Research Of Semi-intelligent Robot's Control And Load Effect

The Semi-intelligent Bomb-Disposal Robot discussed in this thesis is an advanced robot with independent intellectual property rights. It is developed based on the research of similar kind types of robots at home and abroad composed with the newest technologies nowadays, and also draw on the results of previous research in this laboratory.The Robot's manipulator is designed into three booms and five freedoms to meet the actual capture requirements. In this thesis a simplified analytical model has been made out to solve this less freedom manipulator, and for this model we use the D-H method to derive the forward and inverse kinematics algorithm. Finally, we use the algorithm to program the automatically capture part of the robot in the underlying control system. Compared with the original complex spatial matrix calculation, the new algorithm can be certificated more effective. So it provides a good guarantee for the robot to capture the object accurately.The executive mechanism of robot's manipulator is brushless DC, the state feedback control system according to the optimal linear speed control system is a static error system, to reduce the static differences of load and disturbance, and improve the dynamic response characteristics in case of sudden load, in this paper, we introduce the observer to estimate the load torque, and use the reduced observer to compensate the system when a sudden load occur, then we use it with optimal control system to constitute a dual-channel control system.The speed and current dual-loop feedback controller is widely used in nowadays industry, and also could get good performances, but for the robot which is composed by multiple joints and also needs for more high requirements, this kind of controller will show its drawbacks. To ensure the robot will have the same stability and precision to located when loaded or unloaded or under sudden load, in this paper, we design a kind of reduced load torque observer to estimate the load torque, and also use the feed forward method to compensation the designed dual -loop feedback system. The experiement result shows that the feed forward compensation of load observer can successfully eliminate differences of disturbance, and making the speed and current dual-loop system both on following and rejection of disturbance aspects have an excellent performance.The robot's control system uses MatlabRTW rapid prototyping method to developed, it has multi-layer control structure, the host computer uses the Simulink to design the bottom control diagram, and uses the xPC target system to generate the optimization code, downloads the application program to the target computer PC/104,then composites the real time control system. The servo motor feedback control based on the signal of encoder, uses the data acquisition card to collect data, and achieves feedback control of DC motor. Finally let the robot coordinated to the aimed location. The field computer's main task is to collect the field image, calculates three-dimensional coordinate of the object, and also send the control commands to the bottom control layer, and achieves robot intelligent control action.Finally, to test the feasibility and robustness of the designed control system, we do a target Object orientation for crawling experiement,the results shows, the robot could accurately capture the object. Additionally, to prove the load observer we designed could effectively eliminate the loading effects of robot, we do two comparative experiements, it's the robot with load compensation to locate and without load compensation to locate. The results show, the system with observer could more accurately located the aimed target point, and successful eliminate the load effect on the robot performance.