Research On Multi-Channel Interaction System Of Mobile Detection Robot In Complex Environment

The mobile detection robot integrates the capabilities of environmental perception,decision-making and control.It can replace humans to complete some survey tasks in complex environments such as engineering survey and military reconnaissance.The operator can control the remote mobile robot locally to complete grasping and other operational tasks.However,the current teleoperation control grasping is mostly controlled by a single channel,and the grasping control method cannot well reflect the autonomy and the operator’s target selectivity.The application in the field of mobile detection robot is of great significance.This topic analyzes the development status and research trend of multi-channel interaction technology,and uses the mobile platform as the carrier to build a teleoperated robotic arm system alone.The main contents of the work are as follows:The manipulator teleoperation system is designed,and the multi-channel fusion algorithm is used to achieve the goal of teleoperation control.The local end receives the remote Kinect image in real time,and combines with the robotic arm model to construct a remote virtual scene;at the same time,the local end Kinect gesture information is collected and loaded into the virtual scene,the direction indicated by the local end is adjusted to provide control information,and the local end and the remote end communicate wirelessly.transmission.At the same time,in order to improve the efficiency of teleoperation control,a fusion algorithm dominated by the main channel is designed,and a new mechanism is proposed that the main channel controls the fusion task generation and update,and the auxiliary channel controls the fusion task deletion.A localization method is designed in which the local end indicates the selection of the remote target object.The local end collects the remote Kinect image for target detection and loads it into the virtual scene.At the same time,the hand information of the local end is collected for key point detection,the finger-fingertip indication direction is loaded into the remote virtual scene as the target selection direction,and the indication direction is adjusted to select the target object of the virtual scene,and the local indication direction and the slave end are designed.The bounding volume intersection detection mechanism ensures the accuracy of selection.Finally,a multi-channel interactive system for teleoperation of the robotic arm is built,and a force feedback vibration module is designed to provide interactive feedback on the grasping effect of the robotic arm.The system adopts the voice channel as the main channel,and the gesture channel combined with target detection as the auxiliary channel.Finally,the remote operation target grasping experiments in different scenarios are designed,and the experimental results of the single channel and the main-auxiliary fusion channel are compared,which verifies the effectiveness and efficiency of the main-auxiliary channel fusion algorithm in this paper.