Combining Aiding Sensors with Multiple IMUs for Navigation Optimization (CAMINO)

The Navy is procuring small unmanned underwater vehicles (UUVs) to augment battlespace awareness in complex littoral environments, but these platforms require the development of navigation solutions that provide accurate localization at low size, weight, power, and cost (SWAP-C). To enable high-accuracy navigation at low SWAP-C, we developed a prototype system for Combining Aiding Sensors with Multiple IMUs for Navigation Optimization (CAMINO). CAMINO uses a novel hierarchical particle filter to combine noisy measurements from multiple inertial measurement units (IMUs) and direct observations of platform state from aiding sensors (e.g., depth sensors and magnetometers) into an accurate estimate of UUV geolocation and orientation. Our CAMINO system, using two IMUs, achieved more accurate localization results than dead reckoning and particle filtering based on measurements from a single IMU. CAMINO also outperformed nave methods of combining multiple IMUse.g., averaging two dead reckoned locationsdemonstrating the value of CAMINO"s hierarchical particle filter and local density reweighting. Critically, CAMINO achieves this localization accuracy at lower SWAP-C than currently available high-accuracy navigation systems, making CAMINO a promising option for small UUV navigation. In Phase II, we will build and demonstrate an in-water CAMINO prototype capable of processing live sensor feeds and localizing a platform in real-time.