Research On UAV Mission Planning Methods Based On Intelligent Optimization And RRT Algorithm

The tendency of UAV system is aim to be intelligent and autonomous. Mission planningoccupies an important position in designing of the future UAV system, which is also one of the keytechnologies in the realization of the UAV’s autonomous control. UAV route planning and taskallocation are two major problems studied in this paper.For the problem of UAV route planning under three-dimensional and static threat circumstance, akind of multiple heuristic ant colony algorithm is proposed. This algorithm analyzes the condition ofthe distance and the threat distribution among the current position, candidate position and targetposition of UAV. And these above conditions are utilized to be the multiple heuristic information inthe state transition of ants, which can help determine ants foraging behaviors. Moreover, theconvergence of this multiple heuristic ant colony algorithm is proved. In the meantime, an artificialpotential ant colony optimization algorithm (APACO) is proposed, which combines ant colonyoptimization algorithm with artificial potential field. According to the potential force distribution, thestate transition rules comprise deterministic selection and probability selection. The experimentalresults show that the proposed algorithm for the route planning is better than both of simplex artificialpotential field algorithm and ant colony optimization algorithm. Double improved algorithms caneffectively shorten the route planning time, improve planning accuracy, and ultimately achieve anoptimal path.An improved hybrid particle swarm optimization (IHPSO) algorithm is proposed, whichintegrates the Boids model and the Powell algorithm. Collision avoidance mechanism is integratedinto PSO to get rid of the shackles of local minima, and then Powell algorithm is for further localsearch. The convergence of IHPSO algorithm is proved. The test results show that, compared with thestandard particle swarm algorithm, IHPSO has a better accuracy and success rate. In light of themultiple UAVs cooperative route planning problem, a threat heuristic particle swarm algorithm isbrought out. Threat information is added as part of the particle velocity updating formula, which cankeep particles far away from threat area. The two stages of collaborative planning framework areadopted. Firstly, using k-means clustering method to classify particle swarm and get multiplecandidate paths for each UAV. Then, through the coordination variable and coordination function,cooperative constraints can be tackled. According to results of the simulation, the validity of theproposed algorithm is proved.The relation between parameters setting and the performance of RRT algorithm is studied.Meanwhile, the guiding principle for the selection of parameters in RRT algorithm is put forward. On this basis, an improved RRT algorithm is proposed. This algorithm uses chaotic sequence to generaterandom node, and uses the fuzzy inference system to adjust algorithm parameters dynamically. Animproved dual RRT algorithm for route replanning in popup threat environment is proposed. Thisalgorithm can make full use of the existing off-line route information, delete invalid nodes, andpreserve residual random tree uninfluenced by popup threats, then according to the UAV’s currentlocation, execute improved dual RRT growth. A path pruning method for the removal of redundantwaypoints is designed.The Bessel curve is used for improving path smoothness. For unknownenvironment route planning problem, a rolling RRT algorithm is proposed. The execution process is akind of alternated rolling route planning of the UAV flight curve. A random node selection method isproposed, which can guide the tree faster approximate the rolling window boundary. The experimentresults show that the proposed method effectively meets the requirements of the unknownenvironment online route planning, and achieves the satisfied results with the rolling planning.A multiple group ant colony algorithm (MGACO) is proposed for the condition of multipleUAVs attacking multiple ground targets. Design of comprehensive optimum index includes UAVattack benefits, survival probability, distance cost. According to the comprehensive remaining ability,choosing the ant to perform state transition, and then deciding which task arranged to this ant. Using2-opt algorithm for further local optimization. The experiment results show that the task allocationscheme generated by MGACO algorithm can satisfy various constraints and need less iteration. Thetask load of different UAVs is relatively balanced.A route planning system is designed for a small UAV, including data entry, route generation,flight simulation and verification, and route loading and checking etc. This system sets taskspossession by software interfaces, and can realize the route planning, verification and upload. At last,an actual example is given to verify the route planning system.