Toward Bio-Inspired Autonomous Navigation: Long-Range Marine AUV Navigation

The challenge of autonomous marine navigation in GPS-denied environments has gained significant attention due to the growing demand for alternatives to satellite-based positioning. A promising solution is Earth-based localization, which leverages naturally occurring geophysical fields as reference sources, making it particularly suitable for long-range navigation. Among these, the Earth's geomagnetic field stands out as a reliable option due to its stability and global coverage, representing thus a viable alternative to GPS for autonomous missions. The interest in geomagnetic navigation has its origin in the observations of migratory animals, such as turtles and passerine birds, which appear to be capable of using geomagnetic cues to guide their long-distance journeys. Inspired by this natural strategy, we propose a novel, GPS-independent approach to long-range navigation that requires no prior geomagnetic map. The proposed algorithm is a guidance control law for real-time path planning and generates a piecewise trajectory that can be described as an 'Adaptive Biased Random Walk (ABRW).' At each step, the vehicle's heading for the next trajectory direction is randomly sampled from a multi-modal distribution formed by combining multiple Gaussian distributions. The mean and variance of each distribution is adjusted based on the magnetic field measurements collected by the vehicle during navigation. Initial simulation results demonstrate that the proposed algorithm enables successful vehicle movement using only geomagnetic information, without requiring a pre-existing map. This bio-inspired approach provides a robust framework for long-range autonomous marine vehicle navigation, offering valuable insights into geomagnetic-based guidance in featureless environments.