The Feasibility Research To The Combination Of FastSLAM2.0 And PGR

As the science and technology developing, the robot research has been rapidly improved. Robot has come into every side of our lives. In all of the research of robot, the mobile robot is a research focus in recent years. Intelligent mobile robot in unknown environment has self-organization planning, adaptive ability of a robot. Mobile robot localization and navigation study is the basis of these researches. Navigation algorithm is a popular field of robotics. This article is focused on the (Simultaneous Localization and Mapping, SLAM) research, and to combine the FastSLAM2.0 and PGR, discuss the feasibility, and then prove it by simulation with Scilab software.Firstly, the article discusses the SLAM algorithm, FastSLAM2.0 algorithm, particle filters, extended Kalman filter EKF, and then compare them with each other. Highlighting the obvious advantages of FastSLAM2.0 algorithm in convergence, convergence rate, excellent feature of simultaneous observations treatment. FastSLAM 2.0 algorithm has solved that the robot which has not any information of environmental map, and doesn't know its pose information can move to the aim pot independently.Then we study the PGR (Path-generating regulator) navigation rules. It has mainly been used for which is a feedback controller for navigation of two-wheeled non-holonomic mobile robots. Because the non-holonomic mobile robots have non-integral constraint conditions, it is difficult to use a control law to make the mobile robot converge to the target state. PGR (Path-generating regulator) navigation method designs a nonlinear regulator carrying out asymptotic convergence of non-holonomic mobile robots to a given trajectory family. As a result, the mobile robots are convergent to the origin as the target state while generating their path.To achieve better performance of the path-generating regulator, it is necessary to obtain the position and orientation of the robot within certain accuracy. The FastSLAM2.0 can estimate the position and landmarks in the environment and the position and orientation of the robot. So it can be applied to the path-generating regulator. In the data simulation process, firstly Scilab software is used to make the input and sensor data collection programs. And then compile the programs of FastSLAM2.0 algorithms and PGR navigation rules. According to the robot control process, the simulation is been worked, and the excellent simulation results has come out. The result has proved the feasibility to the combination of FastSLAM2.0 and PGR. Meanwhile, we design an improved mobile robot laboratory EMC-700, and it provides a very good platform and hardware infrastructure for the research in future.