International Journal of Computer Applications (0975 – 8887) Volume 95– No. 9, June 2014 42 A Gravitational Black Hole Algorithm for Autonomous UCAV Mission Planning in 3D Realistic Environments A. A. Heidari GIS Division, Geomatics Eng. Dep., College of engineering, University of Tehran, Iran R. A. Abbaspour * GIS Division, Geomatics Eng. Dep., College of engineering, University of Tehran, Iran ABSTRACT This article addresses a novel approach to 3D mission planning of UCAVs in constrained environments. To solve this NP-hard problem, black hole algorithm (BH) is improved by considering stars gravities information. By modelling UCAV properties, aerospace constraints and DTM of environment, proposed mission planner based on black hole optimization algorithm is proposed. Also it provides a comparative study for efficiency evaluation of evolutionary 3D mission planners based on ACO, BA, DE, ES, GA, BH and PSO optimization algorithms. Then mission planning task of UCAV is performed. Simulations show the advantage of proposed gravitational BH mission planner. General Terms UCAV mission Planning, Artificial Intelligence, Autonomous Navigation, Mission Planning, Black Hole Keywords Unmanned combat aerial vehicle (UCAV), Flight simulation, 3D mission planning, Black hole optimization algorithm 1. INTRODUCTION UCAV is from the family of unmanned aircrafts developed for performing reconnaissance missions. Long-range drones have an autopilot system for following predesigned waypoints and continue motion based on planned mission, when they are out of the control of station's communication range. Operational UAVs need human control, but operator tasks are based on UCAV level of autonomy. However, developments of intelligent unmanned flight systems have become a growing trend in many research areas. Trajectory planning is a vital task in autonomous control processes of UCAV navigation, which is responsible for producing optimal trajectories from the launching location to the landing station considering some known constraints in environment. Many tasks should be applied to UCAV control systems for providing autonomous navigation. These steps maybe include scanning environment, DTM generation and mission planning. Mission planning is a complex requirement in the autonomous navigation. Its objective is to find an optimal flight path in proper time, to UAV be able to accomplish several mission tasks. Choosing efficient algorithms for solving mission planning problem is an influential task. Optimal mission planning relies on optimization technics, so it's usually solved offline. Use of UCAVs, which can fly autonomously in aerospace environments, is necessary in several innovative applications. Reliable safe navigation of UCAV in Complex missions has technical challenges and UCAV planning is an essential task. Aerospace applications of UCAVs require exact maneuvers and optimal decisions and robust mission planning algorithms. Complex space around UCAV flight trajectory makes the problem NP-hard. Based on pervious literatures, trajectory planning problem was turned into novel hybrid methodologies based on ICA [1], neural network [2], fuzzy logic [3], ACO [4], PSO [5,6], GA [7] and the artificial potential field [8]. When we have large mission ranges in UCAV flight, trajectory planning will be a large scale constrained optimization process. General methods on 3D trajectory planning could be used to solve this NP-hard problem including A* [9] and D* and rapidly exploring Random Trees (RRT) [10] and other is potential fields, evolutionary techniques include PSO, GA, ACO and multi- objective evolutionary algorithms [11,12]. Every method has its own robustness in certain aspects that is related to the problem complexity. The UCAV mission planning in realistic test fields is a well- known optimization problem, so many algorithms have been designed to solve this multi-constrained problem, such as differential evolution [13], biogeography-based optimization [14,15], genetic algorithm [16], ant colony algorithm [17] and its variant [18,19], cuckoo search [20,21], chaotic artificial bee colony [22], firefly algorithm [23,24], and intelligent water drops optimization [25], also algorithms such as immune GA (I-GA) [26], PSO [27], quantum-behaved PSO (Q-PSO) [28] and master-slave parallel vector-evaluated GA (MPV-GA) [29] have been applied. Black hole algorithm is a swarm intelligence approach inspired from the black hole phenomenon [30]. In this paper, performance of this algorithm will be improved by considering stars gravities information. For this aim, kind of gravitational force among stars is defined and the movement of stars toward the black hole is adjusted based on computed universal gravity during the searching solution space. Then this algorithm is applied to the UCAV mission planning scheme. The structure of the article is as follows. In Section 2 black hole Algorithm (BH) is introduced and then Improved BH is proposed. Section 3 defines the UCAV mission planning problem and section 4 reflects the main results of UCAV mission planning in 3D aerospace. Conclusion is at the last section. 2. GRAVITATIONAL BLACK HOLE ALGORITHM 2.1 Black Hole Phenomenon in Cosmology The theory of a black hole phenomenon is proposed based upon Einstein's general theory of relativity [31]. Black hole is expressed as existence of an infinite curvature in space-time [32]. Any close enough mass to the center of distortions, can't escape from gigantic gravitational field, including light [33]. Based on astrophysical explorations, there are many evidences that show supermassive stars with finished life cycle will be vanished in a form of black holes and make distortion in