基于多核学习的潜射武器环境因子安全性预测

Safety Prediction with Environment Factors of Submarine Launched Weapon Based on Localized Multiple Kernel Learning

摘要: 环境因素是影响潜射武器安全性的重要因素。针对安全性预测对泛化能力要求高的问题，采用局部多核学习方法对环境安全性进行预测。局部多核学习方法的输入为环境因子(温度及其变化率、湿度及其变化率)，输出是安全性预测结果。仿真结果证实，局部多核学习方法可以有效对武器系统环境安全性进行预测。

Abstract: Environment is an important factor to safety of submarine launched weapon. To improve safety prediction correct rate, the paper use Localized Multiple Kernel Learning (LMKL) to predict safety of underwater weapon. The input of LMKL includes: temperature, temperature change rate, humidity and humidity change rate, and the output of LMKL is safety prediction result. The simulation demonstrates that LMKL can accurately predict safety of submarine launched weapon with environment factors.