deletions | additions       diff --git a/textbf_A_Testing_Framework_for__.tex b/textbf_A_Testing_Framework_for__.tex  new file mode 100644  index 0000000..869b7aa  --- /dev/null  +++ b/textbf_A_Testing_Framework_for__.tex 
...
\textbf{A Testing Framework for Multi-UAV Conflict Resolution using Simulation and Multi-objective Search}  Xueyi Zou1 Rob Alexander2 John McDermid3  University of York, York, England, YO10 5GH, UK  We tackle the problem of testing the capability of multi-UAV conflict resolution algorithms. The problem was formalized as a multi-objective optimization problem, where there are two objectives: finding air traffic encounters (1) with the minimum number of UAVs, (2) able to falsify the conflict resolution algorithm. A framework was developed to automatically find the encounters satisfying there criteria using simulation and multi-objective search. A parameterized geometry representation was designed to specify two-UAV encounters in 3D space. Encounters involving multiple UAVs were generated and simulated in a fast-time agent-based simulator by combining different geometry representations. A multi-objective search method based on Genetic Algorithm was designed to find the required encounters. The proposed approach and a random search based approach were used to test a widely-cited open source multi-UAV conflict resolution algorithm, specifically, ORCA-3D. Comparison shows that the proposed approach can find the encounters meeting these criteria more efficiently. The resultant encounters can help to find limitations of the conflict resolution algorithms.  \section{I. Introduction}  The Federation Aviation Administration Wednesday announced its new Pathfinder Program, which will work with two companies on extended line-of-sight and beyond line-of-sight drone flight.  Civilian use of UAVs → Multi-UAV conflict resolution → how to evaluate multi-UAV conflict resolution algorithm → we tackle the problem of testing the capacity of the algorithms.  \section{II. Problem Formalization and the Testing Framework}  two conflict objectives:   1, finding a encounter involving the minimum number of UAVs → meaning this encounter is more likely to happen.  2, finding a encounter that will falsify the conflict resolution algorithm, i.e. lead to a collision.  The less UAVs involved, the more difficult to cause a collision.  The framework: search guided simulation  simulation: agent-based simulation(1), each UAV agent contains a conflict resolution agentsearch: multi-objective search, searching a load of encounters to find ones that fulfill the two objectives. A simulation run is used to judge whether an encounter fulfill the second objective.  \section{III. Encounters Generation}  A. Parameterized Two-UAV Encounter  First fix the position and velocity of ownship. 7 parameters are used to specify the intruder:  B. Multi-UAV Encounter  Fix the ownship, combining the two-uav encounters to generate intruders.  IV. Agent-based Simulation  A. Agents  Different kind of agents: conflict resolution agents, UAVs, global observers (proximity measurer, accident detector)  B. Environment  The size of the environment. What are there in the environment:  waypoints, targets  V. Genetic Algorithm Based Multi-objective Optimization  A. Genetic Algorithm   B. Genome Representation for Encounters  C. Genetic Operators  Gene mutator, gene remover.  D. Fitness Evaluation    E. Pareto Non-dominated Front  VI. Case Study: ORCA-3D  A. ORCA-3D  B. Guided Search  C. Random Search  D. Result and Discussion  VII. Conclusion  A conclusion section is not required, though it is preferred.   Acknowledgments  Xueyi Zou would like to thank the China Scholarship Council (CSC) for its partially financial support for his PhD study.  References  1Vatistas, G. H., Lin, S., and Kwok, C. K., “Reverse Flow Radius in Vortex Chambers,” AIAA Journal, Vol. 24, No. 11, 1986, pp. 1872, 1873.  2Dornheim, M. A., “Planetary Flight Surge Faces Budget Realities,” Aviation Week and Space Technology, Vol. 145, No. 24, 9 Dec. 1996, pp. 44-46.  3Terster, W., “NASA Considers Switch to Delta 2,” Space News, Vol. 8, No. 2, 13-19 Jan. 1997, pp., 1, 18.