| For the sake of beauty and the convenience of daylighting,modern high-rise buildings often adopt glass curtain wall structure on their wall surface,following is the glass curtain wall cleaning problem.The cleaning of the glass curtain wall is a heavy and dangerous work.At present,the cleaning of the glass curtain wall is mainly completed by the cleaning workers on the hanging basket or the waist rope.This manual cleaning method is not only dangerous,but also low in efficiency and high in cost.With the maturity of robot technology,using robot cleaning has become an inevitable trend of development.Based on the analysis of the technical requirements of the glass curtain wall cleaning task,this paper designs and produces a cheap and simple negative pressure adsorption wheel driven cleaning robot prototype,focusing on the research of the robot control system.The whole control system of cleaning robot adopts the form of upper and lower computer,uses high speed low power STM32F429 microcontroller as the machine master controller,and uses negative pressure sensor,ultrasonic sensor,nine-axis sensor,image sensor and other sensors to collect the robot's motion state and the surrounding environment system,and realizes the transmission of robot parameters to the upper computer and control commands to the lower computer through RS485 communication.In order to achieve the purpose of obstacle avoidance,robot adopts more ultrasonic sensors with driver module to realize the autonomous obstacle avoidance.At the same time,in order to facilitate operators to observe the surrounding environment of the robot,especially the cleaning effect of the glass curtain wall,discrete cosine change is used to process the image collected by the image sensor,and then transmitted to the upper computer.Finally,reliable and clear images are obtained through image restoration.The robot usually controls the hardware operation by means of background program,but it is only suitable for the embedded system with less tasks and less real-time requirement.In order to take full advantage of the 32-bit controller performance,meet the requirement of robot in real time,and at the same time convenient for subsequent robot task module extension,the real-time operating system ?C/OS? is transplanted to the control platform,so as to realize real-time task processing.The most important task of glass curtain wall cleaning is to wipe the area that can be reached by no obstacles or robots as much as possible,which is essentially a completecoverage problem.Aiming at the problem of complete-coverage path planning of mobile robot,an improved heuristic rule with ant colony algorithm is proposed.The algorithm firstly constructs a working environment based on the dynamic grid method through sensors on the robot body.A new priority heuristic rule that covers the entire path is constructed,considering the attributes of the grid,the robot steering,the distance between the neighboring grids,and the size of the uncovered area.In order to solve the problem of deadlock,this paper proposes an ant colony algorithm to find the optimal path and escape from the dead zone,so as to ensure the robot achieving complete-coverage and making the coverage path repetition rate as small as possible.Compared with the existing algorithms,simulation results are used to verify that the proposed algorithm has reduced the numbers of deadlock and trajectory repetition rate while ensuring the area coverage rate is 100%,and the work efficiency of robot is improved.Finally,the reliability and stability of the whole control system design are verified through experiments,including communication experiment,autonomous obstacle avoidance experiment and cleaning experiment.The results show that the designed control system achieves the desired effect and provides a reference for the control system scheme of glass curtain wall cleaning robot. |
PDF Full Text Request