The University of Texas at San Antonio Awards
- Project Name: Novel Technology for Detection and Prediction of Spreading of Air-Borne Chemical Agents
Awarding Agency: Deputy Assistant Secretary of Defense for Industrial Policy
Project Summary:
The research aims to: 1. to develop integrated modeling environment for dynamically tracking chemical plumes using cooperative Unmanned Aerial Vehicles (UAV’s) with onboard sensors and multi-scale chemical plume dynamic models and to conduct case students of possible chemical attack scenarios, 2. Data Assimilation – Synergy between model and observations through Validation and Calibration of the developed integrated modeling environment, and 3. Self-Guided UAV systems for testing the control laws and optimal model for chemical sensors.
Outcome:
The project developed an integrated modeling environment for dynamically tracking chemical plumes and analyzed the motion interplay between aerial vehicles and a chemical plume. It developed and designed cooperative vehicle guidance laws and estimation filtering techniques to detect the concentration levels of the plume and analyzed the resulting plume tracking performance the estimation filtering techniques using a simulator with realistic environmental and motion dynamic parameters, and coupled weather-forecasting-chemistry model with UAV dynamics and Chemical Sensor modules with two-way feedback. It modeled simulations based on geography and topography, the different release height of chemical sources, and calibrated/validated the simulator via sensors data, tested the controller and UAV dynamic model and efficacy of plume-chemistry model based on UAV data. - Project Name: Developing a Pattern-Based Measurement Model for Improving Software Reliability
Awarding Agency:US Army Materiel Systems Analysis Activity
Project Summary:
Various software systems have been employed by warfighters to acquire, process and communicate information on the battlefield. The reliability of these software systems is essential for warfighters to make correct decisions. It has been noticed that that focus on defect management (e.g., reduction of defects overtime) and have limitations on modeling defect severity and assessing defects with newly added software components. Therefore, PAM, a pattern-based approach which checks the software artifacts (e.g., class diagrams, code base) for violations of positive patterns and existence of anti-patterns.
Outcome:
Completed the development of the model to quantitatively estimate the reliability of a software project based on the detected positive-pattern violations and anti-pattern instances in software artifacts considering the importance and the activation probability of the patterns and document findings; 2. Developed the initial version of the model to demonstrate behavior and general principles; 3. Completed documenting the model behavior; and 4. Completed documenting the limitations for approach accuracy and success. - Project Name: Developing a Pattern-Based Measurement Model for Improving Software Reliability
Awarding Agency:US Army Materiel Systems Analysis Activity
Project Summary:
Early software-focused systems engineering and testing are critical for reducing and eliminating faults. Failure reduction significantly decreases costs for the Department of Defense and increases the ability of Soldiers to execute their missions. This Phase 2 effort will enhance and refine the Phase I findings and deliverables to include: 1. Learning Engine; 2. Pattern-Failure Correlation Estimation; 3. Code Pattern Detection; 4. Measurement Model (both parameter estimation and reliability estimation); 5. Test Coverage Correlation; and 6. Design Pattern Detection. Particular focus will be placed on characterizing the viability of applying the overall methodology to software code to provide an estimate of software reliability or, as a minimum, an assessment of software reliability risk.
Outcome:
The PAM project achieved the following major objectives: 1) Quantitatively estimated the reliability of a software project based on design and code patterns,2) Support to separate estimation of different aspects of software reliability, which enabled fine-grained prediction to the effect of software failures, 3) Confirmed the violation of positive patterns or existence of negative patterns using test-coverage-based approach, and 4) Evaluated the PAM model on real world software projects to conduct a feasibility study of developing a model to perform quantitative estimation of software reliability. To evaluate whether the PAM model is effective, the project implemented PAM for a certain software platform (e.g., Android system), and tested it on a large number of real-world software projects developed for the platform. - Project Name: Novel Technology for Detection and Prediction of Spreading of Air-Borne Chemical Agents
Awarding Agency: Deputy Assistant Secretary of Defense for Industrial Policy
Project Summary:
In the event of release of a toxic chemical agent in the atmosphere, toxic clouds or plumes form and spread laterally due to turbulence in the atmosphere. Predicting the trajectory of the chemical plume is very challenging. An interdisciplinary approach is being used to overcome the challenges in predicting the source of the attack, and the trajectory or the path of the plume. For this purpose, we are developing a mobile cooperative sensor network consisting of self-guided cooperative UAV’s with on-board meteorological and chemical sensors that can predict (a) source of the plume, (b) concentration-averaged centroid of the plume to identify the trajectory of the plume.
Outcome:
The research outputs are completions of the State of the Art Laboratory, Enhanced Drone Platform, Control Systems, Software Algorithms, Sensors, UGVs, Chemical Sensor Chamber, and Communication Network. In addition, the project produced one (1) PhD dissertation, one (1) Master's thesis, six (6) undergraduate trainnig certificates, five (5) publications, six (6) conference presentations, and two (2) patent applications. - Project Name: Improving Leeway Drift Data for the USCG Search and Rescue Optimal System
Awarding Agency:United States Coast Guard
Project Summary:
The fundamental challenge of estimating and forecasting search areas in the presence of large uncertainties include the following – errors that arise from the current fields, the wind fields, turbulence mixing, wave-effect, poor estimates of the real drift properties of the object, and drag forces acting on the object. The unsteady forcing and variation of drag forces complicates the problem. Without a proper estimate of the basic drift properties and their associated uncertainties, forecasting the drift and expansion of a search area remains difficult. The integration of scientific principles of modeling of marine hydrodynamics, machine learning and nonlinear control theory – for the first time – to develop a novel approach for tracking trajectory of leeway subjected to unsteady forcings by developing an integrated simulating environment.
Outcome:
The proposed research accomplished the following goals: 1. Collection of lee way data by tracking, observing, and correlating the effects of winds and waves on an object's motion, 2. Analyzed lee way data for downwind and crosswind drift components and developed a combined dynamic model that relates the object motion to the environmental conditions, 3. Validated the model by using existing environmental conditions to "forecast" leeway object drift from deployment to retrieval, and 4. Refined the model based on the new results, determined ranges of uncertainties and weight of significant environmental variables in correlation functions.
