Zhang Minglu

Double inverted pendulum system control strategy based on fuzzy genetic algorithm

The double inverted pendulum is a MISO complication system. If only adopting the traditional fuzzy control, then the total of fuzzy rules will bring the exponential increase with the increase of input variables, which is disadvantaged to the real-time control. So in this paper, we firstly adopt the add weighting coefficient to reduce the input variables of fuzzy controller, then using Genetic Algorithm to optimize the weighting coefficient and membership function parameters of fuzzy controller by the off-line way, finally, using the optimizing result to control the double inverted pendulum system. In this paper, we introduce the whole optimizing operations with GA, which are coding strategy, determination of the fitness function, selection, crossover and mutation. In the end, by the validation of simulation platform Matlab gives out the curve of simulation and shows the optimizing control strategy is more effective in the aspect of speed response and stability than traditional fuzzy controller.

Methods to Modular Robot Design

Reseach on modular robot is a hotspot in the robot field. The design methods to modular robot are different to traditional robot. By analyzing the domestic and international modular robot reaserch, the 7 kinds of modularizing methods and 3 kinds of configuring methods to modular robot have been summed up. A bottom-up approach about system design of modular robot is presented to achieve the goal of rapid response of the users.

Design of mobile robot teleoperation system based on virtual reality

As a rising subject, virtual reality is the best human-machine interface technology so far. This paper submits the architectural structure of the robot teleoperation system based on virtual reality and after analyzing the principle of the systems operation, it sets up the model of hierarchical control. Subsequently this paper elaborates critical techniques such as telerobot, creating virtual environment, designing video supervising system and so on. Bring technology of virtual reality in robot teleopration system, not only strengthens the working ability of the actuator but also widens application areas of the device. Manipulator can safely and cozily completes tasks by teleoperating the robot.

Function-based modular robot system

A method making traditional robots into modular robots was introduced. It can be applied to a robot or a couple of robots with similar function. Through the functional decomposition for non-modular robots, identify different modules of the system, so make the system design of modular robot come true. This method has been applied to three types of representative special robots: attack robot, explosives disposal robot, fire-fighting robot. By function structure analysis, get the sharing modules and personal modules. The IDEF0 modeling method has been used to clearly and explicitly define the activities and relations of each modular in the system from the functional point of view, so establish standardized model of modular robot architecture.

Study of Dynamic Transient Stability of a 2-Link Wheeled-Suspended Mobile Flexible Manipulator

This work studies the dynamic transient stability of a 2-link wheeled-suspended mobile flexible manipulator. The dynamic transient stability criterion is constructed based on the reacting forces and skidding resistance forces existing between the tires and the road surface. At last, Computation simulation results are presented to illustrate the study. This criterion is feasible to evaluate the transient dynamic stability and perform mutil-goal optimization of a mobile flexible manipulator with respect to a specified performance.

Topological structure design of cleaning manipulator for aircraft surfaces based on single-open-chain

Based on topological structure design theory and method of Position and Orientation Characteristics, the thesis analyzed the topological characteristics of serial robot with six degrees of freedom and synthesized the single-open-chain mechanisms containing five revolute joints and one prismatic joints. And then two mechanism which summarizes two topological structures meeting 5R1P optimized index has been designed. Theoretical innovative approaches of single-open-chain mechanism for application in aircraft cleaning are put forward.

Modular robot system configuration design based on axiomatic design

Axiomatic Design is applied to the modular special robot configuration system. For a specific design mission, we refine the design objective and the design restraints, and then complete decomposing of FRs and DPs between the functional and physical domains through zigzagging. By adjusting design matrix, configuration design order can be overall planned. By using the design parameters, the corresponding module can be found from the module set easily, and the candidates can be defined for modular configuration design activity.

The study on dynamic scheduling of mobile manipulator based on MAS and neural network

A distributed intelligent decision and management system is erected which incorporates the technologies of decision support system, expert system and database based on the theory of multi-agent. The system is able to support more rational, more effective and more scientific decision-making and schedule strategy, and it also can do attribute to the automation and intelligence of mobile manipulator. Also, in this paper, we use Neural network as the intelligent controller of the hierarchical intelligent system, because it is near to the method of person dealing with information and has the ability of imitating persons image thinking. This paper utilizes the excellent characteristics of neural network in studying and fault-tolerant to solve the coupling between manipulator and mobile vehicle and compensating the torque in order to realize the coordination movements between manipulator and mobile vehicle of mobile manipulator.

Pattern recognition and following of wall by behavior-based mobile robot

The movement of mobile robot along wall is classified into wall type recognition, straight wall following and turning around corners according to the behavior based idea. Mobile robot measures distance between its body and the wall it follows by three ultrasonic sensors, it recognizes the wall type online with neural network, follows straight wall with fuzzy controller and turns around 3 kinds of corners at their optimal turning radius by calculating models for the corners. The results of simulation and experiments according to TIT-I mobile robot indicate that the control strategy for wall recognition and following proposed in this paper is practicable and effective.