遥操作机器人的神经网络校正地图策略

Neural Network Correction Map Strategy of Teleoperation Robot

提出一种利用全身传感装置进行地图定位和校正的遥操作机器人。首先,针对遥操作控制系统进行了深入的研究,包括遥操作要素和技术要点等。其次,根据周围环境地图特征进行了地图建立和定位,提出了一种神经网络的方法进行地图校正。该方法可以学习来自信息捕捉装置的传感器数据信息,分配给每个神经网络的机器人执行器的位置之间的映射。最后,为了在学习过程中收集数据,机器人通过一系列成对的同步动作来校正地图。该方法可以应用于任何环境下机器人地图定位和校准,而不考虑环境物理中呈现出的差异。实验结果表明,该方法为实现机器人的遥操作提供了一种快速、有效、灵活的方法。

This paper proposes a teleoperation robot that uses a full-body sensing device for map positioning and correction.First,this article has conducted in-depth research on the teleoperation control system,including teleoperation elements and technical points.Secondly,according to the characteristics of the surrounding environment map,the map establishment and positioning are carried out,and a neural network method is proposed to correct the map.The method can learn the sensor data information from the information capture device and assign the mapping between the positions of the robot actuators of each neural network.Finally,in order to collect data during the learning process,the robot corrects the map through a series of synchronized actions in pairs.This method can be applied to robot map positioning and calibration in any environment,regardless of the differences presented in the environment physics.The experimental results show that the method provides a fast,effective and flexible method for remote operation of robots.