Open Access. Powered by Scholars. Published by Universities.®
Navigation, Guidance, Control and Dynamics Commons™
Open Access. Powered by Scholars. Published by Universities.®
- Discipline
Articles 1 - 9 of 9
Full-Text Articles in Navigation, Guidance, Control and Dynamics
Improved Ground-Based Monocular Visual Odometry Estimation Using Inertially-Aided Convolutional Neural Networks, Josiah D. Watson
Improved Ground-Based Monocular Visual Odometry Estimation Using Inertially-Aided Convolutional Neural Networks, Josiah D. Watson
Theses and Dissertations
While Convolutional Neural Networks (CNNs) can estimate frame-to-frame (F2F) motion even with monocular images, additional inputs can improve Visual Odometry (VO) predictions. In this thesis, a FlowNetS-based [1] CNN architecture estimates VO using sequential images from the KITTI Odometry dataset [2]. For each of three output types (full six degrees of freedom (6-DoF), Cartesian translation, and transitional scale), a baseline network with only image pair input is compared with a nearly identical architecture that is also given an additional rotation estimate such as from an Inertial Navigation System (INS). The inertially-aided networks show an order of magnitude improvement over the …
Magnetic Field Aided Indoor Navigation, William F. Storms
Magnetic Field Aided Indoor Navigation, William F. Storms
Theses and Dissertations
This research effort examines inertial navigation system aiding using magnetic field intensity data and a Kalman filter in an indoor environment. Many current aiding methods do not work well in an indoor environment, like aiding using the Global Positioning System. The method presented in this research uses magnetic field intensity data from a three-axis magnetometer in order to estimate position using a maximum – likelihood approach. The position measurements are then combined with a motion model using a Kalman filter. The magnetic field navigation algorithm is tested using a combination of simulated and real measurements. These tests are conducted using …
Navigation Using Vector And Tensor Measurements Of The Earth's Magnetic Anomaly Field, Lauren A. Mount
Navigation Using Vector And Tensor Measurements Of The Earth's Magnetic Anomaly Field, Lauren A. Mount
Theses and Dissertations
This research explores the viability of using a navigation system that relies on measurements of the magnetic anomaly field as an alternative to GPS navigation. Previous research has been conducted on developing a navigation system using the intensity of the Earth's magnetic anomaly field as an alternative signal. This research focuses on using vector and tensor measurements, as opposed to scalar measurements of the anomaly field, as a means of obtaining accurate position and orientation solutions. This paper presents two navigation systems. The first uses an Extended Kalman Filter (EKF) with vector measurements of the magnetic anomaly field to aid …
Visual-Ins Using A Human Operator And Converted Measurements, Turner J. Montgomery
Visual-Ins Using A Human Operator And Converted Measurements, Turner J. Montgomery
Theses and Dissertations
A method human operated INS aiding is explored in which the pilot identifies and tracks a ground feature of unknown position over a short measurement epoch using an E/O sensor. One then refers to Visual-INS. In contrast to current research trends, a human operator is entrusted with visually tracking the ground feature. In addition, a less conventional measurement linearization technique is applied to generate “converted” measurements. A linear regression algorithm is then applied to the converted measurements providing an estimate of the INS horizontal velocity error and accelerometer biases. At the completion of the measurement epoch, the INS is corrected …
Navigation Constellation Design Using A Multi-Objective Genetic Algorithm, Heather C. Diniz
Navigation Constellation Design Using A Multi-Objective Genetic Algorithm, Heather C. Diniz
Theses and Dissertations
In satellite constellation design, performance and cost of the system drive the design process. The Global Positioning System (GPS) constellation is currently used to provide positioning and timing worldwide. As satellite technology has improved over the years, the cost to develop and maintain the satellites has increased. Using a constellation design tool, it is possible to analyze the tradeoffs of new navigation constellation designs (Pareto fronts) that illustrate the tradeoffs between position dilution of precision (PDOP) and system cost. This thesis utilized Satellite Tool Kit (STK) to calculate PDOP values of navigation constellations, and the Unmanned Spacecraft Cost Model (USCM) …
Error Characterization Of Flight Trajectories Reconstructed Using Structure From Motion, Daniel C. Alix
Error Characterization Of Flight Trajectories Reconstructed Using Structure From Motion, Daniel C. Alix
Theses and Dissertations
This research effort assessed the accuracy of Structure from Motion (SFM) algorithms in replicating aircraft fight trajectories. Structure from Motion techniques can be used to estimate aircraft trajectory by determining the position and pose of an aircraft mounted camera from a sequential series of images taken during flight. An algorithm is proposed and implemented that successfully reconstructed aircraft trajectory using only a known starting position and a sequential series of images. The error in and reliability of the algorithm was found to be a function of image resolution as well as the amount of overlap and angular separation between sequential …
An Integrity Framework For Image-Based Navigation Systems, Craig D. Larson
An Integrity Framework For Image-Based Navigation Systems, Craig D. Larson
Theses and Dissertations
This work first examines fundamental differences between measurement models established for GPS and those of proposed image-based navigation systems. In contrast to single value per satellite GPS pseudorange measurements, image measurements are inherently angle-based and represent pixel coordinate pairs for each mapped target. Thus, in the image-based case, special consideration must be given to the units of the transformations between the states and measurements, and also to the fact that multiple rows of the observation matrix relate to particular error states. An algorithm is developed to instantiate a framework for image-based integrity analogous to that of GPS RAIM. The algorithm …
Improved Mathematical Modeling For Gps Based Navigation, Salvatore Nardi
Improved Mathematical Modeling For Gps Based Navigation, Salvatore Nardi
Theses and Dissertations
This thesis is concerned with the development of new closed form GPS position determination algorithms that work in the presence of pseudorange measurement noise. The mathematical derivation of two closed form algorithms, based on stochastic modeling and estimation techniques, is presented. The algorithms provide an estimate of the GPS solution parameters (viz., the user position and the user clock bias) as well as the estimation error covariance. The experimental results are analyzed by comparison to the baseline results from the conventional Iterative Least Squares (ILS) algorithm. In typical GPS scenarios, the closed form algorithms are extremely sensitive to noise, making …
Pseudorandom Code Generation For Communication And Navigation System Applications, John F. Brendle Jr.
Pseudorandom Code Generation For Communication And Navigation System Applications, John F. Brendle Jr.
Theses and Dissertations
This research project investigated the design, construction and evaluation of a pseudorandom code generator for communication and navigation system applications. These types of codes include spreading codes, Gold codes, Jet Propulsion Laboratory (JPL) ranging codes, syncopated codes, and nonlinear codes. Such waveforms are typically used in communication and navigation system applications. The code generator uses the Stanford Telecom STEL-1032 Pseudorandom Number (PRN) coder. A coder interface was designed and constructed for manual data entry to the registers of the PRN coder. The code generator is capable of independently clocking and generating all possible codes with lengths up to 4,294,967,295 bits. …