], aerosol backscattering measurements [18
], and time-of-flight imaging systems [21
]. Our approach uses a return-to-zero (RZ) modification to the PN code in which the laser pulse only occupies a fraction of the bit duration, after which the transmitted signal and the correlation kernel are both zero for the remainder of the bit duration [24
When using PN-code modulation, the receiver performs a cross-correlation of the return signal to determine the lidar range. This requires significantly more signal processing compared to single-pulse range measurements. For small body space missions, this data processing must be completed onboard in real-time since it is impractical to downlink the raw data from the receiver for ground processing. Our earlier laboratory implementations of RZPN lidar post-processed the data in software, not in real time [15
]. Prior to this work, real-time digital signal processing of the RZPN data has not been demonstrated. For space missions, there are also signal characteristics such as Doppler shift when ranging to a moving target, pulse broadening, and the effects of the signal sampling rate that have not been previously explored.
Here, we describe the performance simulations for a PN-code lidar, as well as the design and implementation of real-time correlation processing of an PN-coded pulse train using a field-programmable gate array (FPGA). First, we will describe the PN code parameters and considerations for the FPGA implementation of the FFT-based processing approach. We also describe PN code and SALi performance simulations. These were developed to determine the maximum achievable range and performance during various instrument modes, the effects of Doppler shift on the cross-correlation, and to test the method used to estimate surface reflectance. The performance simulation results include range error, signal-to-noise ratio (SNR), reflectance error, and impacts from Doppler shift. Although the performance model and simulations we report here use the SALi instrument parameters, the principles they show are generally applicable to other lidar using PN code modulation. We also performed a PN code ranging simulation using a high-resolution shape model of the asteroid 101955 Bennu as a representative small body target. Finally, we describe the results of implementing the real-time correlation processing using a Xilinx KU060 FPGA.
We have described simulation studies and the implementation of a real-time processing approach for an PN code lidar (SALi) for planetary missions to small bodies. SALi uses a return to zero PN code and is designed to operate as a mapping and survey lidar, with a single-beam Reconnaissance mode out to 400 km and a 16-pixel Mapping mode to 50 km. In addition, it can support descent and surface sampling by providing ranges at a 100 Hz measurement rate with 16 channels.
Simulations of the lidar performance quantified the anticipated performance. In the long-range Reconnaissance mode, SALi obtained accurate ranges to 440 km with a precision of 8 cm below 400 km. In Mapping mode, accurate ranges were retrieved up to 90 km in altitude with a ranging precision of 5 cm at altitudes of 50 km and lower. We also developed a reflectance algorithm using active laser control to determine the surface reflectance to a precision of 3% of the surface reflectance values from 4% to 36%. These simulations have also uncovered several relevant PN code design aspects that have not been previously described in detail. These include the binary aspect of the range error, the correlation acting as a low-pass filter, the FFT length considerations, the Doppler sensitivity limits, and the quantitative reflectance measurement method. Finally, the PN ranging algorithm was tested using simulated measurements of the surface of asteroid 101955 Bennu. The range residual standard deviation was 5 cm with a sub-mm bias.
The FFT-based processing algorithm implemented on a Xilinx Ultrascale FPGA performed the required cross-correlation range measurement at a maximum throughput of 3050 Hz with a latency of 1.07 ms for a single channel. Based on these results, the SALi PN code approach was shown to be feasible as a high-measurement-rate ranging and reflectance lidar well-suited for future missions to small bodies.