Design And Implementation Of Tracking Robot Based On ROS

Robots with object tracking function are widely used in real life,such as security,military and other fields.Among them,target tracking algorithm is the core algorithm of tracking robots,which directly affects the tracking effect of robots.In order to further improve the overall performance of tracking robots,a new target tracking algorithm based on convolutional neural network is transplanted into robots based on robot operating system,and its design and implementation are described in detail.The paper constructs the overall design block diagram of the tracking robot,clarifies the robot function requirements and key algorithms required,such as object tracking algorithm,motion control algorithm,and elaborates the related algorithms.After comprehensively investigating a variety of target tracking algorithms from the aspects of accuracy and robustness,we finally select the SiamRPN algorithm based on convolutional neural network as the target tracking algorithm of tracking robot,and use PID control algorithm as the motion control algorithm.The thesis deeply researches and designs the hardware and software system of tracking robot,and constructs the hardware and software system block diagram.In hardware system part we elaborated the principle,selection and use of some hardware like mobile platform,main control chip,motor drive and various types of sensors in the robot system,including ultrasonic sensor,Kinect,etc.,at the same time we designed some hardware circuits.The software system part is based on the ROS development platform,using Python as the main development language,design and implementation of various functional nodes in the robot system,including motion control nodes,target tracking nodes,ultrasonic obstacle avoidance nodes.Independent of each other,the robot's software system is highly modular and easy to debug and update.Finally,the prototype of the tracking robot was built and debugged and experimented,including PID adjustment experiment,motion control experiment,target tracking experiment and so on.The PID parameters of the mobile chassis are determined experimentally,and the target tracking and obstacle avoidance capabilities of the tracking robot are verified.The experimental results show that the tracking robot designed by this subject can complete the simple target tracking task and has certain obstacle avoidance ability,and initially meet the subject requirements.