Intelligence-Based Multi-Resolution 3D Visual Modeling, Registration And Obstacle Avoidance Capabilities For Unmanned Vehicles

As one of NASA's key motivation, the use of truly autonomous unmanned vehicles (UV) has been hampered by lack of sophisticated and resource efficient obstacle avoidance systems. Solving the obstacle avoidance problem to permit truly autonomous operations in a cluttered and occluded environment has been the subject of considerable research. In Phase I, UtopiaCompression (UC) has developed a layered, intelligent and adaptive system concept that will facilitate UV operations by solving the collision avoidance and path planning problems using inexpensive imaging sensors and modest computational resources. In Phase II, UC shall fully implement this concept. UC's focus is to push the limits of automation in the 3D arena by providing a real-time, end-to-end solution that includes: (i) deriving 3D structures (computing depth) from overlapping (stereo pair) images or video sequence acquired from sensors aboard autonomous agents; (ii) using the derived 3D structure to compute and model a spatio-temporal vector field; and (iii) based on the distance to the moving/stationary objects, plan a navigation strategy that commands the agent to cruise along a safe path from a source to a destination, avoiding the stationary/moving obstacles in the frontal hemisphere. UC's proposed technology provides immediate benefits to NASA and Non-NASA commercial applications.