Representation for Navigation

ABSTRACT: Matrix Research, Inc. proposes to design, implement, and test novel feature representations in Multi-Agent Simultaneous Localization and Mapping (MA-SLAM) systems for enhanced navigation accuracy/reliability. Specifically, we will develop generic, yet compact, multi-modal feature representations, along with a methodology for their extraction in significantly variable conditions, which may include densely and/or sparsely populated dynamic scenes. The design of these local and global feature representations will be tailored for navigation applications in which low-bandwidth, low-rate communications between agents constrains data throughput. Our primary focus is the development of a novel, compact feature representation that enables efficient communication and exploitation by agents within the system. The feature representations, along with the associated mapping algorithm, will be general enough to support multiple sensing modalities for utilization in a multi-source navigation system. We will consider the question of how many agents are needed with respect to the available feature representations, bandwidth constraints, and mapping accuracy requirements. Our analysis will include algorithm simulations for proof-of-concept, as well as a system prototype and test plan for future development. BENEFIT: The primary benefit of successful completion of this effort is a revolutionary new capability for extracting features for mapping systems. This capability has numerous commercial applications in various business sectors such as defense, search and rescue, mapping, mining, and robotics.