| In recent years,various automatic ball picking products have gradually matured and put into use,bringing convenience to users and improving training effects.However,the badminton picking-devices currently on the market are less automated and may cause damage to the badminton during the ball picking process.Therefore,designing a robot that picks up badminton automatically and without damage has certain market value and social significance.Aiming at the problem that badminton is difficult to pick up in a non-damage way,this thesis proposes a new ball picking method,combined with multi-sensor data fusion technology,and developed an automatic badminton picking robot based on SoC.The main work of the system perception part design is as follows: the LIDAR and camera fixing components have been designed to maximize the range of viewing angles for sensor data collection;the interface module,data collection and transmission module are designed to realize the collection and transmission of sensor data;the data fusion program and badminton recognition program have been designed to realize the detection,positioning and attitude recognition of badminton;the ball-picking path planning algorithm was designed based on the idea of ant colony algorithm.The main work of the system motion control part design is as follows: the two-wheel differential moving structure chassis has been designed to make the picking-robot have flexible steering and fast movement functions;a lifting type wind control posture adjusting ball picking device has been designed so that the picking-robot has the characteristics of badminton nondamage picking;the motor control module is designed to realize the speed control and direction control of the moving parts of the ball picking robot;the motion control program is designed to realize finding and picking of the picking-robot.The innovations of the thesis are as follows:(1)Aiming at the problem that there is a small amount of information and unreliable detection when a single sensor detects and locates badminton,the method of generating candidate regions by fusion of information from LIDAR and visible light image sensors is used to achieve high reliability of badminton detection performance,low computational overhead,and also improved badminton positioning accuracy.(2)Aiming at the problem of the optimal solution of the picking speed caused by the randomness and time-varying badminton distribution,a path planning algorithm with the best picking speed is designed according to the ant colony algorithm.(3)Aiming at the problem that the badminton is irregular in shape and fragile,and it is difficult to realize mechanized picking,a lift-type wind-control posture-adjusting ball picking device capable of picking up the badminton without damage is designed to basically pick up the badminton without damage.A number of tests were performed on badminton-picking robots at an average standard density of 4 to 5 badmintons per square meter at indoor standard venues.The results showed that the average ball picking speed was 20 to 30 per minute,and the ball damage rate was less than 5%,its indicators have reached the expected goals.This thesis tested the picking-robot on standard courts,cement ground,and ceramic tile ground.The results show that the badminton-picking robot has strong environmental adaptability,high ball picking success rate,and low damage rate to the ball,which realizes the automatic picking and meets the design requirements. |
