基于神经网络算法的多轴式机车运动轨迹优化  

Trajectory Optimization of Multi-spindle Locomotive Based on Neural Network Algorithm

作　　者：甘卉[1] 孙学进[1] 许崇明[1] 

机构地区：[1]青岛黄海学院,山东青岛266427

出　　处：《煤矿机械》2015年第7期275-278,共4页Coal Mine Machinery

摘　　要：为了提高井下机车的运行效率及稳定性,提出一种基于神经网络算法的运行轨迹优化方法。根据机车多轴控制特点,完成了控制系统硬件设计。通过空间轨迹状态的最优控制理论,建立了多目标动态评价函数,将机车在侧翻约束条件下的轨迹要求作为优化目标,与神经网络算法相结合,实现多目标优化。将优化算法应用于Matlab分析,对机车侧向速度、加速度以及横摆角速度进行数值模拟,结果表明,优化后的轨迹可缩短运行时间,并降低运行的波动性,提高控制精度。In order to improve the operation efficiency and stability of the underground locomotive, a trajectory optimization method based on neural network algorithm is put forward. According to characteristics of locomotive multi-axis control, the hardware design of control system is completed.Through state space curves of the optimal control theory, a multi-objective dynamic evaluation function is established. With the rollover constraints of locomotive and the combination of neural network algorithm,the trajectory is made as optimization goal, which can realize multi-objective optimization.Optimization algorithm was applied by Matlab for numerically simulation as lateral velocity, acceleration and yawing angular velocity,and the results show that the optimized trajectory can shorten the operation time, reduce the operation volatility, and improve the control precision.

关 键 词：神经网络 多目标优化 MATLAB 

分 类 号：TD525[矿业工程—矿山机电]