Get Started
Home
Topics
Search
Library
Research questionHow can lunar robots maintain accurate, bounded-latency instance segmentation under low light, tight compute, and radiation faults?Lunar rovers must interpret scenes in dim conditions while onboard processors impose strict latency and power limits. Radiation can also corrupt inference without obvious failures, so perception must account for both resource constraints and hardware reliability.
AI
Computer Vision
Inference Optimization
Research Paper
Robotics
Small / On-device Models
Technology
Latest papersRecent research connected to this question, newest first.Hardware-Accelerated Instance Segmentation for Resource-Constrained Space Robotics with Criticality AnalysisThe evidence concerns a YOLO-based instance-segmentation model deployed on a Deep Learning Processor Unit for a lunar micro-rover. It includes label-free quantization calibration, statically compiled execution, and software-level criticality analysis; reported results include 309 ms latency, 5.7 W power consumption, recovery of 69.8% of quantization-related accuracy loss, and a 31.7% reduction in global criticality.research paper · Sep 2, 2026
Related questions
How can automated lunar-crater detection remain reliable across crater sizes, illumination, and rugged terrain?How can robots maintain reliable underground localization when smoke, dust, and self-similar corridors undermine vision and LiDAR?How can monocular models estimate relative lunar terrain height reliably for landing hazard localization?How can legged robots maintain accurate proprioceptive body-state estimates when cameras and LiDAR become unreliable?