This paper presents results demonstrating real-time six degree of freedom localization, mapping, navigation and obstacle avoidance in an outdoor environment using only a lowcost off-the-shelf inertial measurement unit and a monocular camera. This navigation system is intended for operation of small unmanned aerial vehicles when GPS signals are unavailable due to obstructions or jamming and when operating in cluttered environments such as urban canyons or forests. A small radio-controlled car is used as a test bed.A bearings-only Simultaneous Localization and Mapping algorithm was implemented to localize both the vehicle and obstacles. This paper describes: (a) the hardware used; (b) a two-step approach for data association (image frame to image frame followed by image frame to map); (c) a technique for landmark initialization for the case where landmarks are located on the ground. Hardware test results demonstrating navigation to a goal in an obstacle strewn environment are presented. The effect of unmodelled sensor biases is examined in simulation.
A method for navigation of a small UAV through an unsurveyed environment (e.g. a forest) is presented. In particular the problem of estimating the state of the aircraft and of obstacles in the environment given limited sensors (an inertial measurement unit and a monocular camera) is addressed. The combination of limited observability and the close proximity of the vehicle to the obstacles lead to significant nonlinearities which cause Extended Kalman Filter (EKF) based approaches to fail. This paper presents an implementation of an Unscented Kalman Filter (UKF) to estimate the locations of obstacles and the state of the UAV based on measurements from the IMU and camera. This solution is applicable to navigation of a 6DOF vehicle in a three dimensional environment and allows generic obstacle avoidance routines to be used. Simulation results are presented for a two-dimensional environment which demonstrate: (a) stability and consistency of the UKF implementation and a comparison with an EKF implementation; (b) fusion with a potential field obstacle avoidance algorithm to enable navigation in an unknown twodimensional environment.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.