Plug and Play Navigation with Universal Interface for Aiding Sensors

We propose to develop a universal interface for flexible integration of a variety of aiding sensors with an inertial navigator toward plug-and-play navigation. In this multi-sensor fusion approach, a self-contained inertial navigator operates as a core sensor, which is assisted by secondary aiding sensors that are based on external radio signals such as GNSS and SOOP and external observations such as features extracted from laser scanners and video cameras. These aiding measurements may or may not be available depending on a specific navigation mission and the operational environment. The proposed approach exploits a synchronized timing of asynchronous measurements packer (STAMP) and a reconfigurable integration filtering engine (RIFE) in order to accommodate the asynchronous nature of aiding sensors' independent operations. A flexible modular structure with numerically stable implementation, which is adaptive to the initial conditions, the sensor error models, and the number, type, quality, and timing of measurements, will be developed using a decentralized (federated) Schmidt-Kalman filter to process complementary multi-rate and possibly out of sequence measurements. It therefore permits, on the run, physical/logical addition/subtraction of new aiding sensors, adjustment/weighting of sensors according to measurement qualities or due to a failure or a change in operational conditions and environments, and optimal selection/scheduling of aiding sensors. In Phase I, we will design the proposed universal interface in terms of its physical layer (the STAMP hardware) and its abstract layer (the RIFE software) and evaluate its performance via computer simulation. In Phase I Option, we will create preliminary specs for hardware and recommend the system configuration and the implementation approach for Phase II prototyping and demo.