Biped Robot Geometry Modeling And Motion Planning Study

Biped walking is an extremely perfect and energy- saving walking way in nature. In recent years, biped robot research is developing rapidly, attracting lots of researchers all over the world. Due to the complexity of its technology, and the high research cost, mudular has become the development tendency in this domain. Biped robot walking study of this thesis is based on the PowerCube module produced by German AMTEC Corporation.Firstly, based on the humanity walking pattern, the number of freedom degree are determined which meet the need biped robot can walk completely. The 3D-model of PowerCube module is designed by Pro/E. Then links that are used to link modules are designed at the same time .With the same number of freedom degree and different position distribution, three kinds structure of biped robot are designed, and their advantages and disadvantages are analyzed. Finally, one of them is chosed as this article's robot structure. For mathematical analysis, according to the actual structure of biped robot, the actual structure is simplified reasonably and the mathematical model is obtained. Based on the mathematical model, the forward kinematics and inverse kinematics are analyzed thoroughly, and formulas of forward and inverse kinematics are derived.At the phase of gait planning, the process of biped robot walking are decomposed, then carry on planning on each step. Based on the mathematical model of biped robot, the biped robot is modeled as an inverse pendulum, and the robot walking gait is analyzed by means of the inverse pendulum movement rule. This article takes the robot as a linear inverse pendulum, and solves the planning functions with responsible constraints. Suppose that the position of the ZMP does't change during the movement. Divide the movement into two parts, computed center of mass and swing leg. Derive the trajectory function by inverse pendulum theory, and obtain the planning result.A new method is proposed for biped robot energy optimization with short property of geodesic, preliminary discussion of this method is made as well. Based on the 2D-inverse pendulum, build the energy equation, and take the least energy as target to change the equation into Riemann surface, then derive Riemann measure matrix. On the Riemann surface, find out the variational tendency of control variables and energy during walking.At last, give out the concrete data belonged to the robot of this article, compute planning results with the analysis above, and draw the result curves. Then 3D-model is inducted into ADAMS software to simulate the kimematical behaviour. The kinematical analysis and gait planning are proved by simulation.