Gait Planning And Dynamic Stability Study Of Biped Robot

Humanoid robot stands for the highest research result in the robot field. As a high degree cross discipline it contains mechanics, electronic, computer science, automatic control engineering, artificial intelligence, bionics and other technologies and is one of the hottest research field. Different from the stable supporting of wheel type, pedrail type, multiped type and other kind of robots when they are moving, stabilization of humanoid walking has always been the main technical difficulty of humanoid robot, especially the fast dynamic walking. To solve this problem, reasonable structure and control system of humanoid robot are necessary, besides, gait planning is also the key. Thus, gait planning has become one of the hot topics about humanoid robots. This paper mainly talked about the dynamic gait planning of designed humanoid robot.Firstly, the paper studies motion planning and states path planning and gait planning in details separately. After analyzing static walking and dynamic walking, the paper studies ZMP dynamic stability criterion which is very important to stability control for biped robot walking. At the same time it studies and discusses two main patterns of walking generation and chooses the one based on linear inverted pendulum.Secondly, combined with designed mechanism dimensions and quality and freedom of motion, the paper studies kinematics modeling by decoupling which into forward motion and sidewise motion. It adopts GCM(Generalized Coordinate Method) to solve forward and inverse kinematics, laying foundation for later gait planning. On the base of kinematics modeling, it makes dynamic modeling with Lagrange dynamical equation, which is convenient to gain driving moment needed with known information of joint rotational angle, thus, in terms of the above analysis, electric engines can be selected.Finally, based on ZMP dynamic stability criterion, the paper adopts cubic spline interpolation to realize the gait planning of biped robot and pay more attention to periodic walking phase. It applies improved two-steps planning which makes sure a fluent moving. With some rational assumption and restrict, it parameterizes some key points of ankle and hip in forward motion and sidewise motion to gain a result with major stability margin making sure that the planning has a good stability. In order to test and verify the periodicity and accuracy of gait planning, the rotational angle of every joints and their first derivative and second derivative designed above are needed. It uses the above data to obtain driving moment. At last, it makes a complete offline planning by pulsing initial starting planning and halting planning.