Design And Implementation Of The Following Robot System Based On Linear Laser Imaging

In "Made in China 2025" issued by the State Council,it is pointed out that robots are one of the key areas and robot technology should be promoted.Service robot is a branch of robot application,and its following technology has a wide application prospect in mobile robot.In this paper,a new following robot system is designed.The method of target location of monocular camera combined with linear laser is proposed.Combined with the optimized green laser segmentation scheme,the interference of uneven light in the windowed room can be effectively solved,and the interference of light blue can be filtered out.The spatial position relationship between the target pedestrian and the robot can be determined only with a single frame image.This method effectively improves the efficiency of vision-processing,and completes the motion control of the robot according to the position relationship between the robot and the target.When the target is lost,the alarm is triggered through Bluetooth communication between robot and mobile phone.The following robot system consists of vision-processing module,motion control module and loss alarm module.In the vision-processing module,a green linear laser with a wavelength of 515 nm is tilted upward in the field of vision of the monocular camera,and the distance between the target pedestrian and the robot is reflected in the height of laser projection to complete the single frame positioning.In RGB color image,G-R and G-B features are extracted respectively.Otsu algorithm is used to pre segment the two features,and then the intersection is used to get the laser segmentation results,and the target position is updated by using the rules of the combination of front and back neighborhood and Bhattacharyya coefficient.The visionprocessing module can also complete the robust tracking of the target under the occlusion of the robot and the target pedestrian,which makes the vision judgment more accurate.The motion control module adopts the command-based motion control mode.By using STM32 single-chip microcomputer to control the dual DC motor,it can realize the close follow of the target pedestrian at different speeds.The feedback of robot's motion speed is obtained by the encoder's signal acquisition.12 V,5V and 3.3V power interfaces are reserved in the power supply part of the motion control module to supply power to the whole system.The loss alarm module uses Bluetooth communication between robot and mobile phone to complete the loss alarm function,and develops a specific application on Android mobile phone to achieve this function.On average,the mobile terminal can receive Bluetooth signal from the robot terminal every 50 milliseconds.If the mobile terminal does not receive the confirmation signal from Bluetooth for more than 3 seconds,it will start the vibration and popup alarm response of the mobile phone.After the system test,in terms of the efficiency of visual algorithm,the simulation is carried out on a computer with Intel I5 CPU and 8G memory.The test result shows that total 349 images can be processed in 16.16 seconds,and one frame can be processed in 46.3 milliseconds on average.The operation efficiency of the algorithm is high.Compared with the classical mean shift algorithm,the algorithm is a little slower,but it can still maintain its accuracy when the target is partially occluded;compared with the optical flow algorithm,the algorithm is faster and can get the target center point at one time.The robot can follow the target pedestrian who can walk in a straight line with a maximum speed of 1.6m/s and turn at an angle with a maximum speed of 2.44rad/s.The following robot design scheme proposed in this paper has the advantages of high efficiency of visual processing algorithm,strong anti-interference ability and high cost performance,and has practical significance and application value.