In context
In early 2023, collaborative robots were pushing toward higher load-to-weight ratios and smoother motion control, but joint modules combining permanent-magnet motors with harmonic drives often suffered from speed ripple caused by torque pulsation and transmission errors. This paper from the Ningbo Institute of Materials Technology and Engineering addressed that challenge with a novel joint design and control strategy.
What was reported
Researchers designed a collaborative robot joint integrating a double-stator permanent magnet synchronous motor and a harmonic drive reducer. The double-stator structure increased torque density by 11% compared with a single-stator design, as verified by dynamometer tests.
To improve speed control accuracy, they modeled the combined effects of friction, motor torque ripple, and harmonic drive transmission error as an equivalent input torque disturbance. They then proposed an adaptive disturbance observer using a dual-loop architecture: an inner disturbance observer compensated for low-frequency nonlinearities and modeling errors, while an outer loop combined a disturbance observer with an adaptive algorithm to estimate and feedforward-compensate periodic disturbances.
Experiments on the joint prototype showed that compared with a conventional proportional-integral speed controller, the root mean square of low-speed steady-state fluctuation errors dropped by 40%–60%, and by about 30%–40% at medium and high speeds. Frequency analysis identified ripple components at multiples of motor speed, linked to cogging torque and harmonic drive effects.
Why it mattered
This work demonstrated a practical path to higher torque density and finer speed control in collaborative robot joints, which is critical for applications requiring precise, low-speed manipulation. The adaptive approach reduces reliance on accurate models, potentially simplifying controller design for industrial robots.
Compared with the conventional proportional-integral (PI) speed controller, the root mean square (RMS) of the joint low-speed steady-state fluctuation errors is reduced by 40%~60%.
Source: 《机器人》期刊 (robot.sia.cn) · Published 2023-03-06 · “基于自适应干扰估测器的协作机器人关节速度波动抑制方法”