This paper presents a flexible actuator with variable stiffness and its associated decoupling control algorithm. The elastic cell consists of two opposite-handed torsion springs and two cam-bearing followers (CBFs). Stiffness regulation is realized by changing the contact point position between the torsion spring and the CBFs. The CBFs are pulled via wire rope transmission and their movement direction is perpendicular to the torsion direction of the springs, thereby resulting in rolling friction. Such a design ensures a wide range of stiffness variations and small energy consumption for stiffness regulation, and contributes to a compact actuator design. Experimental results demonstrate that the proposed mechanism has good stiffness regulation and can reduce stiffness rapidly after collision. The dynamics of load side and motor side are controlled separately by a decoupling control algorithm based on the projection between the virtual torque and the commands of motor position and velocity. The virtual torque required to track the joint trajectory is calculated according to a hybrid proportional-derivative control algorithm, which incorporates a disturbance compensator composed of a smooth saturation function. The spring-damper model is successfully applied to the projection between the virtual torque, and the commands of motor position and velocity algorithm design. The feasibility of the designed actuator and the effectiveness of the control algorithm are demonstrated by trajectory-tracking control experiments under various operating conditions.

This paper presents an angular acceleration planning for joint configuration spaces of vision-based kinematically redundant robot manipulators to achieve real-time tracking tasks in dynamic environment. In accordance with the proved proposition, the joint velocity formed by the inverse kinematics mapping of the defined pose error is regarded as the system error, and utilized to further deduce the planned joint angular acceleration and with kinematic redundancy for obstacle avoidance simultaneously. Highly-complex nonlinearities in the angular acceleration are equivalent to a system perturbation represented by a synthetic form, thereby simplifying the tedious calculations of the angular acceleration theoretically and converting the planning problem into a control problem. Hyperbolic tangent-based super twisting algorithm, as the control input of manipulator system, is designed to resist the synthetic perturbation. The collision-free movement of redundant manipulators for tracking tasks is thus achieved through the integration of the control input in real time. An error-based S-function is proposed as the internal parameter of the planned joint acceleration to prevent the integral saturation. Lyapunov theoretical analysis proves the stability of the proposed super twisting algorithm. Simulation and contrast experiments in dynamic-obstacles environments indicate that the proposed joint angular acceleration planning for kinematically redundant manipulators owns feasibility, smoothness and practicability.

There are a great number of complex relations among different characters in an opera. Retrieving such relations is crucial for performers and audience to accurately understand the features and behaviour of roles. Aiming to automatically extract relations among characters in an opera, in this paper we propose an effective method that can extract character relations from opera scripts. Firstly, we construct a uniform reasoning framework for opera scripts. Based on this model, we propose a deep syntax-parsing method to detect character relations from opera scripts. After that, we propose a new deep learning approach called SL-Bi-LSTM-CRF to extract the objects involved in character relations. The proposed SL-Bi-LSTM-CRF algorithm is a sentence-level relation extraction algorithm based on the Bi-directional LSTM with a CRF layer. With this mechanism, we are able to get a detailed description for character relations. We conduct experiments on a real dataset of opera scripts. The experimental results in terms of precision, recall, and F-score suggest the effectiveness of our proposal.

Wireless localization can provide position information for limitary, aerospace, emergency scheduling, rescue and relief work, fault diagnosis and many other related location-based services system in various application fields. Location information always supports critical decision-making of these application systems. So the localization accuracy is of great importance. As availability, low computation complexity, and low cost, trilateration localization method is applied in various systems. However, as there are numerous negative factors like non-line-of-sight measurement, complex multipath fading, which leads to high level localization error. To improve localization accuracy, an improved trilateration localization method is proposed, which is called improved trilateration localization method with distance estimation quality evaluation and anchor node adaptive selection (IT-QEAS). In this method, the qualities of distances estimation between unknown node and anchor nodes are evaluated, and the anchor nodes with higher quality distance estimation are selected adaptively to form trilateration localization equations, and finally, location result with high accuracy could be gained. The experimental result with actual localization data is shown that the proposed method can get higher localization accuracy with low computation complexity.

This paper deals with the Stewart parallel mechanism for vibration control. The problem arises from a large radio telescope in which the astronomical equipment is mounted on a platform to be stabilized by adjusting six scalable actuators, while the base platform of the mechanism itself is carried by a vibrating cable-car that moves along flexible cables. The characteristic of the Stewart parallel mechanism in this application is that the two platforms are coupled. The Stewart platform has been modeled as a rigid multi-body system. A model with a scale of 5:1 has been implemented and tested. The control effects are evaluated by comparing the root-mean-square responses of the stabilized platform. The simulation and experiment results are given and in good qualitative agreement with each other. The system can reduce the vibration amplitude evidently. The feasibility of the full-scale model scheme has been proved.

Linear Induction Motor (LIM) is widely used in rail transit, precision positioning, and industrial automation due to its high efficiency, simple structure, and low maintenance cost. However, the dynamic performance of linear induction motors is affected by nonlinear characteristics and parameter variations, making accurate state and speed estimation crucial for achieving higher performance control. Traditional estimation methods often have limitations and are difficult to meet the increasing demand for control accuracy. In this paper, an adaptive higher-order sliding mode observer (AHOSMO) based synchronous state and speed estimation method for linear induction motors is proposed to address the shortcomings of traditional estimation methods in terms of accuracy and robustness. Considering the boundary effect of linear induction motors, a higher-precision mathematical model is established. Based on the above model, the proposed observer combines the adaptive law and the higher-order sliding mode method to estimate the rotor fluxes and LIM speed, as well as the speed-dependent parameters $\gamma,\alpha,\beta,\varsigma,\eta,\delta$. The application of a super-twisting algorithm (STA) guarantees finite-time convergence for LIM state variable estimation. By employing the higher-order sliding mode algorithm, continuous inputs can be generated without the chattering phenomenon. Numerical simulation and experimental results show that the proposed AHOSMO can quickly and accurately estimate the synchronous state and velocity of the LIM under different operating conditions. Compared with existing methods, the AHOSMO demonstrates significant advantages in estimation accuracy and anti-interference capability. The research results provide a new technical approach for higher-performance control of linear induction motors, which is of great significance for promoting their application in the field of high-precision positioning.