Intelligent Robotics and Applications

The virtual machine tool and Computer-Aided Manufacturing (CAM) simulation are widely adopted nowadays to lower the cost and save time. Although parallel robotic machines are becoming popular in industry due to its unique advantages in manufacturing application, few methods are available for its simulation. This paper presents a work achieved by combining conventional CAM analysis tool HSMWorks, Computer-Aided Design (CAD) software SolidWorks, and programming tools such as Python and MATLAB to realize the machining movement simulation of a parallel robotic machine. Firstly, an original NC code interpreter is compiled in Python that interprets G-code generated by HSMWorks. Then, necessary coordinate transformation and kinematic calculation are done by using MATLAB. Finally, driving data are imported into virtual machine tool in SolidWorks, and a complete motion simulation environment is then developed. The proposed method is a general approach, which can be upgraded and modified for the simulation of parallel robotic machines with any structure.

In the robotic belt grinding process, the parameters such as robot feedrate and contact force play important roles on the material removal. In order to minimize the machining time and achieve high accuracy, a scheduling method of trajectory and force is proposed to maximize the robot feedrate. First, an optimization model with the constraints of joints speed, acceleration and robot feedrate, acceleration of robotic belt grinding is presented. Then, the feedrate scheduling problem is transferred to a linear programming problem, which can be solved efficiently. The contact force is also scheduled based on an empirical formula of grinding mechanism after solving the optimal feedrate. Finally, the simulation and experiment show the method can effectively achieve trajectory and force generation for robotic belt grinding with high efficiency and accuracy.

Biaxial tables are widely applied in high performance motion control applications for high accuracy. Tracking error is one of the most significant indicators of machining precision, and tracking control is an effective means to eliminate the tracking error. In this paper, to attenuate the tracking error and reduce the chattering phenomenon in the control input simultaneously, a fractional-order integral sliding mode controller is proposed. Compared with the existing sliding mode controller, the proposed control law not only maintains the original robustness against variations but also reduces the tracking error effectively. At the same time, the overshoot can be weakened and the reaching law will converge to the sliding surface more rapidly. Experiments conducted on a biaxial table demonstrate that the proposed control scheme is easy to apply, the tracking error is smaller and the input chatter can be improved significantly compared to the integer SMC.

3D object detection and pose estimation based on 3D sensor have been widely studied for its applications in robotics. In this paper, we propose a new clustering strategy in Point Pair Feature (PPF) based 3D object detection and pose estimation framework to further improve the pose hypothesis result. Our main contribution is using Density Based Spatial Clustering of Applications with Noise (DBSCAN) and Principle Component Analysis (PCA) in PPF method. It was recently shown that point pair feature combined with a voting framework was able to obtain a fast and robust pose estimation result in heavily cluttered scenes with occlusions. However, this method may fail in the mismatching region caused by false features or features with insufficient information. Our experimental results show that the proposed method can detect mismatching region and false pose hypotheses in PPF method, which improves the performance in robot bin picking application.

In this paper, a novel five-axis hybrid manipulator is constructed based on the 2-RPU&UPR parallel mechanism, which has less joints and all the axes of rotation are continuous. Then, the structure of the key components of the new five-axis hybrid manipulator is designed. Next, the analytic expression of the inverse position of the hybrid manipulator is established, based on which, the workpiece of the spherical surface is machined. All rotational degrees of freedom of this five-axis hybrid manipulator are of continuous axes, the analytical expression of the inverse position model can be obtained, so the trajectory planning and motion control can be realized easily. Moreover, the hybrid manipulator has less joints, which can guarantee high rigidity and high precision operation, so it has great application prospects.

Collision avoidance is the major concern for the multi-robots operation. However, few literatures can generate a collision free path as well as a smooth motion profile at the same time. To solve this problem, this paper presents an integrated motion planning scheme for two manipulators working in a shared workspace. In this method, first, a collision free path is calculated in the path planning phase. Then, the smooth trajectory is generated by using a dynamic nonlinear filter. Both the path and the trajectory are calculated directly, thus the computation load is low and the approach can be applied in a real-time manner. Simulation results indicate that the proposed method is effective to realize collision avoidance for industrial robots working in a common workspace.

This paper presents methods that can be widely and easily applied to data transformation process of machining robots. A reverse post processor is first introduced to regenerate original CLS (Cutter Location Sourse) data from post-processed NC (Numerical Control) data including variable axes codes. Then, a promising forward post processor is proposed to produce FANUC robot programs called LS data from CLS data. The proposed reverse and forward post processors allow an industrial machining robot to work based on the NC data that have been used for, e.g., a five axis NC machine tool with a tilting head.

In order to solve the problem of poor consistency and low processing efficiency of artificial grinding, this paper establishes a six-axis robot automatic grinding platform. In the first section of the paper, the qualitative relationship between the process parameters and the grinding quality could be learned from the single factor experiment, and the primitive range of the process parameter domain is obtained. On the basis, an orthogonal experiment is carried out, and a quantitative regression empirical model is established. Then the parameter sensitivity function is deduced based on the model. In regards to the grinding quality and stability constraints, the process parameter domain optimization is carried out. Finally, considering the influence of grinding attitude, the optimal attitude angle is found in the range of process parameters. The results show that the blade grinding system and the grinding scheme are effective.

The flexible robotic manipulator, which has the advantages of lightweight, flexible operation and low energy consumption, is a typical electromechanical coupling system containing the driving system, transmission system and flexible manipulator. Considering the driving motor, transmission system and flexible manipulator as an integrated object, the electromechanical coupling dynamic model of the flexible robotic manipulator system (FRMS) was constructed based on the overall coupling relationship and electromechanical dynamics analysis approach. To reveal the electromechanical coupling mechanism of the FRMS, the speed response characteristics under electromechanical coupling effects are presented. The results indicate that the electromechanical coupling factors have significant impacts on the dynamic property of the FRMS, which are meaningful for the design and control of flexible robotic manipulators.

Laser range finder (LRF) or laser distance sensor (LDS), further referred to as LIDAR (light detection and ranging). LIDAR can obtain environmental point cloud data, while a robot can realize environmental sensing by adoption of the point cloud data generated and LIDAR-based SLAM (Simultaneous Localization And Mapping) algorithm. The precision of point clouds provided by the LIDAR determines that of environmental sensing of the LIDAR-based mobile robot. In this paper, a common correction algorithm has been proposed to correct the inaccuracy of measured point cloud data caused by mobile LIDAR, effectively improving the precision of point cloud data measured by the LIDAR under a mobile state. It also conducts mathematical derivation of the algorithm, presents simulation and real world experiments performed and verifies the necessity and effectiveness of the algorithm derived by experimental results in the paper.

Research on non-twist criterion of rope is the theoretical basis of the application of rope in engineering field. Aiming at the twist problem of ropes between two relative rotation platforms, this paper introduces a kind of avoiding twist mechanism and calculates its transmission ratio of the epicyclic gear train. Based on the twist model and mathematical description of rope, a non-twist criterion is analyzed and proposed. Then, this paper reveals that how the revolving speed ratio acts on rope twist by using the method of anti-rotation and non-twist criterion. Furthermore, the avoiding twist mechanism is extended to the case of three rotation platforms and its kinematic relation for avoiding twist is analyzed. Finally, the application of avoiding twist mechanism on optical imaging of rotation platforms is analyzed.

This paper focuses on the motion planning method for a novel mobile welding robot (MWR), based on the screw theory. The robot consists of a vehicle unit and a 5-DOF manipulator, which equipped a torch at the end of manipulator. In order to finish the welding task, the kinematic motion planning strategy is of great importance. As the traditional strategy which uses inverse kinematic and polynomial interpolation may cause a waste of computing time, the screw theory is chosen to improve the strategy. From the simulation and experiment results, it can be found that the optimal motion planning method is reliable and efficient.

This paper presents the compliance modeling and error compensation for an industrial robot in the application of ship hull welding. The Cartesian stiffness matrix is derived using the virtual-spring approach, which takes the actuation and structural stiffness, arm gravity and external loads into account. Based on the developed stiffness model, a method to compensate the compliance error is introduced, being illustrated with an industrial robot along a welding trajectory. The results show that this compensation method can effectively improve the robot’s operational accuracy, allowing the actual trajectory of the robot with auxiliary loads to coincide with the target one approximately.

Joint clearance is a main cause of generating positioning error in industry robot and the low repeatability and uncertainty make it difficult to compensate. This paper mainly focus on SCARA robot and study the effect of revolute joint clearance in positioning accuracy. First, the positioning error model in XOY plane related to joint clearance is set up and analysis to obtain the extreme error value is made. Then a combined idea to study the joint clearance is presented. From this way, the seemingly random clearance error vector can be predicted by some rough deciding factors. Therefore, if the distribution pattern of clearance vector and the rough decided value are obtained, a compensation is possible. Finally, the Ball-Bar was applied to carry out experiment for invalidation and further correction. From the experiment, some patterns are obtained and show that the clearance is indeed rough decided by some factors.

Large aerospace thin-walled workpieces easily give rise to random deformation in clamping and machining processes, in which the real-time monitoring of wall thickness needs a high-quality normal technology. A high-precision on-line surface normal measurement and a real-time compensation strategy are developed in this paper. Firstly, the deviation between the actual normal vector and the spindle direction of curved workpiece surface is calculated from the data measured by four eddy current displacement sensors which are installed at the front end of the spindle; then, the deviation is converted into the compensation of each axis via homogeneous coordinate transformation and post-processing of tools. Meanwhile online compensation results get finished on the move. A simulated and experimental platform is established on an A-C five-axis machine tool in order to measure deviations of sensors at each point position and record the result of compensation. The application of this method in practical engineering can greatly improve the efficiency of measurement and control.

The Acceleration/Deceleration (AD) feedrate scheduling is widely used to plan the feedrate for CNC machining. Since the toolpath in CNC machining consists of enormous blocks (e.g., G01 or G02 block), the AD scheduling will result in lots of successive feedrate profiles. The feedrate at the junction of the adjacent profiles can be discontinuous, which will saturate the actuator and deteriorate the machining performance. The Bi-Directional Scanning Method (BDSM) is used to make the profiles overall continuous. To alleviate the computational burden, the BDSM is usually applied within a look-ahead buffer. When the buffer is filled with feedrate profiles, the BDSM updates the feedrates in the buffer. Conventional works believe that the BDSM with look-ahead (BDSMLA) will only increase the feedrate in each updating and is always feasible. We find that the BDSM can, however, decrease the feedrates in the buffer. The feedrate decrease will result in overall feedrate discontinuity and make the BDSM infeasible. We also propose a tweak method which can guarantee the feasibility of the BDSMLA. Simulation reveals the feedrate decrease in the BDSMLA and verifies the effectiveness of the tweak method.

The post-processing process for an industrial robot in milling applications suffers from a redundancy problem when converting a 5-axis tool path to the corresponding 6-axis robot trajectory. This paper proposes a feed-direction stiffness based index to optimize the redundant freedom of the robot after identifying its stiffness model. At each cutter location point, the stiffness of the robot machining system along the feed direction is maximized, and an optimal robot configuration is obtained. The optimized robot trajectory via the proposed method has an advantage of improving the machining stability and production efficiency. Experiments verify the validity of the method.

This paper deals with the kinematic analysis and performance evaluation of a 2-URPR-UPR (U, R, P standing for universal, revolute and prismatic joint, respectively) redundantly actuated hybrid manipulator. First, the kinematic analysis of the proposed manipulator is presented, including mobility analysis, inverse kinematics, and singular analysis. Then, the reciprocal of the condition number based on a dimensionally homogeneous Jacobian matrix is used to evaluate the dexterity by changing the operating height and radius ratios of the mechanism separately. Finally, the global conditioning index is carried out to describe global dexterity performance with different radius ratios.

This paper presents a solution for the topology optimization of the active arms for high-speed parallel robots. The guide-weight method is introduced into the topology optimization of continuum structures as a numerically iterative criterion. A new hypothetical material interpolation scheme is established as the theoretical foundation of the proposed variable height method. Based on the guide-weight and variable height methods, an efficient and intuitive topology optimization algorithm for flexible manufacturing which includes subtractive and additive manufacturing is put forward. The procedure of topology optimization algorithm for flexible manufacturing is described in detail. Two typical numerical examples of minimum compliance under the weight constraint are tested. In order to improve the static stiffness and dynamic response performance of the high-speed parallel robots, the presented approach is finally applied to optimize the topology of a parallel robot’s active arms.

A variable stiffness joint using a leaf spring is designed to ensure physical safety. The joint stiffness is often controlled by changing the effective length of the leaf spring. The stiffness model based on the small-deformation theory cannot solve large deflection problems of the leaf spring caused by larger joint deflected angles. The elliptic integral solution is considered to be the most accurate method for analyzing large deflections of beams. In this paper, a stiffness analysis model based on the elliptic integral solution is proposed. The joint stiffness property is analyzed. The simulation results show the joint stiffness is nonmonotonic and strong nonlinear. A stiffness simplified model is presented by nonlinear curve fitting for application simplicity. The experiment is carried out to verify the stiffness property and the stiffness analysis model.

In order to achieve a safe human-robot interaction environment, a variable stiffness joint (VSJ) intended to apply in elbow of robot arm is introduced. The configuration of the VSJ is converted from antagonistic to series by employing a mirror pair of nonlinear elastic transmissions (NETs) which are based on compliant four-bar mechanism (CFM). In this paper, the preliminary mechanical design and the working principle details of the VSJ are presented together with systematic design process of the NET. Simulation results validate the feasibility of the design of the VSJ and the NET.

Parasitic motion is widely considered as an important generic property of mechanisms, which is always believed as a drawback of lower mobility parallel mechanisms. Many researchers have tried to remove or reduce the parasitic motion by optimization or designing new structural layouts. However, in some applications, the parasitic motion can enable lower degree-of-freedom (DoF) mechanisms to complete a higher DoF motion task. In this paper, a new method is proposed to utilize the parasitic motion. By the example of a five-bar mechanism and a two-link mechanism, the optimization method is used to maximize the parasitic motion so that the mechanisms can manipulate objects with a large orientation range. This optimization based mechanism design concept will lead to lower cost, lower complexity of kinematics and control.

During extraterrestrial planet exploration programs, autonomous robots are deployed using separate landers. In this paper, a concept of a novel legged robot is introduced which has inbuilt the features of lander and rover, including landing and walking capabilities as well as being deployable, orientation adjusted and terrain adaptable. Firstly, motion characteristics of the novel legged robot mapping its functions are extracted, which can be divided into global and local motion characteristics. Secondly, structures of legs are designed according to the extracted motion characteristics, mainly composed of upper and lower parts. Finally, numerous structures of legged mobile landers are obtained and presented by assembling the same or different structures of legs.

Weight-bearing exoskeleton can effectively help the wearer to bear heavier burden, while assisting his ambulation. However, current researches in this field are relatively scarce on the passive weight-bearing exoskeleton. This research aims to design an unpowered knee-assisting exoskeleton used for weight-bearing, which can store human metabolic energy and assist human locomotion, through utilization of Teflon string, pulley, and compression spring. Biomechanical analysis of human weight-bearing locomotion shows that knee-flexion angle can be used to identify level walking or ascending movement of a person, where the largest sagittal flexion angle in level walking does not exceed 60º. Hence, the contour of pulley used for winding string is designed to be eccentric, in which the assisting torque varies nonlinearly according to the knee-flexion angle. Through mechanical modeling of eccentric pulley, we predict the assisting torque of the device and compare it with actual experimental statistics. Results show that such passive exoskeleton exhibits multi-stage nonlinear assisting augmentation under different knee-flexion angles; with least possible knee assistance during level walking, and remarkable assistance during climbing.

This paper gives an overview on HIT humanoid robot, which is developed as a research platform for replicating human in special environments. The system has two upper extremities, each upper extremity includes a 7-DOF humanoid arm and a 15-DOF dexterous hand as end-effector. Two humanoid arms and dexterous hand are combined with a 2-DOF dynamic torso and a 3-DOF binocular head to form the humanoid upper body. A wheeled mobile robot is employed to provide mobility. In this paper, we describe the design specification and give an overview on mechanical design of each subsystem. Additionally, we give a short introduction to the hardware architecture and software of HIT humanoid robot.

This paper presents the development of a modular robotic laparoscopic tool for MIS (Minimally Invasive Surgery). A dual continuum mechanism is utilized in the tool design to ensure reliability as well as achieve enhanced distal dexterity, increased payload capability and actuation modularity under a simple construction. Via kinematics modeling, the laparoscopic tool could be maneuvered by a Denso manipulator to perform typical laparoscopic tasks and possesses the desired functionalities for MIS. Advantages of the implemented dual continuum mechanism lead to the performances of this attempt. Motivated by the commercial success of the da Vinci surgical system, this paper presents an alternative design to realize robotic laparoscopic surgeries, which could lead to possible future commercialization opportunities.

Bladder cancer, with the leading number of new cases in all urinary system cancers and a high recurrence rate, poses a substantial threat to human health. Even with the transurethral accessibility, current surgical tools have not fully allowed convenient resection of the bladder tumors. This paper presents the design and the preliminary development of a continuum dual-arm surgical robotic system for transurethral procedures. This development aims at improving the current surgical treatments by providing intravesicular imaging with enhanced distal dexterity. With the proposed system, new surgical techniques for bladder tumor resection could be explored. The clinical motivation, design overview, system descriptions and preliminary developments of this transurethral surgical robot are presented. With the system constructed in the near future, a series of ex-vivo and in-vivo experimentations would be carried out to verify the proposed functionalities.

The technology of taking off from the lunar is of great importance for the returning of spacecraft from lunar. To verify the stability of spacecraft when taking off lunar, the force state of the spacecraft needs to be simulated accurately. Traditional simulation mechanisms are unable to meet the simulation requirements of lunar takeoff. A 6-DOF lunar takeoff simulation device with nine cables is introduced and optimized in this paper. Firstly, the dynamic workspace of cable-driven parallel mechanism, which under the condition that the quality, the vertical acceleration and the horizontal acceleration are 15 kg, 2.4 m/s2 and ±1.1 m/s2 respectively, is analyzed when the acceleration requirements of simulation mechanisms is taken into consideration. Then the installation positions of cables are investigated in detail based on experimental design method and response surface method to achieve optimization results that could get the maximum dynamic workspace. Moreover, the virtual prototype experiment based on the multi-body dynamics model is utilized to verify the accuracy of the optimized results.

Study of methods to design optimal Gough–Stewart parallel manipulator geometries meeting orthogonality is of high interest. The paper will design the Orthogonal Generalized Gough-Stewart Parallel Manipulator (OGGSPM) based on the composite hyperboloids. Moreover, Particle Swarm Optimization (PSO) algorithm is introduced to perform the structure optimization as the Jacobian matrix condition number is used as the evaluation index. As an example, two different optimization cases are presented.

Precision positioning with multiple degree-of-freedoms (DOFs) is the core technology of nanometer manufacturing equipment. In this paper, a 3-DOFs monolithic parallel compliant manipulator is designed for relieving the conflicts among large workspace, high precision positioning and multi-DOFs. The 3-DOFs compliant micro-positioning manipulator has parallel structure and is composed of a static platform, a moving platform and three kinematic chains. Three kinematic chains are arranged symmetrically along the center of the moving platform with 120°, and each chain is actuated by a piezoelectric ceramic actuator through a variable cross-section symmetrical four-bar mechanism with low coupling and better sensitivity. Flexure hinges are utilized as the revolute joints to provide smooth and high accurate motion with nanometer level resolution. Based on the “pseudo-rigid-body model” method, the inverse kinematics model and dynamic model of the compliant manipulator are established. Finally, the natural frequency of the 3-DOFs compliant micro-positioning manipulator is obtained by the ANSYS software.

A Gough-Stewart parallel manipulator using point decoupled design method is orthogonal only at a single point, which means that it has small high-precision workspace. It is hard to break through its’ restriction to be applied in the field of engineer applications with larger workspace. This paper formulates the problem to guarantee continuous orthogonality when the manipulator rotating along the z-axis, called rotation orthogonality. Compared to traditional optimum, the rotation orthogonality leads to better performances with enlarged workspace. A class of dynamically adjusting generalized Gough-Stewart parallel manipulators (DAGGSPM) is proposed. The dynamically adjusting mechanism for rotation orthogonality and the related algorithm are deduced analytically. Through analysis of three typical cases and numerical verifications, the results show that a DAGGSPM can pave the way for high-precision applications with large scale workspace including laser weapon pointing, scanning microscopes and integrated circuit fabrication.

According to the topological structure design method for parallel mechanism (PM) based on the position and orientation characteristics (POC) equation, a so-called Delta-CU, 2-R//R//P(4R)//R⊕1-RSS PM is first presented. It is proved that the PM can achieve three-translation output. Second, both the forward and inverse solutions for the position analysis are derived. Besides, the workspace analysis is then performed based on the inverse position equations. Furthermore, the matching relationship among $$ (R - r), $$(R-r),$$ l_{1} $$l1 and $$ l_{2} $$l2 is obtained to ensure its maximum workspace. The PM has the advantages of simple structure, easy assembly and maintenance, etc., and can replace the Delta PM. This work lays a foundation for the dimension optimization, dynamic analysis and the prototype design of the Delta-CU PM.

Basing on the structure of the pneumatic parallel robot with two degrees of freedom (DOF), this paper studies the control method of synchronous coordinated motion of two cylinders in the parallel mechanism to ensure the horizontal attitude of the end position of the parallel mechanism and the trajectory planning. Two active cylinders in the parallel mechanism cannot be controlled coordinately because of the compressibility of gas, the variety of the load force of parallel mechanism and the difference of cylinder pressure changing in the motion. The paper uses the fuzzy adaptive PID control algorithm with gravity compensation and the object-oriented coordinated method to control the motion of the active cylinder. Firstly, the pneumatic parallel mechanism in this paper is introduced. Then the paper studies on the control method when cylinders swing, which is the fuzzy adaptive PID control algorithm. In the end, two cylinder coordinated motion control method is researched based on the parallel mechanism. The synchronous motion control method lays a theoretical foundation for the study of pneumatic parallel mechanism and improves the pneumatic parallel robot technology.

Based on the finite element method and Lagrange equation, high dimensions, time-invariant, nonlinear rigid-elastic coupling dynamic equations are established, in which the coupling influence of elastic deformation motion and the rigid body motion are considered. A comparative analysis of one-step method and two-step method for solving nonlinear rigid-elastic coupling dynamic equations is given. The simulation results show that the elastic vibrations of the flexible links have important effect on rigid body motion, particularly for angular velocity and angular acceleration. There are some errors in the amplitudes and phases of the elastic displacements and the elastic rotation angle for the two methods, rigid-elastic coupling of flexible links will critical influence elastic deformation displacements and elastic rotate angle of the moving platform, but change rule of the kinematic variables are basically consistent. So, one-pass method is used to solve dynamic equation of rigid-elastic coupling system that can reflect the dynamic characteristics well, it can provide guidance for controller design in future.

Parallel robots are manipulators who use closed kinematic loops to create a movement at the Tool Center Point (TCP). Due to this kind of kinematic loops, it causes the mechanism to have high stiffness, high ratio of load to self-weight and low inertia, even though it will have good repeatability and high dynamics. A very well-known example of a parallel manipulator is the Delta Robot which is commonly used in the pick and place industry. Delta Robots are known to normally use a pivoting working principle for the actuating upper arms, meaning, its design does not make it possible for the upper arm actuators to make a complete revolution. This paper introduces a new type of an advanced Delta Robot with extended RSUR kinematic which allows a complete revolution of the actuators what will lead to the advantages of storing rotational energy in cycles comparatively like a flywheel what is a machine element that stores rotational energy. The extended kinematic torque curve of actuators therefore runs more smoothly and thus the payload by same cycle time increases. The following work will introduce the kinematic model description of such an advanced Delta Robot that is working with a circulating working principle, and discuss several simulation results of such a kinematic to prove the efficiency of the circulating working principle. Furthermore, a comparison with a real Delta Robot hardware and the advanced Delta Robot simulation model is introduced.

Surgical robot with a remote center of motion (RCM) plays an important role in minimally invasive surgery (MIS) field. To make the structure more compact and have higher rigidity and accuracy, a new type of parallel surgical robot with three degrees of freedom (DOFs) is developed. The detailed design of the mechanism is provided in the paper. To better study the characteristics of the mechanism, static analysis is proposed in the paper. Firstly, establish the equations of the force and torque through the analysis of the branch and the mobile platform. Next, obtain the expression of the active force and constraining force of the joints. One numerical example is given. In order to analyze the dynamic performance of the UPR-2UPRR system, the Kane method is used to build the dynamic model. The speed, acceleration, partial velocity and partial angular velocity are calculated. Then the dynamic equation is given after analyzing the generalized active force and inertial force. At last, the driving force of each telescopic rod is given.

On the basis of accuracy analysis of rigid-flexible coupling model of 3-R2H2S parallel robot, the pose error analysis and synthesis is studied systematically in this paper. In terms of rigid analysis, the accuracy mathematical model of 3-R2H2S parallel robot is established by the differential theory based on the inverse solution. The independent and equal effect principle of error is applied to the accuracy synthesis of the robot. Through the computer simulation, the distribution of structure parameter errors is analyzed according to the allowable range of pose error. The simulation results show the active arm error has the most remarkable influence on the pose error of the parallel robot. As for flexible analysis, the rigid-flexible coupling model of 3-R2H2S is established by the software ADAMS and ANSYS. The simulation results of rigid-flexible coupling model show the elastic deformation of the fore arms plays a major role in the robot pose error. The methods proposed in this paper have provided theoretical basis for accuracy synthesis for parallel robot of this type.

Singularity analysis is an important problem in the field of parallel mechanism, how to obtain a concise and analytical expression of the singularity locus has been the focus of the study for a long time. This paper presents a new method for the singularity analysis of the Stewart platform parallel mechanism. Firstly, the rotation matrix is described by quaternion; utilizing length constraint equations of the extensible limbs and properties of the quaternion, seven equivalent equations are obtained, and the variables of position and orientation are decoupled preliminarily. Secondly, by taking the derivative of seven the equivalent equations with respect to time, a new kind of Jacobian matrix is obtained, which reflects the mapping relationship between the change rate of the lengths of extensible limbs and the change rate of variables of position and orientation of the moving platform. Finally, the analytical expression of the singularity locus is derived from calculating the determinant of the new Jacobian matrix; when the quaternion are transformed into Rodriguez parameters, there are only 258 items in the fully expanded analytical expression of the singularity locus. Not only can this method be used to study the singularity problem of Stewart platform parallel mechanism, but it can also be used to study the singularity free workspace of the mechanism.

All joints of the serial mechanism are active ones, while only some joints are active ones for the parallel mechanism. Thus, there is a variety of drive configurations with different active joint selections. Drive configuration will affect the performance of parallel mechanism and the effect deserves further study. A planar 5R parallel manipulator with the capability of adjusting drive configurations in real time is proposed in this paper. Based on the deduced Jacobian matrices, performances of the 5R parallel manipulator are analyzed, considering three symmetrical drive configurations. Typical kinematic performance indices, such as local conditioning index and kinematic manipulability, are adopted to give a complete illustration of the performance variation. Then, the optimal adjustment of driven configurations is proposed and studied, considering local conditioning index. Simulation results reveal that the performance of the 5R parallel manipulator improves obviously through the reasonable combination of different drive configurations in the workspace. The drive configuration adjustment proposed in this paper provides a novel potential way to enhance the comprehensive performance of the parallel mechanism.

This paper presents a macro-micro two-stage parallel experimental platform. The 3RRR macro planar parallel mechanism driven by the YASWAKA Σ-V series motor is to ensure the system’s large workspace. The 3RRR micro parallel mechanism based on compliant mechanism is driven by the pack-aged piezoelectric ceramics actuators (PPCA) which effectively guarantee the precision and accuracy of nanoscale positioning. In fact, the load characteristic is a basic index that affects the performance of the macro/micro combination mechanism. A displacement experiment is designed for the macro parallel mechanism to move in a straight line under different loads. The displacement is measured by the Renishaw laser interferometer. It is proved by the experiment that the precision and stability can be affected under different loads. The load characteristics research results and analysis in this paper are significant to the optimal design of a macro-micro dual-drive precision positioning planar parallel mechanisms for different applications.

Though nature images generally consist of edges and homogeneous regions, most image deconvolution approaches will result in unnaturally sharp image estimations, since they tend to overemphasize sharpness while ignoring smoothness. To balance these two significant image properties, this paper presents a novel analysis-based image deconvolution approach based mainly on alternating minimization and variable splitting. The presented approach focuses on a new method of solving a constrained minimization problem which is derived from a universal model and is used to model the image deconvolution. By alternating minimization with variable splitting, the proposed image deconvolution model is firstly converted to an equivalent model, and then decoupled into a number of sub-problems. These sub-problems are alternately handled using a corresponding iterative method to obtain the solution for the proposed image deconvolution model. The presented approach has been demonstrated to be effective and superior to several state-of-the-art approaches for image deconvolution applications.

This paper presents an object reconstruction method based on a binocular stereo vision system. First, the relative position and orientation between the two cameras of the system are obtained by a binocular calibration process. Then feature pixels are extracted from images of the two cameras of the same scene according to the SIFT (Scale Invariant Feature Transform) feature. Feature pixels on the image of one camera are matched with those of the other camera according to differences between their SIFT features. And the RANSAC (Random Sample Consensus) algorithm is used to eliminate incorrect matched pixels. Then a 3D coordinate point is obtained from each pair of matched pixels. Finally, 3D models are constructed from the 3D coordinate points through triangulation and texture mapping. In the above processes, a uniform method of calculating coordinates of 3D points from pixel pairs is introduced, which is fitted for arbitrarily orientated optical axes of the left and the right cameras. Experiment results show that the proposed method can obtain 3D points sampled from real objects and produce 3D models consistent with reality.

In traditional close range photogrammetry, more than eight coded targets are adopted to determine the relative orientation. Generally, dozens or hundreds are needed for higher precision. It is a tedious task to place them in the scene. In this paper, a new method with less than eight coded targets is proposed. The key idea is the coded pattern not only provides identification information, but also contains substantial location information. A new detection method of coded targets is developed to recognize the coded targets and excavate the location information in the coded pattern. All this location information is adopted to determine the relative orientation. The experiment results show that two coded targets are enough to determine the relative orientation, and three are enough for high precision.

Aiming at the problem that the traditional space control point cannot be laid in the large field of view when calibrate the camera’s parameters, a high precision calibration method of the camera is proposed. In this paper, the rover part of the differential GPS system is loaded on the unmanned aerial vehicle (UAV) and the base part is placed at the origin of the world coordinate system. The UAV is controlled to move within the field of view of the camera which to be calibrated and we can obtain the three-dimensional coordinates of the control point and the corresponding image coordinate. And then according to the “High Accuracy Calibration Method Based on Collinear Equation” to calculate the initial value of the camera’s internal and external parameters, and finally we can solve the accurate calibration parameters after optimization. The measurement accuracy of differential GPS system can reach centimeter level, the control points should distributed evenly in the field of view to solve the exact results. Experiments show that the accuracy of the intersection of 300 m and 800 m can reach 0.6 m and 1 m respectively, so the method have the characteristics of high precision, practicality and simple operation.

In the machine vision, the algorithm of background subtraction is used in target detection widely. In this paper, we reproduce some algorithms such as ViBe, Local Binary Similar Pattern (LBSP), Local Ternary Pattern (LTP) and so on. In view of the problem of inaccurate edge in target detection, we propose a method of combining color and LBSP for background model. It can obtain information both in pixel and texture (marked as improved-LBSP in the paper). On this basis, we propose a new method marked as BFs-method in the paper, which have a new persistence consists of color, LBSP, and time (t). The key advantage of this method lie in its highly robust dictionary model as well as it’s ability to automatically adjust pixel-level segmentation behavior, which improves the ability to remove the shadow of the target and the hole inside the target.

Chessboard detection and corner extraction are imperative during the camera calibration process, which is a fundamental work in computer vision. This paper describes a new chessboard corner detection algorithm using the amplitude spectrum feature of circular sampling at each point as corner response function. A simple thresholding technique is that only the points near a chessboard corner will get positive corner response. However, image noise will bring a lot of false positives. The distribution of false positive response with respect to image noise is given, this distribution also has a property like the three-sigma rule in Gaussian distribution, that can be used as thresholding rule. The only parameter needed is the standard deviation of image noise, which can be measured using some patches of the image for once. Experiment is presented showing the efficiency of the proposed method against noise, compared with existing algorithm under simulated image.

Connected components analysis plays an important role in real-time target localization. In this paper, we propose a fast connected-components-analysis algorithm without equivalence-tables based on FPGA (Field Programmable Gate Array). The algorithm runs in a pixel-scanning mode and labels all run-length pixels. As a result, the geometric features of the connected components can be extracted by only one-times image-scanning. A mapping structure between a run-length and a connected components is designed to replace the equivalence table in order to reduce its resource consumption and promotes its running speed. The experiment results demonstrate that our algorithm occupies less resources than state-of-the-art algorithms and runs faster than personal-computer based algorithm. Besides, this FPGA-based algorithm can deal with any connected components in complicated shape or quantity.

To find an area where the robot can obtain a higher kinematic performance is a meaningful work for path planning and off-line programming. In this paper, a methodology was proposed to determine the high performance area (HPA) of a 6R industrial robot. Monte Carlo method was used to get a point cloud of the reference point of the end effector. The manipulability measure was selected as a performance index to filter the point cloud. A grid was defined to approach the filtered point cloud and its boundary was then extracted and smoothed. An example was presented to show that the proposed methodology is feasible to determine the HPA of a 6R industrial robot.

This paper proposes a recognition of initial welding position for fillet weld. A structured-light vision system is presented. Using a laser scanning method, arc welding robot reciprocates along the end of weld seam with an incremental motion strategy, which makes the recognition of initial welding position fast and accurate. After given the endpoint types of fillet weld, an image processing is described. Firstly, the laser center stripe and the feature point extraction methods are given in detail, and then connected components in the image are extracted by using image segmentation algorithms. Finally, experiments are conducted with calibrated vision system to prove the effectiveness of the proposed method.

A dual-camera assisted method of the SCARA robot for online assembly of cellphone batteries is proposed, to solve the problem of low success rate of SCARA robot used for assembling cellphone battery. In this method, both cellphone battery and cellphone base can be precisely located with the help of the dual-camera, thus the success rate of assembly can be significantly improved. Two steps are mainly included in this method: First, calibrating hand-eye relationships between the SCARA robot and each of the two cameras; second, extracting features of the assembled targets and controlling the robot to rectify the error. The experimental results show that the proposed method can well meet the requirements of cellphone battery assembly and improve the success rate.

In this paper we propose a framework for instance recognition and object localization in cluttered and occluded household environment for robot grasping task. The whole system bases on a coarse to fine pipeline in combination with the state-of-the-art methods of RGBD-based object detection. We build a sparse feature model by extracting structure key points incorporating texture cues in the train procedure. After that, the paper demonstrates how the algorithm decreases the time complexity and simultaneously guarantees the accuracy of the recognition and pose estimation. Quantitative experimental evaluations are presented using both acknowledged ground truth dataset and real-world robot perception system.

This paper compares implementation of multiple view approach of pose estimation on an eye-in-hand robotic system. By combining RGB-D frames from multiple view with the eye-in-hand robotic system, geometry information of target objects can be best recovered thus pose estimation performance can get improved. Two primary approaches for pose estimation, namely 3D point cloud registration and 2D image matching are implemented and compared. For the 3D method, we reconstruct target objects by taking advantage of the eye-in-hand system to get an accurate representation of target objects. For the 2D method, we discuss distance metrics and regression for 6DOF pose and apply RANSAC with it to fuse multiple estimation results. State-of-the-art pose estimation algorithms which cover both the 3D and 2D approaches are implemented and compared. Experiments show that the multiple view approach can provide more accurate and reliable pose estimation results when compared with conventional single view approach.

Most of the plug-in machine adopts the visual positioning system that is used by Surface Mount Technology (SMT), and obtains the bottom image of component. For some electronic components, as the color of pin is similar to the body, the pin is difficult to be distinguished. In order to improve the positioning accuracy of the rotational stereo vision positioning method for plug-in machine, an electronic component pin edge tracking algorithm based on iterative optimization and a sub-pixel edge extracted method based on the cubic curve fitting are proposed. Furthermore, a center point calculation method is proposed for component pin with different shape to improve the accuracy of feature point extraction. Experimental results show that the proposed center point calculation method performs higher accuracy than the method of taking average of borders as center point directly. In addition, the proposed subpixel border extracting method performs better accuracy than the Canny algorithm.

The Iterative Closest Point (ICP) algorithm has been viewed as a standard approach to registering two point clouds. In the process of point clouds registration, the eliminating incorrect point pairs has important effect on the accuracy and stability of registration. In the past two decades, numerous strategies of excluding point pairs have been developed and various combinations of them have been applied to the variants of ICP algorithm. In this paper, an efficient combination of rejection strategies is proposed. It also is compared with other heuristic combinations. As shown in our case studies, the proposed combination can realize more accurate registration without sacrificing computational efficiency.

LiDAR odometry is the key component of LiDAR-based simultaneous localization and mapping (SLAM). However, the low vertical resolution of LiDAR makes it difficult to produce pleasant mapping results. It is even more challenging to reconstruct the surface of dynamic objects from the raw LiDAR input. To address this problem, existing approaches typically divide it into several subproblems like object detection and tracking and then solve them individually, which greatly increases the complexity of LiDAR odometry as well as the SLAM framework. In this work, we propose to address this problem by improving LiDAR odometry with appropriate modifications to the depth fusion process and several additional lightweight components. Extensive evaluations on KITTI dataset and Velodyne HDL-16E laser scanner demonstrate the effectiveness of the proposed method. The results of the improved LiDAR odometry include abundant information about the dynamic objects, which can be used for many high-level tasks such as object recognition and scene understanding.

Particle filters are being widely used in tracking single object for its unique advantages. According to former studies, particle filters can solve the tracking problems in the circumstances of nonlinear and non-Gaussian observation. In this paper, we propose a robot teaching method based on multi-objective particle filter for motion tracking, in which targets can be recognized by the value of H in HSV color space. After the targets are recognized, particle filtering is applied to achieve motion tracking. PnP algorithm is another essential part of this method, which can obtain translation and rotation matrix of the moving objects. All these pose information is sent to robot to reproduce the targets’ motions. The experiment in this paper is divided into three parts, firstly, a single robot arm achieve the linear motion along three-axis respectively, and then move along a particular trajectory. Finally, to achieve both arms tracking hands movements at the same time. When the single target is tracked, the experimental results are better compared with track hands simultaneously.

At present, binocular vision system is widely used in unmanned aerial vehicle (UAV). However, there is a large vibration in the process of UAV’s flight. It will lead to the change of the position relationship in the binocular vision system. To solve this problem, this paper proposes a method based on Bundle Adjustment optimization algorithm. It is based on the camera calibration data calibrated before out of factory. The rigid body transformation matrix between the two cameras is calibrated and optimized by the position of the feature points and the image information around the feature points. A series of experiments are conducted to test the algorithm. The experiment shows that the distance between calibrated 3D points and the ground truth is less than 5 mm when the length between target and binocular vision system is about to 2 m. It has fully satisfied the needs of the subsequent computation of disparity map. The algorithm has been applied to a UAV binocular vision system.

This paper proposes a low computational cost method for abnormal crowd behavior detection with surveillance applications in fixed cameras. Our proposal is based on statistical modelling of moved pixels density. For modelling we take as reference datasets available in the literature focused in crowd behavior. During anomalous events we capture data to replicate abnormal crowd behavior for computer graphics and virtual reality applications. Our algorithm performance is compared with other proposals in the literature applied in two datasets. In addition, we test the execution time to validate its usage in real-time. In the results we obtain fast execution time of the algorithm and robustness in its performance.

This paper introduces a real time non-intrusive method to determine driver fatigue by analyzing eye gaze patterns. Using a standard webcam and a personal computer, the proposed method combines different techniques in order to keep a low computational cost without a loss of performance. Facial features are identified in a reference image to extract a region of interest, around the eyes of the user, and tracked by an optical flow algorithm in subsequent frames. Color segmentation on the resulting images allow the system to extract data needed to determine ocular following, blink detection, frequency, and percentage of eye lid closure over time (PERCLOS). This approach, while simple, proves to be very efficient and accurate for the hardware restricted setup, allowing faster information processing on modest specifications systems. For safety reasons, our experiments are limited to different subjects, simulating fatigue in laboratory conditions as well as a real time test on a moving vehicle to analyze the blinking patterns.

The main goal of this article is a stabilization system for cameras onboard mini Rotor UAVs that minimize the effect of the undesired movements on captured videos through the selection of, an autopilot and a ground control station for data extraction, an inertial stabilized platform (gimbal), an onboard computer for image processing, the communication interface, the software involved, the implementation of the stabilization system algorithm and results.

The future would be full of artificial intelligence definitely. Since Olympic game 2020 would be held in Tokyo, it is overwhelming important to give a navigating service to the tourist from the entire world. Even there would be a large number of volunteers then, there would be a lack of position absent. This paper described a vision system for robot system for airport navigation that set in the airport or other places. This visual system contained the detecting part and human counting part that combined with evolutionary and clustering algorithms and the experiment shows an efficient result in some cases.

In order to improve the static gait performance of a quadruped robot, the method based on the body lateral adjustment has been proposed. In this paper, two kinds of optimization models were designed, which were based on the body intermittent lateral adjustment, and based on the body continuous lateral adjustment, designed a evaluation function with the aim of stable motion and fast walking speed; Besides, the simulations of the virtual prototype model based on those two kinds of optimization models have been finished in MATLAB and ADAMS, and the simulation results were analyzed by the de-signed function; Finally, we can find that the better optimization model to improve the static gait performance of a quadruped robot is based on the body continuous lateral adjustment. Those simulation results revealed that it greatly enhances the quadruped robot’s static gait performance by using the designed optimization model adjusting its body lateral position continuously.

In the field of service robot, implementation of accurate and effective grab is usually the key step in operating process. Common grab of traditional robot is completed by special gripper or holder. Usually these special grippers are composed of mechanisms which are relatively simple, and only used for grab of specific objects, lacking of force feedback and strong universality. Nevertheless, dexterous hand could overcome these shortcomings, which can complete operation with more extensive adaptability. This manuscript designs a kind of double slider-crank mechanism based under-actuated double fingered dexterous hand, describes double-joint finger mechanism and analyzes movements of two grasp modes. According to the grasp of objects in small size, mechanism’s output force performance is analyzed, and movement characteristics of finger mechanism in two grasp modes are described from perspective of joint angles. Degree of adaption is proposed and analyzed, and grasp experiments are designed and performed. At the end of this manuscript, experiments results are showed and analyzed. The results show that the mechanism designed by this manuscript has adaption ability to different sizes and shapes.

In order to overcome the shortcoming that traditional self-adaptive underactuated robot hands cannot perform linear parallel pinching, this paper presents a novel linear parallel and self-adaptive hand with y-shaped linkage mechanism, LIPSAY hand. The LIPSAY hand uses y-shaped linkage mechanisms to achieve linear parallel pinching, and the pulley-belt and double-slider mechanisms to perform self-adaptive grasping. Compared with traditional underactuated hands, the LIPSAY hand can realize the hybrid function of precise linear parallel pinching and self-adaptive grasping. The LIPSAY hand has strong grasping adaptability and high grasping stability, which is suitable for industrial applications.

This paper proposes a novel parallel pinching and self-adaptive grasping hand with triple-shaft pulley-belt mechanisms, named TPM hand. It contains two TPM fingers, and each finger has two phalanges, and mainly consists of a driving shaft, triple pulleys, two belts, and two springs. The TPM hand is capable of grasping objects with parallel pinching and enveloping, and simultaneously adapting to the contours of the objects. Particularly, compared with other parallel pinching and self-adaptive grasping hands, the TPM hand separates the proximal shaft and the driving shaft so that the grasping range can be expanded. Besides, due to the elasticity of flexible devices, objects can be protected in emergencies. New mechanical parameters are added and are conformed to specific domain and function. This function is given in this paper through geometry model. Meanwhile, a feasible analysis is provided to discuss the influence of parameters brought by the departure above. Also, the process of grasping is discussed, and the kinematic analysis is provided.

This paper presents the design of a novel underactuated robot finger with a rotating-idle stroke, called PASA-RIS finger. The PASA-RIS finger consists of multiple gears, a pair of rotating blocks, a pair of limiting blocks, and two springs. The finger has two joints driven by one reduced motor and a transmission mechanism. Grasping force analysis and experimental results show that the PASA-RIS finger has the parallel pinching and self-adaptive grasping function. The PASA-RIS finger can adapt objects of different shapes and sizes and automatically switch from parallel pinching to self-adaptive encompassing.

This paper presents a robot prototype designed to overcome the lack of current security monitoring system: low flexibility, limited monitoring range, disability of copping with sudden special circumstances. It’s designed and built for remote live-video security surveillance and object recognition and capture in highly structured small space environment in industrial field, involving remote controlling, live video taking, distance and position identifying of an object, trajectory planning and object capturing. The robot is controlled by means of human-machine interface. The operator drives the robotic device and performs the related tasks by means of a control user interface developed on android APP. The robot computer operates robot grasping and manipulation automatically, that is, performs the precise localization of a specific object, computes adequate capturing sequence and controls the motion (robot moving and object capturing) of all the mechanical components. Throughout this paper, the specific design of every module of the robotic device is presented. The device has been built. Related laboratory tests have been done to check the robot performance. Results show excellent robot controllability, targeted object finder and live field operations. Some figures and tables showing overall performance are given.

Dynamical identification methods for the industrial robot manipulators are widely and successfully applied to obtain a model that is suitable for controller design. In this paper, the dynamical model of robot was obtained by Newton-Euler method and linearized by a particular approach. A novel glowworm swarm optimization algorithm was introduced to estimate the unknown parameters. The algorithm had been coded in the popular Matlab environment and the procedure was tested in a practical case research to identify the dynamical model of a six degree-of-freedom industrial robot. The results of the identification experiment showed the efficiency of the proposed algorithm.

In order to improve the trajectory tracking precision and reduce the synchronization error of a 6-DOF lightweight robot, computed-torque deviation coupling control strategy based on friction compensation analysis is presented. The mathematical models of the robot which include kinematic model and dynamic model are established. The single joint Lugre friction model is proposed, and the parameters of the friction model are identified by experiment. Since it is difficult to describe the real-time contour error of the robot for complex trajectory, the adjacent coupling error is analyzed to solve the problem. Combined with friction compensation and coupling performance of the robot, computed-torque deviation coupling controller is designed and validated by simulation analysis. A servo control experimental system is constructed, and verified that the synchronization error are significantly decreased and the trajectory error is reduced from (−0.8°~1°) to (−0.230°~0.587°) after the friction compensation is added. The effectiveness of the control algorithm is validated by the experimental results, thus the control strategy can improve the robot’s trajectory tracking precision significantly.

Pneumatic muscles actuator (PMA) is widely used in the field of rehabilitation robot for its good flexibility, light weight and high power/mass ratio as compared to traditional actuator. In this paper, a fuzzy logic-based PD-type iterative learning controller (ILC) is proposed to control the PMA to track a predefined trajectory more precisely during repetitive movements. In order to optimize the parameters of the learning law, fuzzy logic control is introduced into ILC to achieve smaller errors and faster convergence. A simulation experiment was first conducted by taking the PMA model fitted by support vector machine (SVM) as controlled target, which showed that the proposed method achieved a better tracking performance than traditional PD-type ILC. A satisfactory control effect was also obtained when fuzzy PD-type ILC was applied to actual PMA control experiment. Result showed that it takes 25 iterations for the maximum error of trajectory converges to a minimum of about 0.2.

Serial Elastic Actuators (SEAs) have several superiorities over conventional rigid actuators e.g., greater shock tolerance, energy storage, safety, and so on. However, there exist performance limitations due to existence of elasticity in SEAs, i.e., low accuracy in position control, low control bandwidth in force control as coupling stability is guaranteed, and impedance constraints in impedance control. Variable Stiffness Actuators (VSAs) are designed to address the performance degradation but increases complexity in mechanism. As a result, control in SEAs plays a more important role than that in their rigid counterparts. We address this challenge by cascade control severing as the same purpose as that of VSAs’ without losing favorable advantages of SEAs to imitate humanoid manipulation. Then stability and passivity constraints is derived. Performance analysis is conducted in terms of quasi-rigid performance and apparent stiffness. Finally, experiments are carried out to test performance of the suggested control scheme.

Stable walking is the most basic human behavior of humanoid robots and one of the most important research contents in the field of robots. However, reasonable gait planning is the basis for the stable walking of humanoid robots. Therefore, in this paper, we analyzes one of the CPG model and applies it to our own laboratory robot, aim at the problem that it is prone to shock forward and backward in the Robot Athletics Sprint, this paper increased centroid offset control and proposed an algorithm to optimize the walking control parameters for the improvement of the whole gait planning algorithm. Stability margin of ZMP and balanced oscillator amplitude were combined as an optimization target, genetic algorithm was used as a solution tool, the purpose is to get the optimal gait parameters under different input speed. The proposed algorithm was tested on the laboratory robot, the results of simulation and real robot experiments show the effectiveness of our algorithm.

Active compliance control is an effective method for a legged robot to decrease impact disturbances at the moment of foot-ground contact and to realize harmonic locomotion by actively adjusting the leg stiffness and damping in real-time, which will significantly improve the adaptability of the robot to the irregular terrain. However, the design of an active compliance controller remains to be a challenge when the nonlinearity of actuators must be taken into account. This paper presents an active compliance controller for a bionic hydraulic quadruped robot. The nonlinear dynamics of the hydraulic system, including the compressibility of the fluid, the friction of the hydraulic cylinders, and the flexibility of the tubing are well modeled and identified. The nonlinearity of the hydraulic actuators are compensated through friction compensation and feedback linearization. The proposed active compliance controller has been applied to a bionic hydraulic robot prototype. Experimental results indicate that the active controller can handle the impact disturbances from robot feet.

Kinematic redundancy has been widely used in manipulator design. Nevertheless, the control of redundancy is a tough task. This paper proposed a control method for serial kinematic redundant manipulators in consideration of multiple performance criteria. The method is based on the classical gradient projection method, and in order to handle the relations between different performance criteria, a combination of task priority strategy and fuzzy inference is used. The method is applied to a 10 DOFs (degrees of freedom) manipulator as an example to illustrate its efficiency.

In this paper, we propose a boundary control strategy for vibration suppression of two flexible wings and a rigid body. As a basic approach, Hamilton’s principle is used to ascertain the system dynamic model, which includes governing equations and boundary conditions. Considering the coupled bending and torsional deformations of flexible wings, boundary control force and torque act on the rigid body to regulate unexpected deformations of flexible wings that caused by air flow. Then, we present stability analysis of the closed-loop system through Lyapunov’s direct method. Simulations are carried out by using finite difference method. The results illustrate the significant effect of the developed control strategies.

In this paper, an extended Proxy-based Sliding Mode Control (PSMC) which combines sliding mode control with conventional stiff position control, is proposed and applied for the tracking control of a class of second-order nonlinear systems. The well-known chattering problem can be solved by using the proposed control scheme. As a result, the overdamped dynamics which is extremely important for system safety can be achieved without losing the advantage of accurate control performance in the presence of external disturbances and parameter uncertainties. Based on Lyapunov theory, the stability analysis of the proposed scheme is presented and the passivity of the system is also proven. Experiments are carried out to verify the proposed method.

In this paper proposes a numerical methods based control algorithm for tracking trajectories applied in two anthropomorphic robotic arms mounted on an omnidirectional platform, which allows the transport of an object in common. For this, the kinematic modelation of the entire coupled system is performed, considering as the position of interest the midpoint of the operative ends of each manipulator and its formation characteristics. The stability of the proposed controller for tracking trajectories using numerical methods is demonstrated analytically, obtaining that the control is asymptotically stable. Finally the results obtained are evaluated in a virtual reality simulation environment.