Peijiang Yuan

Development of Sensory-Motor Fusion-Based Manipulation and Grasping Control for a Robotic Hand-Eye System

In this paper, a sensory-motor fusion-based manipulation and grasping control strategy has been developed for a robotic hand-eye system. The proposed hierarchical control architecture has three modules: 1) vision servoing; 2) surface electromyography (sEMG)-based movement recognition; and 3) hybrid force and motion optimization for manipulation and grasping. A stereo camera is used to obtain the 3-D point cloud of a target object and provides the desired operational position. The AdaBoost-based motion recognition is employed to discriminate different movements based on sEMG of human upper limbs. The operational space motion planning for bionic arm and force planning for multifingered robotic hand can be both transformed as a convex optimization problem with various constraints. A neural dynamics optimization solution is proposed and implemented online. The proposed formulation can achieve a substantial reduction of computational load. The actual implementation includes a bionic arm with dextrous hand, high-speed active vision, and an EMG sensors. A series of manipulation tasks consisting of tracking/recogniting/grasping of an object are implemented, and experiment results exhibit the responsiveness and flexibility of the proposed sensory motion fusion approach.

Robust Stabilization of a Wheeled Mobile Robot Using Model Predictive Control Based on Neurodynamics Optimization

In this paper, a robust model predictive control (MPC) scheme using neural network-based optimization has been developed to stabilize a physically constrained mobile robot. By applying a state-scaling transformation, the intrinsic controllability of the mobile robot can be regained by incorporation into the control input u1 an additional exponential decaying term. An MPC-based control method is then designed for the robot in the presence of external disturbances. The MPC optimization can be formulated as a convex nonlinear minimization problem and a primal- dual neural network is adopted to solve this optimization problem over a finite receding horizon. The computational efficiency of MPC has been improved by the proposed neurodynamic approach. Experimental studies under various dynamic conditions have been performed to demonstrate the performance of the proposed approach.

Dynamic Balance Optimization and Control of Quadruped Robot Systems With Flexible Joints

This paper investigates dynamic balance optimization and control of quadruped robots with compliant/flexible joints under perturbing external forces. First, we formulate a constrained dynamic model of compliant/flexible joints for quadruped robots and a reduced-order dynamic model is developed considering the robot interaction with the environment through multiple contacts. A dynamic force distribution approach based on quadratic objective function is proposed for evaluating the optimal contact forces to cope with the external wrench, and fuzzy-based adaptive control of compliant/flexible joints for quadruped robots is proposed to suppress uncertainties in the dynamics of the robot and actuators. The dynamic surface control approaches and fuzzy learning algorithms are combined in the proposed framework. All the signals of the closed-loop system have proven to be uniformly ultimately bounded through Lyapunov synthesis. Simulation experiments were performed for a quadruped robot with compliant/flexible joints. The benefits of its tracking accuracy and robustness indicate that the proposed framework is promising for the robots with payload uncertainties and external disturbances.

Intelligent surface normal measurement method of end effector for the aeronautical drilling robot

In order to solve the perpendicularity of robotic drilling in the aircraft assembly, this paper proposed a novel surface normal measurement method for the aeronautical drilling robot. Firstly, vision location module is used to mark the drilling point and get the coordinate of drilling point. Meanwhile, four laser range sensors start to measure the distance between the emitting laser points and their subpoints on the workpiece surface. Thus, the coordinates of four subpoints can be obtained. According to the four subpoints and drilling point, line or circle are used to approach two curves through the drilling point on the workpiece surface. Two tangent vectors of the two curves at drilling point can be obtained. Then the normal vector of drilling point can be calculated from the cross product of two tangent vectors. The angle between the normal vector and drill axis vector can be obtained. If the angle is greater than 0.5 degree, the two eccentric discs will rotate definite angles to meet the requirement. Finally, the simulation results and experimental results on the aeronautical drilling robot show that the surface normal measurement method is valid and effective to meet the requirement in aircraft assembly.

Key Technologies and Prospects of Individual Combat Exoskeleton

With the development of modern warfare, the load-carrying of the soldier is more and more heavy. The overload affects the soldier’s ability and readiness and causes acute and chronic musculoskeletal injuries. Exoskeleton can greatly reduce the oxygen consumption of the soldiers and support energy for transferring, running, and jumping, and enhance locomotor and operational capability of the soldiers. The Berkeley Lower Extremity Exoskeleton (BLEEX), Raytheon XOS, Human Universal Load Carrier (HULC), and Hybrid Assisted Limb (HAL) are the most typical exoskeleton robots. The first three are individual combat exoskeletons in support of U.S. Defense Advanced Research Projects Agency (DARPA). The HAL is mainly used for civilian. We research and analyze the structural characteristics and joints movement of the lower limb and structural design, power system, control system, and so on key technologies of those four exoskeletons. At last, we predict the trend of prospective individual combat exoskeleton.

An effective self-adaptive mean filter for mixed noise

The image is usually corrupted by Gauss noise and impulse noise simultaneously, and its quality will be reduced. Thus filtering image is important in image processing. The traditional mean filter cannot remove impulse noise effectively while preserve the details of image well. And the median filter cannot remove Gauss noise effectively. In this paper, we propose the self-adaptive mean filter to remove mixed noise of Gauss noise and impulse noise. Firstly, dividing pixels of image into good pixels and corrupted pixels based on whether there are noises in their small neighborhoods. And the greyscale value of good pixels are output directly. Secondly, for corrupted pixels, removing Gauss noises and impulse noises respectively based on characteristics of different noise. The results demonstrate that the self-adaptive mean filter can eliminate mixed noise of different density better and preserve the details of image better comparing with the mean filter or the median filter for mixed noise.

Intelligent Double-Eccentric Disc Normal Adjustment Cell in Robotic Drilling

An intelligent verticality adjustment method named double-eccentric disc normal adjustment (DENA) is presented in precise robotic drilling for aero-structures. The DENA concept is conceived specifically to address the deviation of the spindle from the surface normal at the drilling point. Following the concept of intelligent and accurate normal adjustment, two precise eccentric discs (PEDs) with the identical eccentric radius are adopted. Indispensably, two high-resolution stepper motors are used to provide rotational power for the two PEDs. Once driven to rotate with appropriate angles respectively, two PEDs will carry the spindle to coincide with the surface normal, keeping the vertex of the drill bit still to avoid the repeated adjustment with the help of the spherical plain bearing. Since the center of the spherical plain bearing coincides with the vertex of the drill bit, successful implementation of DENA has been accomplished on an aeronautical drilling robot platform. The experimental results validate that DENA in robotic drilling is attainable in terms of intelligence and accuracy.

Application of Universal Kriging for Calibrating Offline-Programming Industrial Robots

The requirement for absolute positioning accuracy has also increased with the increasing use of industrial robots in offline programming. The present study proposed Universal Kriging (UK) for calibrating offline-programming industrial robots. This method was based on the similarities in positional errors. In addition, the method represented the positional errors as a deterministic drift and a residual part, which considered both geometric and non-geometric errors. The semivariogram was designed and the drift was determined to implement UK. Then, the method was applied for predicting positional errors and realizing error compensations. In addition, contrast experiments were performed to verify the practicality and superiority of UK compared with Ordinary Kriging (OK). Experimental results showed that after calibration by UK, the maximum of the original spatial positional errors reduced from 1.3073 mm to 0.2110 mm, that is, by 83.86%. Moreover, the maximum of the spatial positional errors reduced from 1.3073 mm to 0.3148 mm by only 75.92% after calibration using OK. An evident increase was reported in the maximum of the spatial positional errors from 0.3148 mm to 0.2110 mm, with an improvement rate of 32.97%. This is of great significance when accuracy is less than 0.5 mm. Overall, the experimental results proved the effectiveness of UK.

A compensation method based on extreme learning machine to enhance absolute position accuracy for aviation drilling robot

To enhance the absolute position accuracy and solve complex modeling and computational complexity problems in traditional compensation methods for aviation drilling robots, a compensation method based on the extreme learning machine model was proposed in this article. The proposed method, in which the influence of geometric factors and the non-geometric factors of the robot is considered, builds a positional error prediction model based on extreme learning machine. As the input and output training data, the theoretical position and positional errors measured by a high-precision laser tracker were used to train and construct the extreme learning machine model. After the extreme learning machine model was constructed, the positional errors of prediction points could be predicted using the trained extreme learning machine. Then, the drilling robot controller could be directed to compensate for the predicted positional errors. To verify the correctness and effectiveness of the method, a series of experiments were performed with an aviation drilling robot. The experimental results showed that choosing an appropriate number of training points and hidden neurons for extreme learning machine could increase the computational efficiency without decreasing the high absolute position accuracy. The results also show that the average and maximum absolute position accuracy of robot tool center point were improved by 75.89% and 80.93%, respectively.

Trajectory tracking control method and experiment of AGV

Aiming at the trajectory tracking control of AGV in aeronautical manufacturing and automated warehouse, a novel trajectory tracking control method was proposed. Firstly, the kinematic model and trajectory tracking control model of AGV were established. After analysing the reversing problem of AGV in practical situation and the limitations of classic trajectory tracking control method, and based on the similarity between the forward mode and reversing mode of AGV, a novel trajectory tracking control rule was designed and its stability was proved. The trajectory tracking control rule had simpler, less amount of calculation and improved the flexibility and handling efficiency of AGV. Finally, the trajectory tracking control experiments of sinusoid and path consisting of segments and arcs were done on experimental platform. The experiment results verified that the trajectory tracking control method were effective and feasible. The method could meet the requirements of autonomous handing for AGV in aeronautical manufacturing and automated warehouse.