基于DMOEA-APTC算法的无人机在线航迹规划

UAV Online Track Planning Based on DMOEA-APTC Algorithm

为了解决具有时间关联性质的动态多目标优化问题,以无人机在线航迹规划问题为基础,引入时间关联特征概念并建立无人机时间关联动态多目标优化问题模型,提出一种使用自适应预测响应机制和时间关联性优化机制的动态多目标双层优化算法。根据环境变化相关性判断环境变化强弱并启用不同响应机制,快速适应环境变化;优化过程中通过最小二乘法学习历史数据拟合航迹未来预测值,根据预测可靠性自适应选择“仅优化当前”或“同时优化当前与未来”优化模式;使用改进后的切比雪夫分解法对符合偏好的航迹进行决策。实验结果表明:所提算法在复杂飞行环境下降低飞行时长的同时具有更高生存概率,提高了无人机飞行稳定性,更合理有效地处理了在线航迹规划问题。

In order to solve the dynamic multi-objective optimization problem with time correlation,this paper introduces the concept of time correlation feature and establishes the model of UAV timecorrelation dynamic multi-objective optimization problem moedl on the basis of UAV online track planning problem,and proposes a dynamic multi-objective double-layer optimization algorithm using adaptive predictive response mechanism and time-correlation optimization mechanism(DMOEA-APTC).The intensity of environmental change was judged according to the correlation of environmental change and different response mechanisms were used to quickly adapt to environmental change.In the optimization process,the least square method was used to learn the historical data to fit the future predicted value of the flight path,and the optimization mode of"only optimize the current"or"optimize the current and future at the same time"was adaptively selected according to the reliability of the prediction.The improved Chebyshev decomposition method was used to make the decision of the preferred flight path.The experimental results show that the proposed algorithm can reduce the flight duration and have higher survival probability in complex flight environments,improve the flight stability of the UAV,and deal with the online flight path planning more reasonably and effectively.