| This research takes the first Chinese 2-boom tunnel rock-drilling robot as the research object. In order to automatize the rock-drilling rig, this research comes down to kinematics modeling, carriage positioning, kinematic inversing, positioning error analysis, error compensation, and developing a control system of the rock-drilling robot.1. According to the unique characteristics of the boom of the two-boom tunnel rock-drilling robot, which is a coupled redundant manipulator with nine degrees of freedom, combined with the actual working environment characteristics, the Denavit-Hartenberg method is used to establish the modeling of rock-drilling robot and its working environment. The positive kinematics equation from the manipulator to the work object (the tunnel cross-section) is obtained. In allusion to the unique characteristics, a linear decoupled and iteration algorithm is put forward to change the complex position and pose relationship into linear approximation relationship, and decouple the complex and coupling multiple-input/ multiple-output(MIMO) model into many single-input/single-output(SISO) models. And then an iteration algorithm is applied to solve the inverse kinematics problem. A large number of simulation and practical application verify that this method can meet the kinematics precision and real-time requirements of the tunnel rock-drilling robot.2. Carriage positioning is essential to ensure the boom's positioning accuracy. Refered to the laser beam, which is fixed on the wall of the tunnel, the relational matrix between the laser coordinate system and the end-rod coordinate system is derived, and ultimately the coordinate transformation matrix from the rig to the tunnel cross-section was derived, thus the problem of precise carriage positioning is solved. In accordance with the special conditions, two simple positioning ways are respectively got.3. Through a large number of tests, large errors were found in the boring orientation of the manipulator, that has a strong impact on the accuracy of the tunnel construction. In order to analyze the contribution of various error factors, some experimental study and analysis are carried out on errors, which caused by the telescopic arm, the feeding beam, the rotate/flip shaft, and the output signal of sensors. The test results show that these parts could cause larger position and pose error, and among these error factors, the systematic error is the main factor, accompanied by random error. Main error factors include instruction errors, etection error, link flexibility, mechanical drive error and machining error, and so on.4. The position and pose errors are try to be formulized. Compution and analysis were carried on into machining and detection error, the arm flexibility, the feeding beam flexibility, the positioning accuracy of the rotate/flip shaft, etc. And the results show that, the error of the arm positioning comes mainly from the arm flexibility, but the error of feeding beam positioning comes from many reasons, the beam flexibility just takes a small part, and the compressibility of hydraulic oil is the main reason for the error of the rotate/flip shaft positioning.5. The concept of virtual joints is introduced into study on the error compensation of the manipulator positioning. In accordance with the different of the error causes and error characteristics, different methods are applied respectively to compensate the error, for instance, taking compensation directly into coordinate transformation matrix, adding the coordinate transformation matrix, adding the virtual joint, reducing the sensor error by filtering and correcting system error. Because it is difficult to establish the mathematical model of the error caused by the telescopic feeding beam, a GRNN (General Regression Neural Network) method is applied to compute the error, and satisfactory results have been obtained.6. The hardware structure of rock-drilling robot computer control system is designed as master-slave control structure, and the communication problem between the computer and Siemens CPU314 under Windows3.X is settled. With cutting the Chinese Windows3.2 to the minimum operating system, a tunnel rock-drilling robot control system has been developed in Borland C++4.5, and smooth automatically moving of a 2-ton boom is achieved in the process of positioning, and automatically drilling comes true.
|
PDF Full Text Request