توصيفگر ها :
ربات كروي , ديسك اينرسي , نيوتن-اويلر , لاگرانژ مقيد , كواترنيون , منطق فازي , كنترل گشتاور محاسبه شده , مود لغزشي , شبكه عصبي , سطح شيب دار , وقوع لغزش
چكيده انگليسي :
Modeling and Motion Control of a Spherical Robot Soroush Ghaderi ghaderi24@yahoo com January 26 2020 Department of Mechanical Engineering Isfahan University of Technology Isfahan 84156 83111 IranDegree Master of Science Language Farsi Supervisor Mostafa Ghayour Prof ghayour@cc iut ac ir Advisor Saeed Behbahani Assoc Prof behbahani@cc iut ac ir AbstractIn the last twenty years spherical mobile robots have been noticed Spherical robots due to their specialabilities and their symmetrical shape do not overturn during operation and have better maneuverability alsothe closed cover of the robot protects its internal system from dust and other harmful factors and enables therobot to perform missions in environments where humans or other robots are unable to pass through these places Mathematical modeling is always an important part of robotics therefore at first the kinematics modeling ofthe spherical robot was done with three rotor mechanisms Afterwards the dynamic modeling of the sphericalrobot was performed using two different constrained Lagrange and Newton methods Subsequently Lagrangianand Newtonian dynamics of the robot were examined on an inclined plane and the Lagrangian dynamics of thespherical robot were also analyzed when the robot slipping Some works were done on Newtonian dynamics toprepare the dynamic equations for controlling the system The robot was then controlled using computed torque fuzzy computed torque adaptive fuzzy computed torque sliding mode and fuzzy sliding mode methods and theresults of these five control methods were compared At the end of this section the undetermined dynamics ofthe robot were estimated using neural network and the control was done based on the neural network Finally alinear and circular path were designed for the robot and the robot s motion on the inclined plane was investigatedand the probability of slippage on the inclined plane was also considered The results of the linear path that wasdesigned in this study were verified by a previous research The results of these two studies were compatible witheach other All simulations were done with MATLAB and Simulink software One of the innovations of thisstudy is that the Lagrangian dynamics of spherical robot with three inertial wheels has been thoroughly studiedfor the state of pure rotation furthermore when the robot slips the robot s dynamics on the inclined plane arecompletely derived from two different methods and using fuzzy algorithms the solutions were improved andthe undetermined dynamics of the system has been estimated using adaptive neural network algorithms Theuncertain dynamics were well estimated and the robot was well controlled One of the benefits of a neuralnetwork estimator was that the control of the robot was done well this is mainly because the undetermineddynamics were estimated therefore this process did not required much knowledge about nonlinear dynamics Another topics addressed in this study was the investigation of the spherical robot slipping on an inclined planebecause when the robot moves on an inclined plane and for example when it goes up it may slip Keywordsspherical robot inertia disk Newton Euler Constrained Lagrangian quaternion fuzzy logic computed torquecontrol sliding mode neural network inclined plane slipping
