Search

In order to improve the braking performance and safety performance of electric vehicles driven by a hub motor, the cylinder pressure estimation and pressure control of its hydraulic braking system are studied. In this paper, a mathematical model is established for the solenoid valve, a key component of the hydraulic actuator, and the hydraulic and electrical characteristics of the solenoid valve are studied. A state equation is established for the solenoid valve, and the square root volume Kalman filter (SRCKF) algorithm is used to estimate the solenoid valve spool position. The brake fluid flow and brake wheel cylinder pressure are calculated based on the spool position. Finally, a solenoid valve spool position control algorithm based on sliding mode variable structure algorithm is designed, and the brake pressure in the brake wheel cylinder is controlled by adjusting the spool position. Matlab/Simulink-AMESim software simulation and hardware-in-the-loop were used to verify the algorithm. Simulation results show that the brake cylinder pressure can be estimated accurately, and the pressure control algorithm can accurately follow the control target value.

Multilingual abstracts in the five official working languages of the United Nations. (PDF 230 kb)

In the original live-wire algorithm, optimality is defined as the minimum cumulative cost path from a seed point to another, where the cumulative cost of a path is the sum of the local edge costs on the path. Therefore, the optimal path between two seed points is restricted not far away from the straight line between these two points. Consequently, when delineating a boundary with acute concaves and convexes, more seed points are required which can become very time consuming to calculate. This paper proposed an improved live-wire algorithm where the optimality is redefined as the minimum average cost path between two seed points, which frees the optimal path from the restriction of the straight line. In theory, the original live-wire algorithm is only a special case of our improved algorithm. Furthermore, the Canny edges on the boundary of the target object is weighted according to the gray or color value of the seed point. Experiments conducted on a variety of image types have shown that this improved live-wire algorithm requires less seed points than the original algorithm when delineating the same boundary and as a result, reduces the time required to complete the calculation.

A modified PSO for Optimum Design of PID Controller is stated. Particle Swarm Optimization is a new random global optimization algorithm. The algorithm feature is simple, ease to implement and powerful function. But it is easy to trap in the local pole, and search accuracy is low. PID controller principle is concise, physical meaning of its parameters are clear, and theoretical analysis system is integrity and the industrial sector are familiar with it, so in the industrial process control is still widely used. In this paper, the modified PSO algorithm has been used in PID controller to optimize parameters and compared with basic PSO algorithm. Emulation experiments demonstrated that the modified PSO algorithm increase the probability convergence to the global optimal solution.

In the applications of wireless sensor networks, the effects of data reflection largely depend on the accuracy of localization of the information collection node. Therefore, the position information of nodes is very important. Without it, events can not be sensed and measured, and value of application will be lost. To improve the location accuracy of nodes in wireless sensor networks, the paper puts forward research of node localization model in WSNs based on quantum particle swarm optimal algorithm. The results of the simulation show that the method can get significantly higher localization accuracy than the genetic algorithm and get the convergence of optimal solutions faster than the standard particle swarm algorithm under the same conditions.

PID controller for control system parameter optimization problem, we introduced Quantum-behaved particle swarm optimization (QPSO) algorithm, this method has the quantum behavior of particles, which can search the entire feasible region, the state only with particles position vector to describe, and they rely on only one parameter, the convergence speed and global convergence properties are superior to the traditional PSO. Simulation results verify the feasibility of quantum particle swarm optimization and effectiveness.

The development of digital integrated circuit has put forward urgent demands for test technology. Test technology has become a bottleneck in the application of LSI/VLSI. Especially for sequential circuits, it is still a problem which is not resolved completely in theory. By making use of the structure information of circuits, a method of automatic test generation for sequential circuits based on Particle swarm optimization is presented, which is performed by two steps, initialization and fault detection. Experimental results show that the approach can achieve high fault coverage, and CPU times needed for test generations are very short, which shows that it is a method deserving research.

Protein structure prediction, known as an NP-complete problem, is one of the basic problems in computational biology. To get an efficiency approach of protein structure prediction with Toy model, a new algorithm structure based on quantum-behaved particle swarm optimization (QPSO) structure is suggested, which is named as multiple-layer QPSO (MLQPSO). In this structure, population of each generation is divided into elite sub-population, exploitation sub-population and exploration sub-population, respectively using different strategies, sequentially leading to improve the ability of local exploitation and global exploration. Subsequently, the algorithm to predict the structure prediction is evaluated by artificial data and real protein. The experiment shows the MLQPSO is a feasible and efficient algorithm.

The taboo state is introduced in environment to discovery sub-goal. Agent samples trajectories from starting state to goal state, which contain different bottlenecks. Then the different tasks are submitted to agent. According to whether the task is accomplished or not, the bottlenecks among them are discovered. The appropriate bottlenecks are selected as sub-goal of options to be constructed according to the adjacent relationship among them. Simultaneously agent can obtain the initial set and the policies of options. Grid-world tasks illustrate that the agent can automatically construct useful options online with the proposed method, which have capability of accelerating learning and the transference of knowledge among those similar learning tasks.

For many learning tasks, samples are collected over a long period of time. The distribution of the data set is likely to change. And only a little labeled training data is available at beginning of SVM training. So the collected data could not represent the whole data set. SVM should be able to adapt to such changes and situation. In order to achieve an acceptable performance with fewer labeled training samples at beginning of SVM studying, after studying transductive inference method, an Online tranductive SVM based on feedback is well organized based on TSVM. To deal with the expending size of support vector window, the ξα Estimators is introduced, which will maintain the window on the training data by automatically adjustment, The estimated generalization error is minimized. The approach is both theoretically well founded as well as effective and efficient in practice. Two experiments show OTSVM is effective

NMR has been considered as one of the most effective physical system to realize quantum computation. As multiple-quantum operator algebra theory mentioned, any unitary transformation can be decomposed into a sequence of a limited number of one-qubit quantum gates and two-qubit diagonal gates. Based on these backgrounds, we proposed the method to design NMR pulse sequences to implement Generalized Quantum Search Algorithm with arbitrary phase rotation, and, experimental finished the algorithm with different phase rotations respectively in a two-qubit system, on a Quantum Computer Emulator.

In chain based wireless sensor network (WSN), some particular nodes called leaders aggregate the information get from its queue node and send to a base station periodically. Due to the limited battery life, energy efficiency is becoming a major challenging problem in these power-constrained networks. Chain based protocols construct a transmission chain connecting all nodes to save energy dissipation of data transmission. After reviewing several existing chain based protocols, a multichain multileader (MCML) based scheme is suggested, which is implemented by an minimum total energy algorithm for chain construction. This algorithm uses a sequence of insertions to add the least amount of energy consumption to the whole chain. It consumes less transmission power compared to the closest neighbor algorithm. In addition, an entropy based metric, the quality of data aggregation (QoDA), is defined to evaluate the performance of WSN.

In this paper, an approach of image retrieval directly based on shape feature in DCT compressed-domain is proposed. Firstly, according to the edge feature of DCT compressed image, the rough object shape is extracted from DCT compressed-domain to form a binary edge map; then the similarity measurement is redefined by using the moments invariant to improve the retrieval performance. It is supported by theoretical analysis and experimental result that this approach not only is fast, but also has a good performance. And it is also robust to rotation and translation.