| With the successful hosting of the Beijing Winter Olympics,ice and snow sports have once again set off a boom among the public.Among the many sports,curling has attracted a lot of attention with its unique charm.Heilongjiang,as a major province in the ice and snow industry,in order to welcome the Olympic Games and help the Winter Olympics of science and technology,launched the intelligent curling robot project for the 2022 Beijing Winter Olympics and the Ice and Snow Industry in Longjiang,and this project is derived from this.In this paper,a curling throwing robot is studied and analyzed,and the robot motion control based on the visual method is adopted,and the development of the curling robot system is finally completed.Firstly,this paper introduces the hardware selection and composition of the curling robot,explains the implementation plan,and elaborates the workflow of the curling robot’s one-time throwing task.The hardware composition is mainly divided into five parts,namely: the vehicle central control unit,the vision system,the clamping mechanism,the motion mechanism and the rotating mechanism;the workflow is mainly divided into two parts,namely the in-situ aiming stage and the forward acceleration throwing stage.The technical difficulty lies in the control of the motor mechanism based on the vision system.Secondly,the detection and tracking technology used in the curling robot vision system is introduced.It is divided into two parts: the detection technology used to specify the target area and the tracking method of the target area during motion.In addition to the current manual frame selection,the specified target area is also designed with an automatic detection method,which is based on YOLO-v5,and establishes a curling dataset based on narrow field of view for automatic curling detection;based on the consideration of real-time and accuracy,the tracking of the target area uses the KCF method,and combines the experimental data to verify the use effect of KCF.Thirdly,the vision-based motion control method is elaborated.In this project,the motion control method based on the visual servo method based on image features is used,and the motion correction method suitable for curling robots is proposed with reference to the classical PID control theory.Finally,the software architecture and experimental testing of the curling robot system are carried out.It mainly elaborates the software workflow,introduces the experimental design scheme,and analyzes the experimental results to propose targeted improvement schemes.The curling robot of this project has been proved by many field tests to have a high level of ability to complete the throwing task under the conditions of a certain experimental site environment. |
