基于神经网络的机器人运行自动化控制方法研究  

Research on robot operation automation control method based on neural network

作　　者：上官霞 SHANGGUAN Xia(State Grid Fujian Electric Power Co.,Ltd.,Fujian Fuzhou 350003,China)

机构地区：[1]国网福建省电力有限公司,福建福州350003

出　　处：《工业仪表与自动化装置》2022年第5期71-75,115,共6页Industrial Instrumentation & Automation

摘　　要：为提高智能语音机器人的语音识别精度、优化自动化控制效果、合理选择运行路径,提出基于神经网络的机器人运行自动化控制方法研究。分析当运行轨迹不同时智能语音机器人左、右轮速的函数关系,保证运行过程中效率最大化;凭借短时能量、短时平均过零率将语音命令从背景中分离,明确命令帧长度,通过不断修改BP神经网络算法的阈值和权值,尽可能地缩减控制误差,提高网络模型的收敛速度;机器人在接收到语音指令后,及时调整左、右轮速,自主化完成作业任务。将所提方法与其他方法展开对比实验测试,结果表明,所提方法可有效规避障碍物,并且在起始点与目标点之间选择一条最短路径,同时保证控制时间和控制误差为最低。Aiming at the problems of low automatic control efficiency and unsatisfactory path selection of intelligent voice robot,a robot operation automatic control method based on neural network is proposed.The functional relationship between the left and right wheel speeds of the intelligent voice robot is analyzed when the running tracks are different,so as to maximize the efficiency in the running process.By constantly modifying the threshold and weight of BP neural network algorithm,the control error is reduced as much as possible and the convergence speed of network model is improved.After receiving the voice command,the robot adjusts the left and right wheel speeds in time to complete the operation task autonomously.The experimental results show that the proposed method can effectively avoid obstacles,select the shortest path between the starting point and the target point,and ensure the minimum control time and control error.

关 键 词：智能语音机器人 BP神经网络 运行自动化控制 LM算法 收敛速度 

分 类 号：TP185[自动化与计算机技术—控制理论与控制工程]