Font Size: a A A

Design And Development Of Autonomous Ball Collecting Robot Based On Visual Navigation

Posted on:2020-10-22Degree:MasterType:Thesis
Country:ChinaCandidate:L PengFull Text:PDF
GTID:2428330620455962Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
In tennis training and practice,it is time-consuming and laborious to collect the scattered balls all over the court,which turns people's enthusiasm back to dust.Manual operation is still the main method of ball picking mechanism in current market.Although the labor intensity is reduced to some extent,the manual participation in tennis ball collecting is not changed substantially.For this reason,an intelligent ball collecting robot based on computer vision is designed and developed,which is capable of moving autonomously and completing ball picking task independently.Firstly,the function of vision-based ball collecting robot is analyzed to set up the overall design scheme.After completing the mechanical structure design,the ball picking mechanism is verified on the basis of kinematics analysis and experiment.Secondly,the hardware structure of controlling system is designed according to the functional requirements.In respect of software design,the finite state machine method is adopted to realize modular programming.An adaptive neuron PID control algorithm is applied to achieve the close-loop control of attitude during the motion process.Then,an object recognition algorithm based on color and contour features is proposed for tennis ball detection.In order to enhance the accuracy and stability of recognition,an improved least squares method based on contour statistical features is proposed for contour detection.The recognition performance is validated experimentally.Finally,the positioning and path planning of ball collecting robot are studied.The coordinate space mapping relationship is established according to the imaging model to realize the relative localization of detected tennis ball.The combination of dead reckoning and UWB location is applied for the self-localization of the robot.On this basis,the complete coverage path planning algorithm based on rolling window method of local visual navigation and sub-region traversal is proposed.The autonomous ball collecting function based on visual navigation of the robot is verified through the field experiment on tennis court.
Keywords/Search Tags:ball collecting robot, visual navigation, object recognition, localization, path plannig
PDF Full Text Request
Related items