Search

Estimating contact force for a robot manipulator without an force/torque (F/T) sensor poses challenges due to uncertain torques, such as backlash and flexibility. To address this limitation, this article proposes a data-driven uncertain torque model and an overall gray-box structured approach. The contributions of this article are threefold. First, a joint domain unified neural networks (DUNNs)-based model is proposed to compensate for the uncertain torques. This model effectively captures uncertain torques beyond studies focusing solely on individual uncertainty. Second, the DUNNs model receives dynamic and joint domain information, enabling a single DUNNs model to estimate all joint uncertain torques through joint domain knowledge. This approach reduces the model size while maintaining performance. Third, the structure in which the DUNNs model works with conventional static friction models is introduced. This structure improves contact force estimation performance and enhances robustness against untrained data compared with the black-box model. Experimental results verify the method's effectiveness.

This article presents a single shunt current sensing drive method for ultralow inductance permanent magnet synchronous motors that can reliably reconstruct phase currents even at very low speeds. Single current shunt inverters using dc current measurement entail the problem of having dead zones where the phase currents cannot be reconstructed. Voltage injection at the switching frequency can be a solution to avoid this, but at the expense of the newly introduced current ripple. That current ripple is exacerbated in ultralow inductance motors and can cause drive failure, especially at low speeds. To rule out voltage injection, this article presents a drive scheme that combines sequential two-phase excitation with a dc current averaging method using an RC low-pass filter. This permits three-phase current reconstruction from the average of dc current measurements per half-switching period, regardless of the active vector duration. Experimental demonstrations in this article show that the proposed driving method allows reliable current reconstruction at both startup and low speed without a significant current ripple. The current reconstruction and dynamic control performance are demonstrated with both steady-state and transient experiments.

In vector-controlled drive systems, a position sensorless drive with a single current sensor (SD-SCS) is an attractive system due to the minimum number of sensors. However, when the saliency-based SD-SCS is applied, the position estimation accuracy is deteriorated due to the errors in the injected pulsewidth-modulated voltage and the reconstructed current. This article proposes a current-derivative-based sensorless control using a three active vector pulsewidth modulation (TAVPWM) to enhance the saliency-based SD-SCS. First, the TAVPWM scheme, which eliminates the current reconstruction dead zone in the low-modulation region, is adopted to improve the low-speed drive. Next, the derivatives of the phase currents are utilized and analyzed to minimize the position estimation error in the overall operating regions. Finally, the transition strategy with respect to the speed is proposed for the full-speed sensorless drive. Owing to the proposed method, the switching frequency increases up to 16 kHz and the audible noise is virtually eliminated in the saliency-based SD-SCS. Furthermore, the enhanced dynamic performance is verified by various experimental results.