High Precision Sensorless Control of Rotor Mechanical Position Based on Cross Correction of Dual Air Gap Information

Position sensors are indispensable in robotic joint servo systems for acquiring mechanical positions, yet their installation inevitably occupies an axial space and increases system complexity, limiting their applicability in compact robot design where spatial constraints and integration efficiency are critical. Sensorless control reduces mechanical and circuit complexity through hardware simplification, but inherently estimates only the electrical instead of mechanical rotor position information, thus remaining constrained in robot joint control applications. Based on the previously proposed dual-gap dual-pole composite machine (DDCM), this paper systematically analyzes the causes of mechanical position estimation errors and proposes a correction method that utilizes a correction coefficient to reduce these errors and enhance estimation accuracy. Furthermore, this paper derives the applicability constraints of the proposed scheme, demonstrating that its requirements for electrical angle position errors are not stringent, thus enabling wide applicability in conventional sensorless control scenarios. The effectiveness of the proposed method is verified by conducting experiments on a 0.75 kW prototype.