Research On Gait Planning And Gait Control For Biped Robot

Biped robot is a naturally unstable system, the researchers need to optimize the stability of pre-designed gait curves, and then also need precise control to ensure stable walking. Over the years, many scholars put forward a series of models and algorithms to solve these two problems, and made a number of landmark results. However, with the more and more complex mechanical structure and larger and larger degrees of freedom of the robots, gait planning and controlling become increasingly difficult. Therefore, the study of the gait planning and controlling of biped robot is very important. This thesis is mainly about three aspects base on the tow issues above, which are gait planning based on the Linear Inverted Pendulum, gait planning based on the Augmented Linear Inverted Pendulum and the fuzzy on-line compensation based on Zero Moment Point (ZMP) and Center of Pressure (COP) on ground.First of all A Linear Inverted Pendulum Model is introduced in this paper, which is an ideal model proposed by Japanese scholars. With its dynamics equation Gait of biped robot can be planned out. The principles of Linear Inverted Pendulum Model and steps of bipedal gait planning are described in detail in this paper. The data of the gait planned with the above methods is verified through simulations.Some basic principles of Augmented Linear Inverted Pendulum are described in the second part. Some improvements on fitness equation have been made based on the prior achievement. A more suitable off-line gait planning method for Augmented Linear Inverted Pendulum is proposed, and with which a better bipedal gait planning has been done. At the same time, some simulation have been made with virtual prototype used in the first part, and the stability, continuity and humanity of the above two gaits have been analyzed to obtain the principles of making gaits planning in practical applications with Augmented Linear Inverted Pendulum and Linear Inverted Pendulum. As all these done we can get the points to ensure the feasibility and security of gait data in the practical application.In the third part of the paper,a new method is used for online control of gait,aiming to ensure good anti-interference and adaptability of biped robot to the environment.At first15 sensors are used to measure real-time robot ZMP and COP respectively to get the ZMP error and the COP distance with the given ZMP and COP. Then the errors are fuzzified and a fuzzy controller is applied to adjust the robot’s hip and knee angle. This method is demonstrated to have a good feasibility through the simulation tests and analysis.