Study On The Two Wheeled Self-blancing Robot With Multi-DOF

Two wheeled self-balancing robot is a novel kind of mobile robot developed recently.Due to the advantages of light in weight, compact structure, good mobility and zero turningradius, this type of robot is very suitable for narrow and crowed environment. Being usedas a convenient and environmentally-friendly robotics mobile platform, two wheeledself-balancing robot has extensive application prospect and practical value. In addition, twowheeled self-balancing is a complex under-actuated and nonlinear system, so the study oftwo wheeled self-balancing robot can provide valuable references for the control of othernonlinear systems, such as aircraft, space vehicle, precise instrument and robotics.When the traditional two wheeled self-balancing robot is used as a transportationvehicle, the long time standing in a long trip may cause the fatigue of the user. Thetraditional two wheeled self-balancing robot can not keep upright when climbs or descendson a slope, as well as in acceleration or deceleration. This drawback would make theoperator sitting on the robot feel unsafe and uncomfortable. When the traditional twowheeled self-balancing robot transport liquid and precise instruments required keepingupright, this weakness is likely to cause accident. Moreover, when the two wheeledself-balancing robot turns sharply, there is a great risk of overturning due to the centrifugalforce. To solve these problems, a multi-DOF (multi-degree-of-freedom) two wheeledinverted pendulum robot is developed in this paper, and the research in this paper focuseson improving the stability, mobility and safety of the two wheeled self-balancing robot. Themain work of this paper includes:1) To configure the center of the gravity, a multi-DOF two wheeled inverted pendulumrobot is developed and the construction method of the mechanical system and the controlsystem are introduced with details. This robot has two DOFs (degrees of freedom), theDOF of the center of gravity moving forward and backward (CGFB-DOF) and the swingDOF. By the two DOFs, the performance of the two wheeled self-balancing robot areimproved.2) To meet the requirements of precise pose measurement, the processing and fusion of the sensor signal is studied. Based on the principle of pose measurement by using the gyroand the accelerometer, the influence of the sensor installed location is analyzed in detail andthe measurement result is modified accordingly. The Kalman filter technology is adopted tofuse the signal from the gyro and the accelerometer. The synthesized method proposed inthis paper can obtain the precise robot pose information with a very low cost.3) The kinetics and dynamics model of the robot is built via analyzing the motioncharacteristics of the multi-DOF two wheeled self-balancing robot. The dynamics model isdecoupled as the dynamics of the straight line motion and the dynamics of the turningmotion. Furthermore, the dynamics of the straight line motion is simplified under certainconditions. Based on that, the dynamics complexity is reduced and the controller design isfacilitated. On the foundation of the dynamics model, the steady-state characteristics of therobot motions are analyzed.4) The transmission mechanism friction model is built based on the identification ofthe friction parameters. Based on the friction method, the transmission mechanism frictionwas compensated in the output of the balancing controller, which is a robust controllerdesigned by utilizing the sliding mode control algorithm. Then, the control strategies areproposed on the straight line motion and the turning motion. On the premise of theimplementation of the balancing control, the control of the straight line motion is realizedby moving the slider dynamically. In the respect of the turning motion control, a turningcontroller is designed and the turning stability is improved by swinging its body when therobot turns. The dynamics simulations conducted in the integrated environment of ADAMSand MATLAB show that the controllers are effective.5) The physical experiments are carried on with the multi-DOF two wheeledself-balancing robot platform, such as balancing experiment, running experiment on the flatground and on the slope, and the turning experiment. These experiments havesystematically verified the robot performance of self-balancing, mobility, climbing, steering,and the improvement due to the extra DOFs. In addition, these experiments have validatedthe effectiveness of the dynamics model and the control method proposed in this paper.