Downloads & Free Reading Options - Results

A primary aspect involved in the navigation of a wide variety of vehicles, such as ground vehicles and aircraft, is the use of inertial measurement units. The purpose of the inertial measurement unit is to provide an internal measure of the linear acceleration and angular velocity of the vehicle in order to be used by a navigation system in the dead-reckoning of the vehicle s position and attitude. Due to the nature of sensors, the measurement of the linear acceleration and angular velocity are corrupted by several error sources, such as bias and noise, amongst others. In order to assess the coupled impact of these error sources on the navigation accuracy for an arbitrary vehicle, a simulation and analysis tool is developed that can model several inertial measurement unit mechanizations, simulate the error sources encountered in the acquisition of measurement data, emulate the navigation software, and perform a range of statistical analyses via Monte Carlo simulation. The developed tool is applied to a ground vehicle navigation scenario in order to assess the performance of different sensors and provide conclusions as to when each sensor should be used to achieve the best navigation performance.

The theory and assumption used in developing equations for the error analysis of a general class of inertial navigation systems are described. The computer program developed for their solution is described from a user's point of view. Its application includes the synthesis and/or analysis of inertial navigation systems used in ballistic missile or terrestrial space missions. The program is designed to allow studies of both pure inertial and aided inertial navigation systems, the latter being the process of updating navigation data via data from external sensors.

This report develops a method for analyzing the errors incurred in optimum estimation schemes. The optimum control law for modern control systems is usually a function of all the states of the system. In many practical cases it is necessary to extract information concerning the states from the measurement of the output. This extraction process is called estimation. Due to random measurement noise and the random noise components of the states of the system, the estimate of the states is not perfect. A set of computer programs has been developed that provides a way of evaluating optimum and sub-optimum Kalman estimation techniques. The random noise processes have been described in a statistical manner, in order to specify the random noise quantitatively as a function of time.