论文部分内容阅读
围绕巡线机器人电磁导航,分析了电磁传感器阵列信息与机器人位姿之间强非线性映射的特点,提出了采用反向传播(BP)人工神经网络方法对机器人上的电磁传感器阵列信息进行数据融合的方法。该方法利用神经网络方法的非线性拟合逼近特点,通过离线学习训练,建立电磁传感器阵列信息与机器人机械臂相对导线的空间位姿的非线性映射关系模型,实现巡线机器人自主导航。该方法可有效减小单个电磁传感器信息检测误差对机器人导航控制的影响,提高机器人抓线控制的准确性。最后,通过实验验证了该方法的可行性和有效性。
The characteristic of strong nonlinear mapping between electromagnetic sensor array information and robot pose is analyzed around the electromagnetic navigation of patrol robot. The BP artificial neural network method is proposed to fuse the electromagnetic sensor array information on the robot Methods. The method uses the nonlinear fitting approximation feature of neural network method. Through offline learning training, a nonlinear mapping relationship model between the electromagnetic sensor array information and the relative pose of the robotic arm relative to the conductor is established, which realizes autonomous navigation of the robotic inspection line. The method can effectively reduce the influence of the detection error of the single electromagnetic sensor on the navigation control of the robot and improve the accuracy of the robot grasping control. Finally, the feasibility and effectiveness of this method are verified by experiments.