Robot joint motors (hereinafter referred to as “joint motors”) are characterized by miniaturization, light weight, low speed, high torque, and low torque fluctuation [
1]. When the robot is walking, running, jumping, and other actions, the joint motor is often in short-time high-overload operation and the motor torque overload multiplier can reach more than 3–5 times the rated load; at this time the motor is in high saturation operation of the magnetic field, and the magnetic saturation of the motor will lead to increased current harmonics, torque to current ratio coefficient drop, and large torque fluctuations. In order to achieve smooth operation of quadruped robots, it is extremely important to optimize torque fluctuations under overload conditions for articulated motors that are often subjected to high overload operating conditions.
Joint motors often use integrated modules of motor and retarder in order to obtain sufficient torque, and this method can increase torque by reducing the speed of the motor. However, the modular design will increase the overall size and cause difficulties in assembly, while a larger reduction ratio will reduce the response speed of the motor. Therefore, by optimizing the motor stator and rotor parameters to increase the motor output torque, the speed of response can be improved by reducing the reduction ratio, and the overall size of the joint motor can be reduced, while the overall weight can be reduced.
In the design of high overload motor, the traditional design method of electromagnetic scheme can no longer meet the needs of high overload working conditions. The design method of a high overload motor can be obtained by comprehensive analysis of the influence law of stator-rotor structure parameters such as pole-slot combination, magnet thickness, stator split ratio, tooth width on output torque [
2,
3], and the temperature rise variation law under different operating conditions [
4]. Y. K. Cao et al. of Wuhan University [
5] combined the characteristics of quadruped robot joint motors, and comprehensively analyzed and compared the effects of four slot-pole combinations on motor performance from the perspectives of average torque, cogging torque, winding factor, back-emf, air-gap flux density, and unbalance force. Scholars from the Chinese Academy of Sciences analyzed the effects of electrical and magnetic loads on the overload capacity of the motor, as well as the degree of variation of stator current, rotor electric density and iron core magnetic density under overload conditions from the electromagnetic torque of the asynchronous motor, and elucidated the variation law between electromagnetic parameters [
6]. In addition, the loss distribution of the motor under overload operation was statistically processed by establishing the copper and iron loss calculation models for the optimal design of asynchronous motor overload capacity enhancement [
7]. S. W. Hwang of Korea [
8] designed a surface-mounted permanent magnet synchronous motor from the electromagnetic and thermal perspectives based on the lumped parameter thermal network according to the electromagnetic performance and dimensional constraints of wearable robots. S. Zhang and Q. Chen identified the key structural parameters affecting the air-gap flux density and the total harmonic distortion of the output torque by multi-target sensitivity analysis, which improved the topology of the motor and enhanced the electromagnetic characteristics of the motor [
9,
10]. In terms of reducing motor torque fluctuations, magnetic circuit adjustment [
11] and direct torque control [
12] minimize torque fluctuations by improving the number of stator slots [
13] and stator shape [
14]. Overload capability and torque fluctuation are the focal points of joint motors, but how to optimize the motor body design to improve motor output torque while reducing torque fluctuation under overload conditions has been less studied; moreover, due to the mutual coupling of the stator and rotor magnetic fields when the motor is loaded, the conventional calculation method cannot accurately separate the fluctuating component of the load torque.
To address this situation, this paper makes the following contributions to the design methodology of joint motors for quadruped robots with high overload and low torque fluctuation as the design goal: