Private lunar missions are faced with the challenge of robust autonomous navigation while operating under stringent constraints on mass, power, and computational resources. This work proposes a motion-field inversion framework that uses optical flow and rangefinder-based depth estimation as a lightweight CPU-based solution for egomotion estimation during lunar descent. We extend classical optical flow formulations by integrating them with depth modeling strategies tailored to the geometry for lunar/planetary approach, descent, and landing--specifically, planar and spherical terrain approximations parameterized by a laser rangefinder. Motion field inversion is performed through a least-squares framework, using sparse optical flow features extracted via the pyramidal Lucas-Kanade algorithm. We verify our approach using synthetically generated lunar images over the challenging terrain of the lunar south pole, using CPU budgets compatible with small lunar landers. The results demonstrate accurate velocity estimation from approach to landing, with sub-10% error for complex terrain and on the order of 1% for more typical terrain, as well as performances suitable for real-time applications. This framework shows promise for enabling robust, lightweight on-board navigation for small lunar missions.

Moon exploration is undergoing a potentially transformative moment. Over the course of six weeks, three different lunar landers began a rocket-fueled space journey to learn more about Earth's nearest neighbor. All three landers are operated by private, and relatively newly-formed companies. That's a marked shift away from space exploration of the 20th century, which was dominated by state-backed, public institutions like NASA. If they complete their missions, these space upstarts could help pave the way for future planned human moon missions, and possibly, even a not-too distant lunar economy.

CAPE CANAVERAL, Fla. -- A Tokyo company aimed for the moon with its own private lander Sunday, blasting off atop a SpaceX rocket with the United Arab Emirates' first lunar rover and a toylike robot from Japan that's designed to roll around up there in the gray dust. It will take nearly five months for the lander and its experiments to reach the moon. The company ispace designed its craft to use minimal fuel to save money and leave more room for cargo. By contrast, NASA's Orion crew capsule with test dummies took five days to reach the moon last month. The lunar flyby mission ends Sunday with a Pacific splashdown.

A Japanese company is pushing ahead with plans to launch a private moon lander by the end of 2022, a year packed with other moonshot ambitions and rehearsals that could foretell how soon humans get back to the lunar surface. If the plans hold, the company, ispace, which is based in Tokyo, would accomplish the first intact landing by a Japanese spacecraft on the moon. And by the time it arrives, it may find other new visitors that already started exploring the moon's regolith this year from Russia and the United States. Other missions in 2022 plan to orbit the moon, particularly the NASA Artemis-1 mission, a crucial uncrewed test of the American hardware that is to carry astronauts back to the moon. South Korea could also launch its first lunar orbiter later this year.

Hardware and neural architecture co-search that automatically generates Artificial Intelligence (AI) solutions from a given dataset is promising to promote AI democratization; however, the amount of time that is required by current co-search frameworks is in the order of hundreds of GPU hours for one target hardware. This inhibits the use of such frameworks on commodity hardware. The root cause of the low efficiency in existing co-search frameworks is the fact that they start from a "cold" state (i.e., search from scratch). In this paper, we propose a novel framework, namely HotNAS, that starts from a "hot" state based on a set of existing pre-trained models (a.k.a. model zoo) to avoid lengthy training time. As such, the search time can be reduced from 200 GPU hours to less than 3 GPU hours. In HotNAS, in addition to hardware design space and neural architecture search space, we further integrate a compression space to conduct model compressing during the co-search, which creates new opportunities to reduce latency but also brings challenges. One of the key challenges is that all of the above search spaces are coupled with each other, e.g., compression may not work without hardware design support. To tackle this issue, HotNAS builds a chain of tools to design hardware to support compression, based on which a global optimizer is developed to automatically co-search all the involved search spaces. Experiments on ImageNet dataset and Xilinx FPGA show that, within the timing constraint of 5ms, neural architectures generated by HotNAS can achieve up to 5.79% Top-1 and 3.97% Top-5 accuracy gain, compared with the existing ones.

This time, though, they'll have one noble mission: to build shelter the first human colonizers will inhabit. A team of Japanese scientists is working to make this a reality. They started a company called ispace with the intention of launching a private space mission to the Moon. And they want to get started on it soon: the team is planning its first mission for late 2019, and a second in 2020. The Moon is smaller than our other planetary neighbors, but it could become a second home for an exponentially growing human population back on Earth.