The unmanned aerial vehicle (UAV) has significant advantages over the ground vehicle, where it can achieve a high degree of manoeuvrability, high speed response time and the ability of large area coverage. The problem considered here is that of an UAV localization and mapping using an extended Kalman filter (EKF), interval analysis (IA), covariance intersection (CI) and Hough transform (HT) for a partially known environment. The map is known partially in the sense that the obstacles and the land-marks are known to some extent. The vehicle is localized with respect to the known obstacles and to recognise the unknown obstacles to update the map of the environment. The focus is to develop an approach which can give a guaranteed performance of sensor-based localization and mapping that would increase the safety of the aerial vehicle and to produce a better performance in building the map of the environment. The guaranteed performance is quantified by explicit bounds of the position estimate of the vehicle. Generally, the UAVs carry the required sensors such as inertial sensors, accelerometers, and gyroscopes, to measure the acceleration and the angular rate, while the obstacle detection and the mapmaking is carried out with time of flight sensors such as ultrasonic or laser sensors. Most of these sensors give overlapping or complementary information, which offers scope for exploiting data fusion. This task of data fusion is accomplished by combining the measurements from different sensors that are obtained from two different methods namely EKF and IA, and by processing these measurements with a data fusion algorithm using the CI principle. This fused information is used with the measurements from the time of flight sensors such as laser sensor to update the map of the environment by applying the HT technique. The algorithms are complementary in the sense that they compensate for each other's limitations, so that the resulting performance of the sensor system is better than of its individual components, which in turn, improves the accuracy and richness in updating the map of the partially known environment. This proposed intelligent sensor system can provide a mathematically provable performance guarantees that are achievable in practice.

This paper addresses the robust estimation of a covariance matrix to express uncertainty when fusing information from multiple sensors. This is a problem of interest in multiple domains and applications, namely, in robotics. This paper discusses the use of estimators using explicit measurements from the sensors involved versus estimators using only covariance estimates from the sensor models and navigation systems. Covariance intersection and a class of orthogonal Gnanadesikan-Kettenring estimators are compared using the 2-norm of the estimates. A Monte Carlo simulation of a typical mapping experiment leads to conclude that covariance estimation systems with a hybrid of the two estimators may yield the best results.

No abstract

No abstract