Magnetic robots have shown great potential in small-scale applications due to their wireless control mode. However, existing efforts only deal with solid magnetic materials which could not deform. In this paper, we focus on integrated locomotion and deformation of a class of magnetic soft robot made of ferrofluid. To this end, the magnetic model and dynamics model that takes the nonlinearity into account are first established. Then the corresponding motion controller are proposed, based on the results of feedback linearization and frequency domain test results. Furthermore, an extended state observer (ESO) is designed to reduce the perturbation due to model uncertainties. By integrating the control strategies of locomotion and active deformation, we demonstrate that the soft robot possesses the capability of conducting complex tasks such as passing through narrow environment and transporting multiple objects. Various experiments are also performed to demonstrate the effectiveness of the proposed control methods.

Piezoelectric hysteresis is a critical issue that significantly degrades the motion accuracy of piezo-actuated nanopositioners. Such an issue is difficult to be precisely modeled and compensated for, primarily due to its asymmetric, rate, and input amplitude-dependent characteristics. This article proposes a novel method to deal with this challenge. Specifically, a polynomial-based fractional order disturbance model is proposed to accommodate and characterize the complex hysteresis effect. In this model, the rate dependence is captured by a general method of implementing curve fitting in Bode magnitude plot. The inverse model for control purposes is immediately available from the original one. The proposed method does not require expensive computational resources. In fact, this article shows that this controller can be easily implemented in an analog manner, which brings the advantages of high bandwidth and low cost. Extensive modeling and tracking experiments are carried out to demonstrate the effectiveness of the proposed method. It is shown that the piezoelectric hysteresis nonlinearity can be significantly suppressed over a wide bandwidth.

Nanopositioning stages driven by piezoelectric actuator have been employed across a wide spectrum of applications. However, the challenging issues they are suffering include hysteresis and creep effects, along with lightly damped mechanical resonances. In view of these issues, this article presents a novel hierarchical antidisturbance control solution. The proposed strategy seamlessly integrates three antidisturbance components, i.e., a feedforward-type disturbance observer (DO), a feedback-type DO, and an error-based active disturbance rejection controller (EB-ADRC) into a single framework. Such a strategy loosens the bandwidth constraint of DO imposed by mechanical resonance and speeds up the response time of typical ADRC without relying on high-order derivatives of reference input. The inclusion of these antidisturbance components does not complicate the controller tuning procedure. In fact, it is shown that only one parameter needs to be tuned in practice. Moreover, the stability analysis of EB-ADRC is provided within DO framework. The superior control performance of the proposed strategy is confirmed via extensive comparative studies with integral resonant control, positive position feedback, and linear active disturbance rejection control strategies.

Charge control is a well-known sensorless approach to operate piezoelectric actuators, which has been proposed for more than 30 years. However, it is rarely used in industry because the implemented controllers suffer from the issues of limited low-frequency performance, long settling time, floating-load, and loss of stroke, etc. In this paper, a novel controller circuit dedicated to overcome these issues is presented. In the proposed scheme, a grounded-load charge controller with decoupled configuration is developed, which separates high-frequency and low-frequency paths, thus achieving arbitrarily low transition frequency without increasing the settling time. Based on this, a self-compensating configuration is further proposed and integrated into the controller circuit, which makes full use of controller output to improve its own control performance at low frequencies. Experimental results show that the presented charge controller can effectively reduce more than 88% of the hysteretic nonlinearity even when operating close to the transition frequency. To demonstrate its practical value, we then integrate it into a custom-designed high-speed atomic force microscope system. By comparing images obtained from using voltage drive and charge controller, it is clear that the piezoelectric hysteresis has been significantly reduced in a wide bandwidth.