With the development of the city,the rooftop photovoltaic power station gradually has a large scale.In order to improve urban space utilization and light duration,rooftop photovoltaic panels are mainly distributed on the top of large factories and large warehouses.The construction features of photovoltaic panels are wide distribution area and small inclination angle,making it difficult to arrange large-scale cleaning devices.However,regular cleaning of photovoltaic panels is of great significance to improve power generation efficiency.As labor costs increase year by year,replacing workers with cleaning robots to clean photovoltaic panels has become an inevitable trend of development,but there is no in-depth research and design for cleaning robots for rooftop power stations.To solve this problem,this paper studies the design of roof photovoltaic panel cleaning robot and its autonomous control system.Firstly,this paper presents the design of omni-directional mobile robot for cleaning large roof photovoltaic panels.The robot uses a water bag and roller brush cleaning mechanism to clean the panel.The whole body of the robot has an integrated lightweight design,which is portable,does not need a complex installation process,and realizes the characteristics of flexible movement on the photovoltaic panel through the rotatable wheel mobile mechanism.Then,in order to realize the autonomous cleaning of the robot,the control system of the omni-directional mobile robot is designed in this paper.According to the characteristics of the mobile mechanism of the cleaning robot,the kinematics modeling is established,and then the cleaning trajectory is planned by the combination of full coverage search algorithm and a * search algorithm.Aiming at the problem of the friction drop of the photovoltaic panel after being wetted,the wheel slip condition of the photovoltaic cleaning robot was analyzed,and the model was built under disturbance.Aiming at the phenomenon of robot slip,this paper establishes the dynamic relationship between slip rate and longitudinal motion based on the magic tire formula,and realizes the online estimation method of traction coefficient based on Kalman filter and dynamic window method.Finally,according to the planned trajectory of the cleaning robot on the photovoltaic panel,a trajectory tracking algorithm suitable for low friction is designed in this paper.Because it is difficult to carry out accurate horizontal and longitudinal force control on the photovoltaic panel,a horizontal and longitudinal separation tracking control algorithm based on the kinematic model is proposed.Friction constraints are added to the kinematic model.For the case that the nonlinear constraints are not solvable,the cascade PID control of cleaning speed is carried out in the longitudinal direction,and then the predictive control model of transverse error is solved in the transverse space.Finally,the robot achieved the autonomous cleaning task of full coverage of the photovoltaic panel,and kept the speed tracking error below 5%and the trajectory tracking error within 0.08 m. |