Covariance bounds for augmented state Kalman filter application

Novel insights into the covariance bounds of an augmented state Kalman filtering (ASKF) application are provided. These are obtained through empirical investigations based on a scenario where a dynamic vehicle senses a static feature for the purpose of mapping that feature and simultaneously localising the vehicle. Numerical results indicate a relationship between the Riccati matrices of the vehicle and feature. Generalisations to multiple features, multiple vehicles and decentralised networks are considered. The relationships derived are applied to a simple system design example