GPS-denied aerial flight is a popular research topic. The problem is challenging and requires knowledge of complex elements from many distinct disciplines. Additionally, aerial vehicles can present challenging constraints such as stringent payload limitations and fast vehicle dynamics. In this paper we propose a new architecture to simplify some of the challenges that constrain GPS-denied aerial flight. At the core, the approach combines visual graph-SLAM with a multiplicative extended Kalman filter. More importantly, for the front end we depart from the common practice of estimating global states and instead keep the position and yaw states of the MEKF relative to the current node in the map. This relative navigation approach provides simple application of sensor measurement updates, intuitive definition of map edges and covariances, and the flexibility of using a globally consistent map when desired. We verify the approach with hardware flight-test results.

In this article we detail the fundamentals of a new approach to GPS-denied navigation for aerial vehicles in confined indoor environments. We depart from the common practice of navigating within a globally referenced map, and instead keep the position and yaw states relative to the current node in the map. The approach combines elements of graph SLAM with a multiplicative extended Kalman filter (MEKF). The filter provides quality state estimates at a fast rate and a graph SLAM algorithm maintains a pose graph. We provide specific details for the relative MEKF. We verify the relative estimation approach with hardware flight test results accompanied by comparisons to motion capture truth. We also provide flight results with estimates in the control loop.

Multirotor helicopters are increasingly popular platforms in the robotics community. Making them fully autonomous requires accurate state estimation. We review an improved dynamic model for multirotor helicopters and analyze the observability properties of an estimator based on this model. The model allows better use of IMU data to facilitate accurate state estimates even when updates from a sensor measuring position become less frequent and less accurate. We demonstrate that the position update rate can be cut in half versus typical approaches while maintaining the same accuracy. We also

Accurate platform localization is an integral component of most robotic systems. As these robotic systems become more ubiquitous, it is necessary to develop robust state estimation algorithms that are able to withstand novel and non-cooperative environments. When dealing with novel and non-cooperative environments, little is known a priori about the measurement error uncertainty, thus, there is a requirement that the uncertainty models of the localization algorithm be adaptive. Within this paper, we propose the batch covariance estimation technique, which enables robust state estimation through the iterative adaptation of the measurement uncertainty model. The adaptation of the measurement uncertainty model is granted through non-parametric clustering of the residuals, which enables the characterization of the measurement uncertainty via a Gaussian mixture model. The provided Gaussian mixture model can be utilized within any non-linear least squares optimization algorithm by approximately characterizing each observation with the sufficient statistics of the assigned cluster (i.e., each observation's uncertainty model is updated based upon the assignment provided by the nonparametric clustering algorithm). The proposed algorithm is verified on several GNSS collected data sets, where it is shown that the proposed technique exhibits some advantages when compared to other robust estimation techniques when confronted with degraded data quality.