A Master of Science thesis in Electrical Engineering by Bassel Mohamad Al Homssi entitled, "Portable Electromagnetic Surface Profiling System," submitted in June 2016. Thesis advisor is Dr. Nasser Qaddoumi and thesis co-advisor is Dr. Bassam Abu-Nabah. Soft and hard copy available.
Due to its potential accuracy and speed, the use of profiling and detection sensing systems has been gaining rapid popularity in the industry. Whether they are implemented in coordinate measurement machines or encoded in custom-designed measurement systems, real-time accuracy and precision in determining the sensor position and orientation are crucial elements to the performance of inspection sensor measurement techniques. Uncertainty in tracking these sensors within a measurement system can adversely affect the quality of 3-dimensional surface profiling techniques. Recent advancements in micro-electro-mechanical systems and their applications in multiple-axis inertia measurement units (IMUs) have been offering relatively high accuracy and precision in determining linear accelerations and angular velocities within the operating range of inspection measurement systems. This effort targets taking a step forward towards integrating IMUs to offer robust and portable inspection measurement systems. Utilizing the gyroscope as a feedback, different forward models will be analyzed to accurately extract gravitational acceleration from IMU's accelerometer measurements at random real-time orientations simulating realistic measurement environments. Accuracy in the forward model lends itself for real-time assessment of a sensor's position and orientation as part of the inverse model, which will be investigated with different physics-based data processing techniques. Forward and inverse models and their real-time transformational matrixes allow taking a point from the sensor coordinate system to the world coordinate system. The developed models are tested for dimensional accuracy against known input profiles to show the potential capabilities and limitations of the proposed effort. Using the proposed algorithm, the drift bias error was minimized drastically demonstrating potential in usage as an alternative to bulky profiling machines used conventionally.