Results for “Augmented Reality” — Page 9

NASA NTRSDate not supplied

The Gandalf Staff: A Mobile Tool for Lunar Exploration (2nd year of development)

The Gandalf Staff is a mobile tool designed to be a flexible device supporting crewed and uncrewed operations on the lunar surface. The core of the device is a 24v battery with communications and data storage systems. Initial optional components supporting crewed Extra-Vehicular Activity (EVA) include a LiDAR and 360˚ camera. These provide 3D mapping of the traverse for documentation, and to aid future planning. The mapping also creates outreach opportunities for the public to “stand beside” the astronaut in Virtual Reality (VR). The staff provides external lighting for field site illumination in the south polar region low sun angle environment. Navigation instruments for crew position determination with Lunar Search and Rescue (LunaSAR) are also included. The staff itself can be used as a walking aid or as a splint for Incapacitated Crew Rescue (ICR). As a stand-alone device, the staff operates as a long duration untended science platform collecting environmental data and sending it to a lunar base station. The stand-alone mode requires connection to an auxiliary power source (e.g. solar array) and energy storage system (e.g. battery), so it could become an electrical recharging station. To make rapid progress in the 1st year, and also to demonstrate innovative project management techniques, NASA guided a private industry partner, T STAR, in leading Capstone Engineering student teams at Texas A&M University (TAMU) for proof-of-concept development and testing. These teams developed the power system and demonstrated successful integration of LiDAR, WiFi communications, and external lighting subsystems. Another industry team at Jacobs Technology prototyped a tripod to hold the staff upright. For the 2nd year (FY’22), NASA will again collaborate with partners to prototype enhanced power and lighting concepts. Year 2 will also add new capability for LunaSAR and geophysical science instrumentation using a heat probe. The heat probe is based upon Apollo heritage but modified to measure subsurface volatile ice regimes at the Artemis landing site. Components of the Gandalf Staff can be developed, tested, and deployed independently, or on the integrated staff, rovers, or utility trailers. The project supports crew safety, lunar sample curation, mission science, and public outreach goals of NASA. PLAIN TEST SUMMARY: Gandalf Staff is a 24v battery powered mobile tool for the lunar surface supporting crewed geologic field site exploration, or as a stand-alone science platform. As a tool assisting astronauts conducting field geology, the Gandalf Staff provides external lighting to illuminate shadowed regions on the surface. It also documents the process of collecting rocks and dust samples using a LiDAR (laser range instrument) and cameras. The staff provides a beacon for emergency location of astronauts and augments communications when the astronauts are blocked by large boulders. As a standalone science platform, the Gandalf Staff provides a power system for instruments and sensors to measure the lunar environment over long periods of time. The staff gathers data and sends the results to a lunar base station for analysis.

External Lighting↗

NASA NTRSDate not supplied

From Simulation to Reality With Random Noise

The challenging environment of autonomous vehicle (AV) navigation necessitates certain functions be performed by deep neural networks. Optimizing these models involves collecting vast quantities of domain-specific training data and ensuring that the dataset is representative of expected conditions. High-fidelity simulation plays a vital role in making this process feasible, allowing a wide range of scenarios to be explored at low cost. However, learning from simulation introduces subtle biases into models, which can degrade real-world performance in unpredictable ways. This effect can be mitigated with learning schemes specialized to bridge distributional shifts (transfer learning). Given the complex nature of these methods, the underlying models, and their environments, meaningfully evaluating performance is notstraight forward. Many unrelated factors can effect an improvement in generalization accuracy, but a full ablation analysis is often difficult. To tease out signal from noise, it is necessary to understand how transfer learning performance is affected by noise itself. The goals of this paper are (i) to establish a domain randomization baseline for a simple classification transfer learning task and (ii) to validate the RRAV testbed as a platform for further research in sim-to-real learning. We generate imagery from a simulation of NASA Ames Research Center and train a small convolutional neural network (ConvNet) to classify position relative to a centerline. Further models are trained with different types of noise progressively added to the data. The models are deployed aboard the on-site test vehicle to test real-world performance. In our experiments, we find that such naive domain randomization raises sim-to-real accuracy from 64% to 79%, while training directly on real data yields an 89% accuracy ceiling. These results suggest that the isolated mechanism of domain randomization can significantly improve generalization.

simulation↗