Navigating by Means of a Position Potential
Benlin Xu, Petr Vaníček · NAVIGATION Journal of the Institute of Navigation · 2000
ABSTRACT: We have designed a new navigation algorithm that relies solely on measurements of vehicle position, velocity, and their error statistics. We consider the statistical confidence regions of position fixes as “sources” tending to “attract” the undetermined trajectory to pass through these regions. Based on these position fixes and their error statistics, a real-time potential field is constructed in which a mass particle with variable mass inertia is forced to move. To make the algorithm flexible to accommodate a changing navigation environment, we leave some parameters free and determine their values using a sequence of past observations and the least-squares criterion of discrepancies between position fixes and computed trajectory. A comparison with a Kalman filter navigation algorithm on real-life trajectories is also presented.