Yantao Tian

Multi mobile robot path planning based on genetic algorithm

Genetic Algorithm has implicit parallelism and global optimization, however its computation speed constrains its on-line application. This paper attempts to apply improved genetic algorithm to multi mobile robot path planning. By using based-knowledge genetic operators, the performance of genetic algorithm is improved greatly. Simulation results showed that the improved genetic algorithm can satisfy real-time demand. In the future, it is used on-line to replan the multi mobile robot path.

Motion Control of Ball and Plate System Using Supervisory Fuzzy Controller

Ball and plate system is a typical under-actuated system with inherent nonlinearity. Motion control of the system included the set-point problem and the tracking problem along desired trajectory. A supervisory fuzzy controller was proposed for studying these problems, which was composed of two layers. The low layer defined a SIRM fuzzy rule and an importance degree for each input item, and the high layer performed supervisory task for balls velocity and acceleration. When the ball was close to halt, or at high velocity, the supervisory fuzzy controller can overcome disturbance. We implemented this controller to ball and plate system setup in authors laboratory. The experiment results showed little steady-state error and good transient performance than conventional fuzzy controller. Therefore, the control scheme is proposed feasible, and acquired conclusion can be used to investigate similar nonlinear under-actuated dynamics system

Fault diagnosis of nonlinear system based on generalized observer

The fault model of the nonlinear system and the generalized nonlinear observer are presented in this paper. The algorithm compensates the influence of faults on the remnant order by means of adjusting the gain matrix of the observer to track the fault parameters of the system on line. It is proved that the nonlinear observer can realize fault diagnosis. In fact, here the problem of fault diagnosis is converted to the problem of model matching. The fault estimation can be exactly and timely. At last, experiments of three-tank system illustrate the effectiveness of the proposed algorithm.

Successful treatment of giant condyloma acuminatum with combination retinoid and interferon-γ therapy.

Giant condyloma acuminatum, or Buschke–Löwenstein tumour, is a unique variant of anogenital warts. It is characterized by locally aggressive behaviour but rarely metastasizes. Many treatment strategies such as radical surgery, radiation and chemotherapy have been used to treat it but their efficacy is often poor and the recurrence rates are high. We report a case of 16-year-old girl who was treated with oral retinoid combined with intramuscular interferon-γ. All lesions cleared within three months. During a follow-up period of more than two years, no recurrence has developed. This relatively painless, non-scarring treatment may represent a novel therapeutic option.

Algorithms of task-allocation and cooperation in multi mobile robot system

Until now, many methods are used to solve the problems of task-allocation and path planning in multi mobile robot system. There are two new algorithms presented in this paper. The task-allocation algorithm resolves the puzzle of task getting and accomplishing with the least cost based on the optimal notion. The environment of the system is constructed with the topological graphics. Path-planning algorithm programs an optimal path from the initial point to the destination that satisfies some performances based on the idea of depth-first search. And we designed an emulational test bed based on the multi-robot material flow system of the storages and docks. Then we studied the results of the emulational experiments about the presented algorithms.

Multi-objective path planning for unrestricted mobile

Problem of multi-objective path planning is investigated in this paper for the ball and plate system. The purpose of multi-objective path planning is to obtain the safe and shortest path for the ball to track. Workspace is represented by distance map and hazard map. Weights for multi-objectives are calculated by entropy method for each grid node. Dijkstra algorithm is employed to solve the multi-objective path planning problem finally. As illustrated by simulation results, the path obtained by multi-objective method proposed in this paper is much safer compared with single-objective A∗ algorithm.

Output regulation of the ball and plate system with a nonlinear velocity observer

This paper proposes a nonlinear velocity observer for output regulation of ball and plate system. Ball velocities are estimated by state observer instead of first order difference method. Dynamics of motion resistances are included and estimated in augmented state space. Simulation and experimental results on ball and plate platform have shown that nonlinear state observer provided reliable estimation of velocities. Accuracy of velocity estimation is promoted with state space augmentation than the original system. Regulators with a nonlinear velocity observer do stabilize the ball and plate system.

Gabor feature representation method based on block statistics and its application to facial expression recognition

Facial expression representation based on Gabor features has attracted more and more attention and achieved great success in facial expression recognition for some favorable attributes of Gabor wavelets such as spatial locality and orientation selectivity. A large number of Gabor features are produced with varying parameters in the position, scale and orientation of filters, which cause huge computational complexity. In some existing methods, useful discriminatory information may be lost due to down sampling Gabor features directly. To reduce the loss, a Gabor features representation method based on block statistics is proposed in this paper. In addition, Support Vector Machine is used to match the features. The effect of this method is demonstrated by template matching test on JAFFE database, and the comparative test results show that this method can yield better recognition accuracy with much fewer Gabor features as well as less CPU time of feature matching.

Design of GPS/INS Integrated Navigation System Based on Multisensor Information Fusion

Two kinds of design methods of GPS/INS integrated navigation system are presented, that is GPS/INS integrative system and integrated navigation system based on information fusion. By analyzing mathematics model of the first system error, we found that inertia velocity error would cause code loop track error. Consequently, the pseudo range meterage error is interrelated with inertia velocity error, which is complex to calculate. If this relativity is ignored in the state equation and observation equation of this system, it must affect the estimated precision of the Kalman filter and maybe the system is unstable. Therefore, the second integrated navigation method is introduced. It adopts information fusion, federal Kalman filter and covariance. And consider the above pertinence to analyze the navigation performance of the integration system. Then it gives the flow of federal Kalman filter algorithm. By analyzing, the conclusion is that velocity error of integrated navigation system drops from 0.5m/s to below 0.05m/s, and improves the precision and reliability of the navigation system effectively, well continuity and real-time capability. It provides an effective way for data analysis and process of fusion navigation system

A new start method for rolling piston compressor based on DSP to improve the load performance

Starting performance of refrigeration compressors is very important to compressors. The research on starting performance of rolling piston compressors is the research on starting performance of motors actually. An improved starting sheme was proposed in this paper. Simulation and experiment results indicate that the reliability and stability of the starting performance of compressor are improved.