| To ensure the safety production of the coal cleaning plant, inspection work isindispensable. But long boring inspection work made the work quality decrease,which even will make the normal inspection times cannot be guaranteed. Based onthis, the inspection robot was designed to finish the inspection work at the fixed timeand places, which will improve the safety and efficiency of the inspection work. Therobot used the infrared distance measurement system to finish the automatic obstacleavoidance, used the visual system to finish the people detection, which protected thesafety of the workers. The contents of the paper consist of the fowling several aspects:(1) Considering the actual work environment, the robot worked in the coalcleaning plant, so the wheeled robot was selected to be the motion platform and itsstructure was designed. The robot used the chain transmission to complete thefour-wheel drive. And compared with belt transmission, there is no elastic slip andskid phenomenon, which will improve the efficiency and the average transmissionratio. To enhance the ground adhesion of the wheels, improve the damping effect anddecrease the shock to the car body, the robot adopted the independent suspensionsystem.(2)The realization of the robot control system. The robot control system mainlyconsists of camera, PC interface and lower computer. PC interface was completed byusing C#in the VS2008development environment, which consists of6modules. Andevery module independently communicates with lower computer by the ways ofRS485communication module. The work needed to do on the lower computerconsists of the controller smallest system design and motor drive controller design. Allthe above content will be introduced in detail.(3)The robot adopted potential field rejection method and multilayer perceptionstrategy method. In the paper, the potential field rejection method was used to makethe obstacle avoidance strategy table combined with the actual experiments, whichsimplified the amount of processing information and improved the reaction speed ofthe robot. The application of multilayer perception strategy method further simplifiedthe control of the robot motion condition. Four layers of perception layer defined fourstate of motion, which can help robot to efficiently judge its action.(4)The visual system helped the robot to realize the obstacle avoidance andinspected the moving worker near the running machine. The paper mainly analyzedthe algorithm about the motion detection and segmentation and at last chose the background deduction method based on Gaussian mixture model to inspect movingobject and the experiments verified that the method had good detection effect.At last, the robot was made and its functions were realized. And the experimentsproved that the robot had good working performance, which realized satisfieddetection effect. |
