基于改进神经网络的无人机起飞滑跑性能

文中通过对无人机的起飞滑跑过程,及起飞滑跑阶段的控制策略进行研究,对传统起飞滑跑性能理论计算方法的,局限性进行分析,提出一种基于改进神经网络算法的,无人机起飞滑跑性能计算方法。通过建立改进的神经网络模型,对各种环境条件下的发动机推力进行计算,依据飞行试验结果,利用单参数分析换算法,可以预测出不同环境条件下的无人机起飞滑跑性能。通过多架次飞行试验表明,基于改进神经网络算法的,无人机起飞滑跑性能计算方法精度较高,该方法与传统理论计算方法相比,更贴合工程实际应用,还可应用到无人机复杂任务环境,或新使用环境下的适应性分析中,达到降低飞行风险的目的。

Take-off Running Performance of UAV Based onImproved Neural Network

In this paper, the take-off nunning process and contrl strategy of UAV are studied, andthe limitations of traditional theoretical calculation methods of take-off nuinning perfomance are ana.lyzed. A new calculation method of UAV take-off running perfonmance based on improved neuralnetwork algorithm is proposed. By establishing an imprved neural network model and calculatingengine thrust under various environmental condiionals,the take-off running perfommance of UAV un.der different environmental conditions can be predicted by using the single-parameter analysis andconversion algorithm based on flight test results. The flight tests show that the method based on theimproved neural network algorithm is more accurate in calulating the take-off and rumning performance of UAV. Compared with the traditional theoretical calculation method . this method is more suitable for engineering application. This method can also be applied to the adaptability analysis of UAVin complex mission environment or new environment to reduce the flight risks.