Authors: Lane Fuller, Robert Karl, Jr., Bruce Anderson, Max Lee-Roller
Advanced Scientific Concepts, Inc., 125 Cremona Drive, Suite 250, Goleta, CA 93117
ABSTRACT
Due to the increasing amount of Resident Space Objects (RSOs) in the Earth’s orbit, the ability to quickly and accurately extract information about them has become an important element of Space Situational Awareness (SSA). The Global Shutter Flash LiDAR (GSFL) has been proven to be an effective and highly reliable sensor to acquire multi-dimensional data about RSOs over a wide range of in-orbit distances from less than a meter to 5km, with the potential to extend this range to 50km. Various information products are available from advanced GSFL embedded processing depending on the mission needs and concept-of-operations (ConOps). This information is available on-platform in real or near real-time, or for transmission off-platform for further processing and analysis. In order to support development of the GSFL and associated system interaction, a comprehensive simulation testbed has been developed. With this tool, scientists and engineers can work through concept development and feasibility studies, mission ConOps, and system design decisions at the highest levels. The simulation testbed fits into the Model Based System Engineering (MBSE) approach in a number of ways, including functional/behavioral and performance modeling. Additionally, the simulation testbed is an important part of Machine Learning (ML) and Artificial Intelligence (AI) workflows to generate the large organized point cloud datasets needed for Deep Learning Convolutional Neural Networks (DL-CNN), for example. As a tool for software development, the simulation testbed is used for unit and regression testing during development and sustainment, with realistic real-time dynamic scenarios and accurate GSFL organized point cloud data. The simulation testbed has been constructed with layers of abstraction and developed in an object-oriented manner for continuing capability extension and refinements as its use expands.
Advanced Scientific Concepts (ASC) invented 3D GSFL imaging for space and terrestrial applications. ASC’s novel array technology has allowed for its development of compact GSFL cameras that collect a full frame 3D organized point cloud with a single short 10ns laser pulse, with no moving parts. Because a full frame of organized 3D point cloud data is natively captured with each laser flash, the real-time data is immune to motion blurring and related distortion. The organized point cloud data is in the form of an array of vectors of 3D information, consisting of x, y, z, range, and intensity {X, Y, Z, R, I}, where each vector natively maps one-to-one to its corresponding 2D time-of-flight pixel in the focal plane array (FPA) plane. Because this optimized format is native to the GSFL’s novel design, it does not require computationally expensive point cloud rectification to achieve the organized point cloud format, which is optimal for use in DL-CNN models, for example. ASC’s GSFL camera has been demonstrated for rendezvous and birthing aboard the Dragon capsule, on‑orbit testing aboard the Endeavor Space Shuttle, and in deep space operations aboard the OSIRIS‑Rex mission, achieving a Technology Readiness Level 9 (TRL-9).
The GSFL point cloud simulation testbed was designed with a layered approach, with levels of abstraction ranging from orbital and line-of-sight models with frames of reference to detector physics at the lowest level. These layers are outlined as follows:
- RSO trajectory, orbital, line-of-sight, and frames of reference modeling and visualization
- RSO CAD/mesh model libraries (i.e., NASA’s 3D Models Library)
- RSO model characteristics (i.e., area-to-mass and attitude control according to an ASO taxonomy)
- Optical Imaging System (field of view (FoV), flash laser energy, diffuser patterns, and ray tracing)
- RSO materials (bidirectional reflectance distribution function (BRDF)
