Summer at SAIC’s Maritime System Solutions Division
By Zachary J. Harris ‘11
This summer, I worked for SAIC’s Maritime System Solutions Division in Lynnwood, WA. This was my third internship with the office, and each internship proved to be more exciting than the last.
SAIC recently purchased a new Autonomous Underwater Vehicle (AUV), and I was asked to analyze the Factory Acceptance Test (FAT) data. The FAT had just been conducted the last week of June, and I was the first to take a detailed look at the data. This was terribly exciting, because the vehicle design had never been tested at sea before. The vendor had engineering data and had run vehicle simulation, but nobody really knew how she would behave in the field; because an underwater vehicle is out of sight for most of the test, it is difficult to evaluate her performance from the deployment craft. I was given the opportunity to understand and characterize the vehicle’s behavior (path, heading, pitch, roll, etc) before anybody else knew what had happened, including the people present at the test.
During the FAT, the data were recorded by the vehicle’s navigation and sensor suite and, upon surfacing, transported via data link to the laptop running the mission control software. The data were then exported to an Excel format for analysis. From Excel, all 12000 rows of the data were imported into Matlab as an array. The purpose of my analysis was to characterize and quantify the vehicle’s performance, particularly regarding the navigation and control system. Matlab’s graphing functions and data processing tools for massive amounts of data are far superior to those of Excel, and easily allow for multiple graphs to be overlaid. I was given the leeway to decide what graphs and information were necessary to characterize the vehicle’s performance, and, with the help of senior engineers, to determine what those graphs meant in terms of the control system design. The ultimate goal is to use the analysis to improve the vehicle’s control system and update a navigation simulation to reflect the vehicle’s physical behavior.
The three deliverables for the summer were the Matlab code written for the analysis, a written report describing my analysis method and results, and a final presentation to the engineering team. The code was to be as universal as possible, allowing for analysis of the many future tests that will be conducted on the vehicle. For me, the greatest aspect of this internship was the perfect blend of autonomy to perform the task as I saw fit, and guidance from experienced engineers when I needed help. Frankly, I never thought data analysis could be so exciting.