The robot is structurally a space open chain mechanism, in which the motion of each joint is independent. In order to realize the motion track of the end point, the motion coordination of multiple joints is required, and the control system is much more complicated than the ordinary control system.
The characteristics of the robot control system are as follows:
(1) The control of the robot is closely related to the kinematics and dynamics of the mechanism. The state of the robot's hand and foot can be described in various coordinates. The reference coordinate system should be selected according to specific needs, and appropriate coordinate transformation should be done. Forward kinematics and inverse kinematics are often required, in addition to the effects of inertial forces, external forces (including gravity), and centripetal forces.
(2) Even a simpler robot requires at least 3 to 5 degrees of freedom, and a more complex robot requires a dozen or even dozens of degrees of freedom. Each degree of freedom generally includes a servo mechanism that must be coordinated to form a multivariable control system.
(3) The computer realizes the coordinated control of multiple independent servo systems and enables the robot to act according to the will of the person, and even gives the robot a certain "smart" task. Therefore, the robot control system must be a computer control system. At the same time, computer software is taking on a daunting task.
(4) Since the description of the state and motion of the robot is a nonlinear mathematical model, as the state changes and the external force changes, its parameters also change, and there is coupling between the variables. Therefore, it is not enough to use only the position loop, and it is necessary to adopt a speed or even an acceleration loop. Gravity compensation, feedforward, decoupling or adaptive control methods are often used in the system.
(5) Since the movement of the robot can often be done in different ways and paths, there is an "optimal" problem. For higher-level robots, artificial intelligence can be used to build a huge information base using computers, and to control, make decisions, manage and operate with the help of information bases.
According to the sensor and pattern recognition method, the working conditions of the object and the environment are obtained, and the optimal control law is automatically selected according to the requirements of the given index. In summary, the robot's control system is a coupled, nonlinear multivariable control system closely related to the kinematics and dynamics principles. Because of its particularity, both classical control theory and modern control theory cannot be copied. So far, robot control theory is not complete enough and systematic.