Abstract: It is very difficult to select appropriate networked control system parameters. The adaptive control methods of fuzzy dynamic characteristic modeling and characteristic modeling are easy to be used in engineering field, and they can ensure high precision control according to practical demands. The networked control performances of a DC motor control system with different time delays and packet drop ratios were researched, and the adaptive control methods based on characteristic modeling and fuzzy dynamic characteristic modeling were proposed for the control system. Simulation results show that packet drop ratio has a greater influence on the networked control system with characteristic modeling but time delay has a greater influence on the networked control system with fuzzy dynamic characteristic modeling. The proposed methods are effective to the networked control system with time delay and packet drop ratio.
Abstract: A new sparse echo state network (ESN) with a leaky integrator, which is expected to has more neurophysiology characteristics, was proposed and trained using the online supervised learning method so as to make the modeling and prediction of the matching decision-making problem. To evaluate the matching decision-making performance of the network, three kinds of test datasets were set up and an estimation method based on the maximum correlation coefficient for the actual output and the desired one was present. Simulation experimental results show that the proposed model can achieve a better decision-making performance with a less training time. Meanwhile the model has a better robustness on spiking interval change, shifting, and network noise.
Abstract: A nonlinear PID-like intelligent controller based on nonlinear proportional, integral and differential components was studied. Firstly the three components are respectively expressed as the trigonometric functions of error signals, and then these three independent nonlinear functions synthesize the intelligent controller. Through online adjusting the weights of the three independent nonlinear functions, the intelligent controller can control nonlinear objects and is independent of the nonlinear object model. Simulation results show that the intelligent controller possesses an excellent non-linear control performance.
Abstract: A method is introduced for designing a controller based on the feature data of a control system. According to the input and output data of a controlled object only, this method identifies an impulse response by means of the convolution theorem. The impulse response data are taken as the feature data of the control system. With the controlling signal in a certain scope, the output reference data and the fastest control signal are calculated according to the feature data to meet the requirements of the fastest response.Then the ideal feature data of the controller are worked out on the basis of the control goal and the fastest control signal. Finally the linear parameters of the controller are identified by the feature data. Simulation results show that the control result has the characteristic of the fastest reference control.
Abstract: Based on the backstepping technique, introducing the integral-type Lyapunov function and utilizing the approximation capability of neural networks, an adaptive neural network control scheme was proposed for a class of stochastic strict-feedbagk nonlinear systems with unknown virtual control gain. Compared with existing literatures, the proposed approach relaxes the requirements of the control system and cancels the restriction of the unknown function, By the Lyapunov method, it is shown that all error variables in the closed-loop system are bounded in probability. Simulation results illustrate the effectiveness of the proposed control scheme.
Abstract: A kind of improved mRMR SBC was proposed by using K-means clustering and incremental learning algorithms to enlarge the scale of training samples. On one hand, the testing samples are labeled using the K-means clustering algorithm and are added to the training set. A regulatory factor is introduced into the process of attribute selection to reduce the risk of mislabel resulting from K-means clustering. On the other hand, some samples that are most helpful for improving the current classification accuracy are selected from the testing set and are added to the training set. Based on the enlarged training set, parameters in the Bayesian classifier are adjusted incrementally. Experimental results show that compared with mRMR SBC, the proposed Bayesian classifier has better classification results and is applicable for solving the classification problem for the high-dimensional dataset with little labels.
Abstract: A general scheme for the automated classification of gait patterns based on time-frequency analysis was proposed to discriminate acceleration signals characterized by high dimension, non-linearity, strong coupling and high time-varying acquired under different terrains and motion patterns of lower limbs. A three-axis acceleration sensor was mounted on a crus to acquire acceleration signals in the sagittal, coronal and cross-sectional planes separately. By using a 5-order Daubechies wavelet base, the features were extracted from time-series acceleration signals and further dimensionally reduced by employing linear discrimination analysis (LDA). The reduced features were classified by the decision tree and the support vector machine (SVM). From experimental results, both classifiers can achieve the high classification accuracy ratio over 90% and for the specified gait the ratio can be up to 100%, indicating the rationality and effectiveness of the proposed methods for feature extraction and dimension reduction.
Abstract: Due to urgent demands for real time relative motion patterns mining applications, an efficient cluster-recombinant (CLUR) algorithm for real time discovering closed swarm patterns was proposed. The algorithm maintains a candidate swarm list, and at each timestamp carries out cluster analysis on moving objects using the clustering algorithm based on density, and according to the clustering results it recombines the maximum moving object set and records the corresponding maximum time set, further constructs a candidate swarm pattern and then finally updates the candidate swarm list up to date by using three update rules and an insert rule. The rules greatly reduce the redundancy of the candidate list and improve the efficiency of the algorithm. At the end of each timestamp, the current closed swarm patterns can be real time obtained by closuring checking rules. Comprehensive empirical studies on large synthetic data demonstrate the correctness, real time and efficiency of the CLUR algorithm. The CLUR algorithm can be applicable to real time relative motion pattern mining systems.
Abstract: A novel map matching method based on fuzzy neural networks was proposed. This method integrates digital road map information and GPS/DR position data, and two important variables, the projection distance from the positioning point to the candidate link and the angle difference between GPS/DR heading and the link bearing of the candidate link, are selected as input signals for fuzzy neural networks. A four-layer fuzzy neural network was designed and the improved learning rule was acquired for the fuzzy neural network. Experimental results show that the proposed algorithm has very good performance for matching the position of car running to the correct link under normal traffic conditions.
Abstract: Eight extended feature prototypes were presented by combining rectangular feature blocks and triangular feature blocks. In consideration of the fact that the amount of eye image blocks is far less than that of non-eye image blocks during a scanning block passing through face images, a fast eye location detection scheme based on AdaBoost algorithm combining rectangular feature blocks and triangular feature blocks was proposed. After most of non-eye blocks are excluded through the foregoing strong classifiers, most eye image blocks and a few of non-eye image blocks are detected through the rear parts of the cascade classifier, which can reduce the detection time and boost the detection speed. The experiments further show that the scheme has better detection performance and positive detection rate compared to the case only employed Haar features.
Abstract: A novel quadruped robot with variable structure was analyzed by kinematic methods. First, the velocity of a single leg was studied. Then the kinematics of the whole robot was discussed, and a globe velocity equation was proposed and established with considering the transformation and motion of the robot's body and four legs. At last, the method of velocity decomposition control, which is usually used in the serial-chain manipulator control, was generalized to the legged mobile robot (serial and parallel hybrid mechanism) control through applying the globe velocity equation. The contraction of the robot's body with four feet fixed on the ground was taken as an example to demonstrate the feasibility of this approach.
Abstract: A simulation model of a biped robot with torso was established using Matlab, the passive dynamic walking of this model was studied, the stability of the walking was analyzed, and a full-state linear feedback walking stabilizer was designed. Simulation results show that the model can walk down a slope without any actuator and there are two unstable walking gaits on the same slope. The results also indicate that the model can realize stable walking by adopting the full-state feedback stabilizer.
Abstract: The design of structural parameters and the simulation of motor-driven 6-DOF parallel manipulators were researched to solve the problems such as more restriction conditions, complicated interference and longer design periods during parallel manipulator design. Taking the Stewart parallel manipulator as an example, two methods of solving the workspace of parallel manipulators were discussed. Firstly, the workspace was solved by the geometric method, and then a method to solve the workspace and the load-bearing capacity by the virtual prototype technique was proposed. It is such a method that the three-dimensional design is developed through Pro/E software and the three-dimensional model is diverted to Adams software for dynamic simulation. Simulation results show that this method avoids tedious calculations and reduces the design period on condition that the simulation results are reliable.
Abstract: A rescue robot often can not directly find the target shouting for help which is in other rooms or invisible places in the building by using visual, ultrasonic or infrared ray sensors because of smoke or collapsed walls. The sound can diffract over obstacles due to the long wavelength of audio signals. Combined with the speech recognition technology, an audio navigation system was developed for the rescue robot. This navigation system makes it feasible to guide the rescue robot run to the target shouting for help in global motion control. Experimental results verified the feasibility of the navigation system.
A reinforcement learning based adaptive PID controller was presented for the attitude stabilization of a kind of bionic underwater robot with two bionic undulating fins. The scheme of the reinforcement learning based adaptive PID controller was given concretely including the control law and the parameter adaptive method based on reinforcement learning. Simulation experiments of yaw angle stabilization based on actual model parameters were carried out. The results indicate that the stabilization performance of yaw angle is improved distinctly after several iterations of learning control and the controller can overcome ordinary disturbances in short time, exhibiting its preferable adaptability.
Abstract: Taking the traveling wave equation of the fish body as a dynamic basis and according to the motion function equation of the fish tail, carangiform robotic fish named"BLRF-I"series were designed. Because of modular design in the robotic fish, the number of tail driving motors can be expediently changed and the difference among the single-link, two-link and three-link biomimetic robotic fish is only the number of tail driving motors. The cruising speed and minimum turning radius experimental results of the robotic fish show that the cruising speed grows with the number of tail driving motors increasing and the minimum turning radius reaches its minimum value while the number of tail driving motors is two, demonstrating that the number of joints has effect on the cruising performance of the robotic fish. The relation equation between the number of joints and the swimming velocity was proposed.
Abstract: The attitude control of a flexible satellite with system parameter uncertain and disturbances was investigated. A variable structure intelligent controller based on the radial basis function (RBF) neural network and the cerebellar model articulation controller (CMAC) was designed. By using the merit of insensitivity to model error, parameter uncertain and disturbances of variable structure controller systems, in combination with the characteristics of quick approximation to unknown functions and good generalization ability of neural networks, the controller overcomes the influences of parameter uncertain and disturbances, and achieves the effective control of the flexible satellite. Simulation results show that the proposed control algorithm can improve the attitude accuracy and response speed of the flexible satellite.
Abstract: A real-time fault detection and identification (FDI) scheme of time-variant signals for a complex system was studied. A sliding-window Mallat wavelet fast transform was first introduced to avoid depending on the signals in all periods for the classical wavelet transform, and the computing effect was improved, which makes sense that the real-time fault detection is effective. Secondly, aimed at the problem that it is difficult to identify the fault by using time-variant signals, an improved dynamic recurrent neural network (IDRNN) was utilized to identify the fault intelligently after detecting the fault. Finally, the scheme, including fault detection based on the sliding-window Mallat wavelet and fault isolation based on the optimized IDRNN, was applied into a satellite attitude control simulation platform to verify the online diagnosis result. Experimental results show that the sliding-window Mallat wavelet fast transform is consistent with the classical wavelet transform in real-time scenarios, IDRNN has a better generalization ability for identifying time-variant signals, and the scheme including the sliding-window Mallat wavelet and IDRNN can implement detecting the faults and classifying the multiple faults based on real-time monitoring signals for the complex system.
Abstract: An improved differential evolution algorithm was proposed for solving the online path planning problem of unmanned aerial vehicle (UAV) low-altitude penetration in partially known hostile environments. The algorithm adopts von Neumann topology and improves its structure to maintain the diversity of the population, prevent the population from falling into local optima in the early evolution and speed up the convergence rate in the later evolution as well. The mutation operator of differential evolution is improved to speed up the convergence rate of the algorithm, so that the optimal solution of the multi-objective optimization problem can be found quickly; the coding method combined the absolute Cartesian coordinates with the relative polar coordinates is used to improve the searching efficiency. The simulation experiment of online path planning for UAV low-altitude penetration shows that the proposed algorithm has a better performance than the unimproved differential evolution algorithm.
Abstract: A distributed virtual simulation system of the hypersonic vehicle X-38 based on the client/server model was presented through using Java and virtual reality modeling language (VRML). Some key technologies in the virtual simulation system, such as 3D scene reconstruction and event modeling, distributed network architecture design and database management as well as scene interface implementation, were discussed in detail. The practical running result of this virtual simulation system is good. The simulation system has good portability and scalability. It is easy to deploy under the large-scale environment and to implement the secondary development.
Abstract: In order to deal with the negative effects of load disturbances and parametric perturbation on the field oriented control system of synchronous motors, an active disturbance rejection controller (ADRC) was transplanted into the speed control of synchronous motors. A novel flux observer based on ADRC was proposed to solve the problems such as the direct current bias and parametric perturbation of conventional flux observers. Simulation results show that the ADRC ensures very good robustness to load disturbances and parameter variation with fast response, small overshoot, and good static and dynamic performances. The improved flux observer not only restrains effectively the direct current bias and parametric perturbation, but enhances the precision of the flux.
Monthly, started in 1955 Supervising institution：Ministry of Education Sponsoring Institution：University of Science and Technology Beijing Editorial office：Editorial Department of Chinese Journal of Engineering Publisher：Science Press Chairperson：Ren-shu Yang Editor-in-Chief：Ai-xiang Wu ISSN 2095-9389CN 2095-9389
