[author_cn_name].[cn_title][J].空军工程大学学报:自然科学版,[year_id],[volume]([issue]):[start_page]-[end_page] 基于模糊免疫神经网络PID算法的全向底盘控制方法-Control Method for Omni Directional Chassis Based on Fuzzy Immune Neural Network PID Algorithm
文章摘要
孙浩水,王小平,王晓光,林秦颖.基于模糊免疫神经网络PID算法的全向底盘控制方法[J].空军工程大学学报:自然科学版,2018,19(4):59-65
基于模糊免疫神经网络PID算法的全向底盘控制方法
Control Method for Omni Directional Chassis Based on Fuzzy Immune Neural Network PID Algorithm
  
DOI:10.3969/j.issn.1009-3516.2018.04.011
中文关键词: 全向底盘  模糊神经网络  PID算法  免疫算法
英文关键词: omni directional chassis  fuzzy neural network  PID algorithm  immune algorithm
基金项目:航空科学基金(20145196023)
作者单位
孙浩水,王小平,王晓光,林秦颖 空军工程大学航空工程学院,西安,710038 
摘要点击次数: 84
全文下载次数: 121
中文摘要:
      针对全向底盘控制的实际需求,提出了基于模糊免疫神经网络PID算法的智能控制方法。首先根据神经网络算法和模糊算法的结构特点建立了模糊神经网络模型,并使用误差反向传播的方法对模型进行训练;然后使用免疫算法确定学习率,实现了对PID参数的动态整定,并对底盘路径跟踪控制器参数进行整定,以实现对底盘的精确运动控制;最后建立了底盘的运动学模型,基于Matlab平台进行了相关算法的仿真,并基于Linux Ubuntu系统下Tensorflow框架搭建并训练了神经网络模型,进而实现了整体算法。轨迹跟踪试验表明:当底盘沿不同方向以5 m/s的速度进行轨迹跟踪时,最大误差为4.88 cm,平均误差为0.25 cm,该算法能够有效地对底盘进行控制,满足全向底盘控制的要求。
英文摘要:
      Aiming at the actual demand of omni directional chassis control, an intelligent control method based on fuzzy immune neural network PID algorithm is proposed to set parameters of the path tracking controller of the chassis to realize the precise motion control of the chassis. According to the structural characteristics of the neural network algorithm and the fuzzy algorithm, the fuzzy neural network model is established and the model is trained by the method of error back propagation. The immune algorithm is used to determine the learning rate and then, the dynamic tuning of the PID parameters is realized. After that, the kinematic model of the chassis is established. Then, based on Matlab platform, the relevant algorithm is simulated. Finally, the neural network model is established and trained via the Tensorflow structure under the Linux Ubuntu system and the algorithm is then implemented. When the chassis carries out trajectory tracking in different directions at a 5 m/s speed, the maximum error of the process is 4.88 cm, the average error is 0.25 cm. Simulations and experiments both show that the algorithm can effectively control the chassis to meet requirements of thecontrol of the omni directional chassis.
查看全文   查看/发表评论  下载PDF阅读器
关闭