Driven by the rapid development of computer and automation technologies, the performance of industrial robots has continuously improved. From their initial application in assembly and logistics handling, they are now widely used in high-precision applications such as machining, welding, and polishing. From 2013 to 2017, China's robot industry experienced a period of rapid growth, with its scale expanding continuously and an average growth rate of 30%. According to the "China Robot Industry Report (2018)" released by the China Electronics Society, the Chinese robot market reached a size of US$8.47 billion.
In industrial robot applications, servo systems using traditional control structures generate significant mechanical vibrations during high-dynamic trajectory commands, leading to reduced positioning accuracy and response speed. Prolonged vibration can cause excessive torque transmission from the robot's reducer, damaging the shaft system and reducing its lifespan; in severe cases, it can even render the robot unusable. As the power output system for industrial robots, the servo system should possess high instantaneous overload capacity, tracking accuracy, responsiveness, and good versatility and expandability. Therefore, reducing vibration is a key research focus in current industrial robot joint servo control systems.
In academia, control methods are primarily designed based on vibration models, often incorporating additional state feedback, including the reducer's flexible torque and the end effector position information of the industrial robot arm. Proportional resonant control, by feeding back the reducer's flexible torque into the control loop, effectively raises the system's natural mechanical frequency and moves it away from the system's control bandwidth, thus avoiding vibration. Sliding mode control and model predictive control have also been applied in robot servo control, but these algorithms differ significantly from traditional control structures, are more complex to implement, and heavily rely on model accuracy, thus limiting their widespread practical application.
Currently, the main method used in engineering applications is to add low-pass filters and notch filters to the existing control structure, and suppress vibration by setting an appropriate frequency. Adding a low-pass filter often significantly reduces the system bandwidth, which can reduce vibration to some extent, but sometimes there is a phenomenon where the mechanical arm vibrates significantly even though the motor side does not vibrate.
Notch filters can filter out signal components of a specified frequency. By setting an accurate notch frequency, command fluctuations caused by mechanical resonance characteristics within the servo control loop can be effectively filtered out, thus ensuring vibration-free motor output torque. However, when an industrial robot arm moves in space, the equivalent inertia of each joint is constantly changing, causing its resonant frequency to change accordingly. When the notch frequency of the notch filter is inconsistent with the resonant frequency, it cannot effectively suppress vibration and may even lead to system instability.
To suppress vibrations in low-stiffness robot servo systems, Huang Xuanrui, Song Yuyang, and Xiao Xi from the Department of Electrical Engineering and Applied Electronics at Tsinghua University, along with Li Qiusheng from Tsinghua Energy Dechuang Electric Technology (Beijing) Co., Ltd., proposed a vibration suppression algorithm for industrial robot joint servo systems based on the elastic transmission model of a single joint and the principle of internal model control. The research results were published in the 3rd issue of the 2019 *Journal of Electrical Engineering*, titled "A Vibration Suppression Algorithm for Industrial Robot Joint Servo Systems Based on Internal Model Control".
Low-stiffness industrial robot joint equivalent testing platform
This algorithm achieves vibration suppression by adding a vibration damping filter between the position and velocity loops of a traditional servo controller. It requires no additional sensors or complex control algorithms, and does not alter the basic control structure of the traditional servo controller. Simulation and experimental results show that this method is simple and feasible, effectively suppressing vibrations generated during positioning in low-stiffness robot joint systems. It also exhibits strong parameter robustness, is easily implemented in existing servo control systems, and has engineering applicability.
Internal model control (IMC) is a control method developed from chemical process control. It boasts advantages such as simple design principles and good robustness, and has gradually attracted the attention of scholars both domestically and internationally. Controllers designed based on IMC often have the structure of traditional feedback PI control, with an additional control loop added. The PI controller is used to adjust the system's following performance, while the additional control loop enhances the system's disturbance rejection capability. Therefore, it also offers the advantage of intuitive parameter tuning.
Disclaimer: This article is a reprint. If there are any copyright issues, please contact us promptly for deletion (QQ: 2737591964). We apologize for any inconvenience.
