A dynamics model was built for the steer-by-wire(SBW) system to study the road feel generation mechanism to deal with the need of simulative generation of road feel of SBW system. Based on the analysis of the assist characteristic the returnability and the active damping control strategy of the electric power steering(EPS) system, an integrated road feel simulation control algorithm was proposed. In the algorithm, the EPS rack force information was acquired by measuring the steering motor current. The hardware-in-the-loop tests were performed under on-center and lemniscate test conditions. The results show that the proposed road feel simulation method guarantees the steering easiness at low speeds and road feel clearness at high speeds, satisfying the driver need for road feel feedback.

Starting with the whole vehicle energy consumption, the principle of energy flow of electric vehicle was analyzed. The relational functions of electric power consumption to the driving condition, the vehicle parameters, the component efficiency and braking energy recovery ratio are established. By calculating the feature weighting of the driving condition, the potentials of energy-saving of different vehicles under different driving conditions can be analyzed to provide theoretical foundation for matching of parameters of electric vehicle. The function of braking energy recovery ratio was built to calculate the vehicle energy-saving rate based on the braking force allocation algorithm. The simulation model was built on Matlab/Simulink platform to calculate the energy-saving potential and energy recovery efficiency. In comparison with AVL/Cruise simulation, the calculation method of this paper is more effective in analyzing vehicle energy-saving potential.

To obtain the parameters in the Fancher leaf spring model with the step method and experiments often leads to large randomness in the engineering praxis. A one-time non-linear fitting method to obtain all parameters without estimation was proposed to reduce the calculation error. The Fancher model was also discussed to deepen the understanding of the mechanism and its differential equations in the aspects of application scope and limitation. The results of simulation were compared with those of leap spring characteristic experiments showed the proposed parameter fitting method was characterized by high accuracy, being able to simulate the non-linear hysteresis characterstics of the leaf spring.

In order to speed up the vehicle inspection in the production line, the latter was simulated and optimized by the software Flexsim. A method to analyze the existed unballance factors in the production line and to ballance it was proposed based on the genetic gene theory. The simulation results show that it needed 183 min to finish the vehicle inspection for the unballanced production line, while after the ballance optimization of the production line by the genetic gene theory, it needs only 141 min. The effectiveness of the proposed method was proved.

The flow-guiding effect of the stator wheel in the fluid circulation of the hydrodynamic torque converter was analyzed and a bladeless simulation method was proposed to obtain the torque characteristic and circulation flow performance of the converter under idling condition of the stator wheel. The idling speeds of the stator wheel were monitored at some set monitoring points. The comparison of the results with the traditional simulation resuts and the existing experiment data showed that the proposed bladeless simulation method is characterized by higher simulation accuracy with less computation work of the stator wheel idling simulation.

An innovative design method of the machine tool for manufacturing axisymmetric asphere was put forward. Based on the configuration charactersitic of axisymmetric asphere surface, the relationship of position and orientation between the tool and the workpiece was analyzed. The matrix of position and orientation of the tool relative to the workpiece was set up. The requcired motion units were proposed and the motion function scheme of the machine tool for manufacturing the axisymmetric asphere surface was generated. The design results were optimized and the structure configuration of the machine tool was obtained according to the expressions of motion structure.

Referring to the existed skill of centroidal Voronoi tessellation(CVT) generation, the curvature function of surface was studied and modified to satisfy the demand of the density function of CVT, and the convergence of cost function was regarded as the end criterion of algorithm, a new sampling method was proposed based on CVT. The adaptive distribution of the sampling points was realized by the proposed method. The method was comparatively validated with the traditional sampling methods by simulation tests. The results show that the method is better than the traditional ones, and it is stable and efficient for the common sculptured surface.

A model optimization was performed for the crashworthiness of the thin-walled taper tube with square cross section which often appears as an energy absorbing part of the crash box. Taking the crash force efficiency and the specific energy absorbing part of the crash box. Taking the crash force efficiency and the specific energy absorption as evaluation indexes, a multi-objective optimization model was built in the form of weighted the distribution of inducing grooves on the energy absorption and crash force of the structure were analyzed to choose the feasible region of inducing grooves. Taking the number, the non-uniform intervals and depth of inducing grooves as optimization parameters, choosing the sampling points reasonably, the surrogate models were constructed by the cubic polynomial response surface method and the radial basis function method respectively. The optimization design was performed by the particle swarm method and the optimal number and distribution of inducing grooves were obtained. The effectiveness of the proposed method was validated by the finite element simulation analysis.

A co-simulation platform was built using softwares GT-Power and Simulink to analyze the engine transient performance. The platform was used to simulate the transient process of a turbocharged engine, and a modified turbocharger variant was put forward to improve the transient performance of the engine. The simulated results were verified by engine tests, showing that the time for the load step of constant speed and increasing torque was reduced by 31%, and the vehicle accelerating time from 0 to 60 km/h was reduced by 8.7%. The test result trend was in good agreement with that of simulation, proving the feasibility of predicting the engine transient performance by simulation. The proposed simulation method may be used to analyze and evaluate the transient performances of the engine and the whole vehicle at the early stage of engine performance development, guide the matching of the turbocharger with the engine.

The diamond-like carbon (DLC) film was deposited on the valve train tappet of a FAW(First Automobile Works) production engine by dc magnetron sputtering which using methane and argon as the precursors. The influence of the tribological properties of the DLC valve train tappet on engine fuel economy was studied by the test in the universal material tester and the electrically driven valve train test rig. The test results show that friction factor of the DLC valve train was reduced by 67% and the ability of wear resistance was enhanced. Mechanical friction loss of engine valve train was decreased by 6%. DLC coating of engine component makes engines more efficient, so this technology is suitable for the low carbon technical route.

The evolution of underground temperature field of the pumping and injecting wells area with different aquifer structures was investigated experimentally. The transport behavior of the thermal-wet of different aquifer media and their thermal interaction were analyzed under same total water flow rate and well geometric parameters. The experimental results showed that the effect of aquifer rockyness variation on the occurrence time of the thermal breakthrough is significant while relatively small on the later thermal interaction. For the aquifer with larger size of gravel, the thermal breakthrough advances significantly. The effect of aquifer distribution on the thermal breakthrough is less than thermal interaction. For the homogeneous aquifer medium, the water flow rate of each well should be evenly distributed to postpone the thermal breakthrough and weaker the thermal interaction. For the inhomogeneous aquifer medium, each well should have individual flow adjustment to improve the thermal-wet balance and avoid the adverse influence of aquifer in homogeneity.

Against the traditional green wave control which considers only to maximize the public greeen wave bandwidth(GWB), a new phase-offset optimization way was proposed which considers the public GWB as well as the additional bandwidth for a part of the connected intersections to maximize the integrated GWB. The corresposnding objective functions of the control were designed, the calculation method of GWB and its optimization process were given by combination of diagram method and trial-and-error method. The case study showed that although the public GWB obtained by proposed method appears a little reduced than that given by the traditional control method, the proposed control method is better in control effect.

The real-time traffic flow data, crash data, crash data and weather condition data were extracted and collected from a 23-km segment of freeway I-880 N in the state of California of the United States. A real-time crash risk prediction model on freeways was built based on the traffic flow data and weather data using the Logistic regression model. The data analysis results showed that the weather condition variables had significant impact on the likelihood of crash occurrence on freeways. The odds ratios for rainy and foggy days were 6.4 and 4.4 respectively, indicating that the crash risks for rainy and foggy days were 5.4 and 3.4 times respectively higher than that for clear days. A Logistic regression model was also built based on only real-time traffic flow data for the comparison purpose. The analysis results show that the prediction accuracy of the model with weather variables was 71.7%, while it was 66.5% without these variables. The weater condition variables significantly enhance the prediction accaracy of the real-time crash risk prediction model on freeway.

Vehicular experiments were conducted on Chang-Yan expressway to collect the required data. Using mean RGB of landscape and drive time as independent variables, mean heart rate(MHR) as dependent variable, the function between the three variables were developed based on trend surface analysis method. Then sensitive analyses of the two independent variables were conducted based on single factor method. The results show that mean RGB color is negative correlated with MHR, and with the increase of color brightness, the fatigue level of the driver decreases. Drive time is positive correlated with MHR, and with the increase of drive time, the MHR also increases, which indicates that the driver becomes more fatigue. The results also show that MHR is more sensitive to drive time than to landscape color.

The finite element method was used to build a finite element model for the external prestressed simply supported steel beam to simulate dynamically the natural frequencies of the beam under different prestress forces and different prestressed reinforcement eccentricity arrangements. Based on the results, the behavior and the extent of the effect of the prestress force on the natural frequencies were analyzed and the quantitative relations between the prestress force and the natural frequencies of the bridge were obtained, giving a theoretical support for the idcntitication of the effective prestress force in the prestressed simply supported beams.

The natural vibration analysis and the stability analysis were performed for a large-span half-through concrete filled steel tube(CFST) arch bridge. The effects of the design load, rise-to-span ratio and arch surface inclination angle on the natural vibration behavior, the elastic stability and the elastic-plastic stability of the large-span CFST arch bridge were investigated. The results show that reducing the rise-to-span ratio or increase inclination of the arch surface improve effectively the transverse out-of-plane vibration frequencies of the structure, and make the structure stability change in different way. Integrated consideration of the natural vibration and the elastic stability to evaluate the parameter selection of structure is more reasonable than that considering the natural vibration only. For the investigated 360 m span half-through CFST arch bridge, the rise-to-span ratio should be around 1/4, and the arch surface inclination should be less than 10°.

In view of the fact that the visual measure system with single camera can not measure the shield between the measured objects each other, a networked technique was proposed for this kind of measure system. The optical characteristic spots on every network control point were measured by the single camera measure system, and the measured coordinates were normalized into the same world coordinate system to get the global data, and the overall measurement was achieved. Applying the weighted data fusion algorithm, distributing the weight factors according to the stability theory of eigenvector, fusing the global normalized space coordinates, and the ultimate expression of fusion estimate was derived. The experiment in comparison to the simulation was performed in a CMM and the results showed that the maximal absolute error of the space coordinates is 0.035 mm after normalization by quaternion and Euler angles algorithms. The proposed algorithm apppears more accurate and more stable than the point set center fusion algorithm, and in the application of networked normalization of measure system, the quaternion algorithm is better than the Euler angles algorithm.

In this paper, a Loop Tightness Algorithm (LTA) is proposed. First, it finds the network loops and calculates it tightness value quickly. Then, it obtains the communities of the networks based on the tightness values. Finally, it reveals the relationship between the network loops and the community structure. The LTA is tested and validated by means of synthetic networks and real networks.

In model-based diagnosis, computing minimal hitting sets is an important part of the representation of candidate diagnostic results. In this paper, a new method is proposed to obtain the minimal hitting sets by matrix. In this method, the related information of the conflict sets is stored in a matrix. The frequencies of elements of the set cluster are taken as the heuristic information to compute the minimal hitting sets. The data structure of the algorithm is simple and the implementation of the algorithm is easy. The introduction of heuristic information reduces the number of the generated nodes. This algorithm can generate and only generate all the minimal hitting sets. Experiment results show that the proposed algorithm has a better computational efficiency than other algorithms.

The broad boundary region is an important method for indeterminate region representations. Existing qualitative direction relation models often have a huge number of basic relations, which is quite difficult to deal with, and the partitions of the direction tiles do not meet well with human habits. To solve these problems, the cone-based cardinal direction relation is selected as the base model. The direction between two broad boundary regions is represented by a quad tuple, whose elements are the basic cone-based direction relations of different combinations of the internal and external crisp regions composing the broad boundary regions. The constraints inside the quad tuple are illustrated; the inverse and composition operations are discussed; then the rules of how to transform a constraint network of broad boundary regions into the network of crisp regions are given. The proposed model reduces the number of relations and is more in line with human cognition.

There exist representative models for spatial topological relationships, such as the Region Connection Calculus (RCC) model and Intersection model. However, most of the models are suitable for the spatial relation between two regions, but few for three regions. In this paper, we extend the 4-intersection matrix model to 8-intersection model to represent the spatial relationship among three simple regions. Using this model we can obtain 109 topological relationships among three simple regions, which are 56 more than that can be obtained by the 4-intersection matrix model. We proved the exclusivity the completeness of the 109 topological relationships. We also present an algorithm to show that at most there are 109 feasible topological relationships among three simple regions.The conceptual neighborhood graph of 109 ternary topologiral relations is also given.

A polynomial algorithm of codiagnosability verification of discrete event systems was proposed, which can be used to verify F-codiagnosability and NF-codiagnosability of system respectively. Algorithm is achieved by constructing a test automaton which extracts the fault path and the normal path to compare. In addition, the traditional codiagnosability was divided into two case, it can improve the efficiency of the algorithm. Both complexity analysis and case study have shown that our algorithm has more lower computational complexity. The experimental results verify the effectiveness and efficiency of the method.

A new Global k-Prototype (GKP) algorithm is proposed for clustering mixed numeric and categorical data. First, the algorithm randomly selects a sufficiently large number of initial prototypes to account for the global distribution of the data sets. Then, it progressively eliminates the redundant prototypes using an iterative optimization process with an elimination criterion function. Systematic experiments were carried out with data from widely used datasets in this area. Experimental results and comparative evaluation show the high performance and consistency of the proposed algorithm. Compared with other well-known mixed data clustering algorithms, the proposed algorithm significantly improves the clustering accuracy.

Existing E³ algorithm has the drawback of long convergence time, which can not be efficiently applied in practice. To overcome this problem, an improved E³ algorithm based on parallel sampling with multiple agents and learning experience reuse is proposed. In order to build an approximate model soon, in the exploration phase, the multiple agents explore the environment in parallel and collect the information of the environmental model. In the exploitation phase, the optimal value function is retained and reused to speed up the convergence. Experiment results show that the improved algorithm consumes much less to converge to an almost optimal policy and has great advantage over other two parallel reinforcement learning algorithms.

The network structure of most of current multiple mobile robot systems is static when they are running. So it is difficult to alter the number of the robots. Most of the systems need complicated configures when launching. In this paper, a multi-robot system, UMRS-1, is proposed, which uses distributed architecture, supporting heterogeneous mobile robots. In UMRS-1, member robot can join in or quit from the system dynamically. The architectures of member robots and UMRS-1 are analyzed. It uses KQML-based Contract Net Protocol (CNP) as cooperation protocol. The multi-robot hunting problem is discussed to show the communication and cooperation mechanisms of UMRS-1.

Using ant colony algorithm to solve Traveling Salesman Problem (TSP) has certain disadvantages, such plunging into local minimum, slower convergence speed and so on. In order to find the optimal path accurately and rapidly, an improved ant colony algorithm is proposed. The improved algorithm uses adaptive adjusting mechanism of pheromone decay parameter to adjust the convergence speed and ensure the global search ability. With the help of common path, the computation time is reduced and the better solution can be obtained. Experiments show that, in a relatively few iterations, the percentage deviation of the average solution to the best known solution is 0.46%, and the percentage deviation of the best solution to the best known solution is 0.23%. The improved algorithm has high accuracy and convergence speed.

According to the digital representation, content retrieval and culture calculation of paper-cut, an eXtensible Markup Language (XML) data storage model is constructed. The model can not only realize two-way conversion from the paper-cut image to XML data, but also prepare data for paper-cut content retrieval, data mining and culture calculation. To evaluate the feasibility and advantages of the model, its supportive ability to find known cultural patterns is tested. Results show that, by adding necessary structure and character information, the data sets based on the model possess the ability to support data mining and culture calculation. The model design procedure of this research is also applicable to other non-text resources.

To improve the performance of Wireless Sensor Network (WSN), a concentric ring cluster routing algorithm based on coverage (CRAC) is proposed. This algorithm divides the WSN into rings using the coordinate of concentric ring by Sink. Meanwhile, it works out the coverage partitions of every ring and the rotational clusters according to QoS of application requirements. Different sensor nodes of the rotational clusters in the same ring can alternate into sleep mode to save energy. By judging the energy equilibrium equation, each ring can decide itself whether to be the sending ring. The data gathered is collected to the sending ring after it is in parallel transmitted to the path of transmission. Then, the data is sent to Sink node by LEADER node. Simulation results show that the proposed CRAC can greatly reduce both delay time and energy dissipation, prolong network lifetime and make the network energy bourn evenly.

A new Virtual Network Embedding (VNE) algorithm is proposed, which improves the original sub-graph isomorphism search process, and overcomes the defects in existing VNE algorithm. First, a node resource evaluation method is proposed, which takes both node resource requirement (capability) and topology attribute into account, to improve the mapping order of the virtual nodes. Second, the algorithm improves the selection process of candidate substrate nodes when mapping the virtual nodes, which enhances the quality of virtual link mapping. Third, the algorithm improves the selection of resource allocation sub-area in substrate network by considering the total resource capability in the sub-area. Experiment results show that the proposed algorithm performs better in mapping quality, revenue, acceptance ratio and runtime compared with existing algorithms.

We propose an algorithm, HSRDS, which achieves the sensitive association rule hiding based on data stream. In this algorithm, we adopt Transformative FP-Tree (TFP-Tree) as transitional structure between original data set and sanitized data set, and complete the sensitive association rule hiding of the newest data combined with sliding window. Furthermore, in the progress of hiding sensitive association rule, we inject two thresholds that are σ and δ, to guarantee that the sensitive association rule can be hidden successfully, at the same time the side effect, resulting from the hiding sensitive rule, can be minimized.

To solve the defects of the traditional (α, k)-anonymity model has the defect that it does not satisfy the personalized requirement of different privacy preserving degree for different sensitive attributed values. To overcome this defect, the concepts of personalized privacy sensitive factor and personalized privacy preserving requirement degrees for each sensitive attribute value are introduced. Then the personalized (α, k)-anonymity model is defined formally. Meanwhile, an entropy-based classification approach for personalized privacy anonymity is presented to solve this personalized (α, k)-anonymity model. Experiment results show that the proposed method not only produces similar information loss and time cost to the existing (α, k)-anonymity algorithm, but also meets the requirements of personalized service and achieves more reasonable privacy preservation.

To solve the problem of face detection under rotation in image plane, an improved Scale Invariant Feature Transform (SIFT) algorithm is proposed. With the invariant properties of rotation, scale, translation and illustration, the SIFT algorithm is used to detect the rotation face in image plane. First, a pruning algorithm, combining the template and contrast, is used to increase the efficiency and filter the invalid key points; and a higher accuracy main direction algorithm is proposed. Then, the AdaBoost algorithm is used to train a common face classifier and compute the matching ratio of key points. Finally, the SIFT algorithm detects the face under rotation in image plane, and normalizes the face by the main direction of key points. Experiment results indicate that improved algorithm preserves the high detection ratio, and performs better in normalization precision and detection efficiency.

According to the strengths and weaknesses of the L₂-norm Support Vector Machine (SVM) and the L₁-norm SVM in the classification analysis of cancer gene, a Doubly Regularized Support Vector Machine (DRSVM) is applied to the DNA microarray classification based on the Bhattacharyya distance, which is used to eliminate most of the unimportant genes and gain a few highly correlated important genes for classification. A quadratic polynomial loss function changes the constrained optimization into unconstrained and differentiable optimization, which can be computed by Brogden-Fltcher-Goldfarb-Shanno (BFGS) algorithm. Experiment results on two kinds of tumor gene data sets show that this method is effective and feasible.

An iris recognition method based on improved empirical mode decomposition is proposed. First, the normalized iris image is decomposed based by row and then column to generate the different layer intrinsic mode components of the image. Second, the feature image is obtained by binarizing the components useful for the iris recognition. Third, the Hamming distance matching vector is obtained by horizontal and vertical shift match. Finally, the improved standard deviation of the matching vector is calculated, which is used as the threshold for iris recognition. This method is tested using CASIA1, CASIA2, CASIA3-Interval and MMU1 databases. Experiment results show that this method can extract the binary feature effectively, with faster speed and higher correct recognition rate.

A novel kind of DNA computing model for graph vertex coloring problem is proposed based on microfluidics. This model can not only improve the reliability of DNA computing, but also reduce the computing time and experiment operation. The hardware structure of the model is described. The operation steps are introduced using a case study. The simulation results show that this computing model can be used to solve any instances of graph coloring problem automatically.

In order to increase the prediction accuracy and generality of video quality assessment, a no-reference video quality assessment method based on distortion estimation is proposed. In this method, first, the local distortion is calculated using the statistical characteristics of the difference between contiguous pixels and the global distortion is calculated by measuring the detail loss of the video after Gaussian Filtering; then, the video distortion is estimated by the combination of the local and global distortions. Second, the video complexity is calculated through intra and inter frame prediction. Finally, the objective video quality is obtained utilizing both the video distortion and complexity. Experiment results show that using the proposed method can achieve better accuracy and generality compared with other methods.

This paper proposes a hybrid carrier Code Division Multiple Access (CDMA) multi-antenna system based on weighted-type fractional Fourier transform to improve the compatibility of existing systems. In this system, CDMA technique is applied in a certain fractional Fourier transform domain, and the signals after spreading are coded by space-time block coding and mapped to the corresponding antenna ports. The proposed system can switch single carrier or multi-carrier CDMA multi-antenna system under specific selected condition. Moreover, the signal energy distribution of the hybrid carrier system is more uniform compared to the single carrier or multi-carrier system, which can resist channel fading effectively, resulting in better system performance.

The random phase coding method of Transform Domain Communication System (TDCS) is studied. To improve the multiple access capability, a new Spread Phase Coding TDCS (SPC-TDCS) scheme is proposed. The framework of the SPC-TDCS is given. The selection of the spread sequence and the cross correlation of the spread basic function are studied, the system performances of the proposed SPC-TDCS are analyzed. Simulation results demonstrate that the proposed scheme has good system performance in a multiple access environment and guarantees the minimum interference among different users at the same time.

This paper focuses on the problem of eavesdropping adversaries in network coding field. Based on Maximum Rank Distance (MRD) codes, a new strongly secure linear network coding algorithm is proposed. In this algorithm, message is pre-encoded at the source node by using coset coding scheme that uses MRD code over extension filed. Meanwhile, focusing on the intermediate nodes, linear network code over the base field is used. It is proved that the proposed algorithm can transform a non-secure linear network code to a strongly secure code. In other word, the algorithm is a general strongly secure linear network coding algorithm. Also, the proposed algorithm can reduce the coding field size of intermediate code compared to current strongly secure algorithm.

A large number of real data traffic volumes per cell in five minutes at Gn interface of China General Packet Radio Service (GPRS) core network are collected for exploring the traffic model. The results show that most cells at most time of a day are at low speed; on the contrary, a very small number of bursts centralize in a few cells in at a little time. Scaled exponential model is suitable to fit by intuitive Q-Q plot and a related discrepancy measure. Traffic volumes samples at different cells and at different time are adopted to verify the validity of the traffic model. To a certain extent, the above conclusions are found universal.

As the transmission bit rate is limited for the multipath fading effect in troposcatter communication, Orthogonal frequency division multiplexing (OFDM) technique is introduced to eliminate the Inter-Symbol Interference (ISI) in high transmission system. First, the troposcatter multipath fading channel is modeled. Then, the performance of OFDM signal over tropo channel is analyzed theoretically, and the components of the received signal are investigated. Finally, the Symbol Error Rate (SER) of the considered system is achieved. To improve the SER performance, the Reed-Solomon coding technique is adopted. Simulation results show that high transmission rate is possible as the OFDM scheme is employed in troposcatter communication. With the coding technique, the SER performance of the OFDM scheme over tropo channel meets well with military communication requirements at high transmission rate.

A universal Progressive Edge Growth (PEG) algorithm with the Kronecker product is proposed for constructing Quasi-Cyclic Low-Density Parity-Check (QC-LDPC) codes. The Kronecker product is introduced for constructing the basic matrix. The check matrix of QC-LDPC codes is designed based on the demand of the node degree distribution. It is proved that the girth of the check matrix is larger than Girth-8. Simulation results indicate that the proposed algorithm performs well to maintain the low density and bite error rate characters of the constructed QC-LDPC coeds. It can be used for the design of regular and irregular QC-LDPC coeds.

The corner-type feature based on photometric invariants and the Histogram of Oriented Gradients of Edge Magnitude (HOG-EM) statistical feature are proposed to overcome the interference of the texture-rich non-text regions to the text localization algorithm. A two-stage multilayer text localization algorithm in complex scene is presented on the basis of the two novel features. In the proposed method, first, edge map is obtained and eight layers of binary maps in the Hue Saturation and Lightness (HSL) space are generated according to the characteristics of the HSL space. Then, nine layers of sub-maps are formed to gain text candidate blocks with multilayer connected component analysis. Finally, the two novel features mentioned above are extracted to remove the non-text block from the text candidate blocks and keep the text. Experiments indicate that the proposed scheme can efficiently remove non-text texture-rich regions, decrease false-alarm rate and obtain reasonable accuracy and recall rate.

According to the output signal characteristics of the profile detecting system, special object recognition method based on sparse representation, combined with the theory of sparse representation and principal component analysis, is proposed. First, using principal component analysis, the method extracts the main components of the sample signal in order to eliminate redundant information. Second, the signal is transformed into the same size of the feature vectors, which is then projected to the lower dimensional space to construct a dictionary. Finally, the testing samples are sparsely represented and recognized by the dictionary. Numerical simulations and experiments show that the proposed method has good classification effect in lower dimensional space, and good robustness for the system with some damage sensors in the actual situation.

An Adaptive Square Root Central Difference Kalman Filter (ASRCDKF) algorithm is proposed, which is used in the initial alignment of Strapdown Inertial Navigation System (SINS) with large azimuth misalignment angle. The basis of this algorithm is central difference transform, and square root of covariance is used instead of covariance in recursive calculation based on the theory that square root filter can overcome filter divergence. In addition, adaptive estimation principle is introduced in this algorithm. The ASRCDKF algorithm not only overcomes the shortcomings that extended Kalman filter can generate linear error and it needs to calculate Jacobian matrix, but also reduces calculation amount to ensure numerical stability. Furthermore, this algorithm solves the problem that traditional filter algorithm relies on a prior knowledge of the system dynamic model and noise statistic characteristics. Simulation results fully verify the effectiveness and superiority of the ASRCDKF algorithm in SINS initial alignment with large azimuth misalignment angle.

An adaptive sliding mode control method is proposed for the multivariable rigid-flexible coupling system of manipulator handling a flexible payload. The position control and vibration suppression are conducted under the system with parameter uncertain and unknown disturbance. The dynamic behavior of the system is described by a distributed parameter model in terms of partial differential equation, which avoids the drawback of the lumped parameter model. The adaptive sliding mode controller is designed based on the energy dynamics of the flexible payload. The sliding surface is selected as the coupling of the joint angle error and angular velocity error with the root strain and shear force of the flexible payload. The asymptotic stability of the closed-loop system is proved using the LaSalle invariance principle. Simulation results demonstrate the effectiveness of the proposed method.