Research On Gait Planning And Control Of Bipedal Walking Robot

This paper describes mainly the study of controlling a humanoid biped walking robot.The core of the paper is the algorithm of gait planning and control for the humanoid biped humanoid robot.Based on the theory of robot research at home and abroad,this paper summarizes and innovates the theory of a humanoid biped walking robot.The paper researches on kinematic model,the gait planning algorithm of the three-dimensional inverted pendulum,vibration control of robotic legs,attitude control of the robot and sensing detection.The main tasks and fruits of this paper are as follows:First,the exponential forward kinematic model of robotic legs is established by introducing Lie Group and screw theory in this paper.Combining the exponential product formula of the forward kinematics model,a method is put forward to obtain the evaluation of joints’ angle by two steps.Step 1 evaluates three angels of the knee and ankle by the Geometric Method.Step 2 finds the evaluation of three angels of the hip through the matrix analysis.Similarly,the paper establishes the kinematic model of robotic arms and head parallel mechanism,which provides the theoretical basis for movement control of the robot’s body.Next,through research on the mathematical model of 3D inverted pendulum,the paper provides a concise method of parametric expression concerning notion law of 3D inverted pendulum,namely,a representation of the hyperbolic function.What’s more,the paper proposes a predictive control algorithm of gait placement after research on the walking algorithm of the forward,left and right direction,based on motion law of hyperbolic sine and hyperbolic cosine.Aimed at the study of the gait planning algorithm of the robot’s starting,stopping and moving,the paper presents the algorithm of limitless differential tracer and the algorithm of fastest motion planning.The two algorithms,not only make the trajectory smooth,but also get the target position and speed.Furthermore,in order to effectively control the leg vibration,the paper adopts the algorithm of auto-disturbance-rejection(ADR).And then the extended observer is improved by introducing the model of the system,which greatly enlarges the range of using the ADR algorithm.And the study effectively solves the conundrum like observing the signal of angle and angular velocity by different of physics feedback signal.Therefore,the actual experiment greatly improves the stability and robustness of the system.In addition,utilizing the angle deviation of robot’s posture,the robot is able to keep balance on a slope by the algorithm of balance control after controlling the angle of robot’s attitude through the PI.For the accuracy of the attitude angle,a new complementary filtering fusion method is proposed.The method not only takes the gravitational acceleration direction as a reference,but also makes full use of the magnitude of the gravitational acceleration.The method thus reduces the influence of the acceleration on the angle calculation.The robot acquires the information of ground traffic by the laser radar.In order to get an exact linear equation from the laser radar’s data,the research adopts a fitting method based on the shortest distance between a point and a straight line.The vector is used to represent the straight line,and the fitting vector is calculated by calculating the minimum eigenvalue of the data point covariance matrix.The eigenvector corresponding to the eigenvalue is the direction vector of the straight line,and the method is very easy to be generalized to any N(N> 1)linear fitting in space and hyperplane fitting in N-dimensional space.Finally,experiments of the robot gait planning and control are carried out.For experiment and debugging,the debugging machine is established based on the Labview and Matlab,and then all the hardware data and working status of the robot are monitored and used to debugging computer programs before the experiment.The experiment verifies validity and correctness of the kinematics model,the gait planning algorithm and the algorithm of the robot control.