Robot control can be divided into joint space control and Cartesian space control. For serial multi-joint robots, joint space control controls the variables of each joint, while Cartesian space control controls the variables of the robot's end effector. Based on the different control variables, robot control can be categorized into: position control, velocity control, acceleration control, force control, force-position hybrid control, and vibration control, among others.
Depending on the task, robot control methods can be divided into four types: point-to-point control, continuous trajectory control, force (torque) control, and intelligent control. This article introduces these four control methods based on the task being performed.
1. Point-to-Point (PTP) Control Method
Point-to-point control has a wide range of applications in the fields of mechatronics and robotics. Typical applications of point-to-point control systems include CNC machine tools tracking the contours of parts in the machinery manufacturing industry, fingertip trajectory control of industrial robots, and path tracking of walking robots.
During control, industrial robots are required to move quickly and accurately between adjacent points, but no specific rules are set for their trajectory to reach the target point.
Positioning accuracy and the time required for movement are the two main technical indicators of this control method. This control method is easy to implement and does not require high positioning accuracy; therefore, it is often used in operations such as loading and unloading, material handling, spot welding, and inserting components onto circuit boards—where only the accurate orientation of the end effector at the target point is required. This method is relatively simple, but achieving a positioning accuracy of 2–3 μm is quite difficult.
Point-to-point control systems are actually a type of position servo system. Their basic structure and composition are basically the same, only the focus is different, and their control complexity also varies. According to the feedback method, they can be divided into closed-loop systems, semi-closed-loop systems and open-loop systems.
2. Continuous trajectory control (CP) mode
Under PTP point control, the initial and final speeds are 0, and various speed planning methods are possible during the process.
CP control is the continuous control of the pose of the end effector of an industrial robot in the workspace. The velocity at intermediate points is not zero, and the movement is continuous. The velocity magnitude at each point is obtained through velocity look-ahead. Generally, continuous trajectory control mainly uses velocity look-ahead methods: forward velocity limit, angular velocity limit, backtracking velocity limit, maximum velocity limit, and contour error velocity limit.
This control method requires the device to move strictly according to a predetermined trajectory and speed within a certain precision range, and the speed must be controllable, the trajectory smooth, and the movement stable in order to complete the task.
Industrial robots perform continuous and synchronous joint movements, allowing their end effectors to form continuous trajectories. The main technical indicators of this control method are the trajectory tracking accuracy and stability of the end effector's pose. This control method is commonly used in robots performing arc welding, painting, deburring, and inspection operations.
3. Force (torque) control method
As the application boundaries of robots continue to expand, vision-based capabilities alone are no longer sufficient to meet the needs of complex practical applications. In such cases, it is necessary to introduce force/torque control output or to introduce force/torque as a closed-loop feedback quantity into the control.
When performing tasks such as assembly and object handling, in addition to accurate positioning, the applied force or torque must be appropriate. This necessitates the use of a torque servo system. The principle of this control method is essentially the same as that of position servo control, except that the input and feedback signals are force (torque) signals instead of position signals. Therefore, a force (torque) sensor is required in this system. Sometimes, proximity and sliding sensors are also used for adaptive control.
Since the contact between the robotic arm and the working surface is often an unknown and complex curved surface, the perception of force/torque should also have multi-dimensional capabilities.
4. Intelligent control method
Intelligent control of robots is a control method that features intelligent information processing, intelligent information feedback, and intelligent control decision-making. It obtains knowledge of the surrounding environment through sensors (such as cameras, image sensors, ultrasonic transducers, lasers, conductive rubber, piezoelectric elements, pneumatic elements, limit switches, and other electromechanical components) and makes corresponding decisions based on its own internal knowledge base.
The development of intelligent control technology relies on the rapid advancements in artificial intelligence, including artificial neural networks, gene algorithms, genetic algorithms, and expert systems. In recent years, intelligent control technology has made significant progress; fuzzy control theory and artificial neural network theory, as well as their integration, have greatly improved the speed and accuracy of robots. Major applications include multi-joint robot tracking control, lunar robot control, weeding robot control, and cooking robot control.
Robot intelligent control can be further subdivided into: fuzzy control, adaptive control, optimal control, neural network control, fuzzy neural network control, expert control, etc.
With the support of intelligent control technology, industrial robots can truly become intelligent, but this is also the most difficult to achieve, as it is heavily reliant on algorithms and components.
Currently, most industrial robots are still at a relatively basic stage of spatial positioning and control, lacking significant intelligence and having a long way to go before achieving true intelligence. Therefore, Chinese robotics experts, based on their application environments, categorize robots into two main types: industrial robots and intelligent robots.
