Balance Mechanism In Unstable Conditions And Gait Planning Of Humanoid Robots

Compared with other mobile robot systems(such as the wheeled or flying robots),the bipedal motion of the humanoid robot is a classical hybrid motion system.It also has many degrees of freedom and strong nonlinear dynamics.The stability and flexibility of humanoid robots have always been a difficulty in the field of humanoid robots.First of all,the fact that the humanoid robots can only have unilateral constraints on the environment leads to difficulty in balancing them.Moreover,associated timing and the coordination control of redundant joints are other important topics in controlling biped robots.Beyond these,the main challenges are being human-like,energy-efficient,versatile and of course agile like humans.Nowadays,the biped robot’s locomotion is still far from the level of the human’s.In view of these difficulties,the main research contents of this paper are as follows:1.Research on the contact model between the legged robot and the ground.The theoretical derivation and the implementation of a foot-ground contact dynamics system for legged robot physical simulation are presented.The physical simulation imposes high requirements on the contact dynamics for fidelity and accuracy.Thus,we explored the contact mechanism and developed a contact model.The ground reaction force(GRF)by the kinetic equation with constraints is theoretically analyzed first.Considering the limitation of the numerical analysis method,we propose a new contact model based on Hertz theory and clarify the fundamental differences between the model and reality.Due to these differences,we optimized some key parameters by the genetic algorithm.Then,we expanded the contact model to nonhorizontal and nonstationary surface contact.At last,to present the usefulness of this contact model and analyzed the dynamic contact behavior,several cases are simulated.The contact force based on this model highly accords with the results of theoretical analysis and includes all phenomena that may occur during a locomotion cycle.2.Momentum-based balance strategies for the robot under external disturbances.This paper regulates the combination of linear and angular momenta and proposes two balance strategies: Fixed-Support Strategy and Change-of-Support Strategy.The ground reaction force according to the linear and angular momenta is determined firstly.When the value and location of the desired GRF cannot be realized simultaneously,the robot should choose to take steps or not.In the fixed-support strategy,the robot attributes higher priority to linear momentum and starts to rotate the upper body,reminiscent of the standing balancing behaviour of humans.In the change-of support Strategy,the reference reactive strides are computed by modelling the robot as a rimless wheel with two spokes.An effective computational method for real-time calculations by approximating and linearizing the dynamics of the model is proposed.To validate the strategies,some experiments are conducted by simulations.3.A 3D aperiodic gait planning method with the double support phase.A trajectory generation algorithm with which a 3D biped robot can perform aperiodic gaits by modifying only a small set of gait parameters is presented.In addition to the double support phase,the gait can transfer smoothly from one single support phase to another.The sagittal and coronal dynamics are decoupled,firstly.In the sagittal plane,two simplified models are used to generate the walking reference trajectory during the single support phase and double support phase.An extra template model is added to describe the motion in the DSP.For the coronal plane,only the LPM is utilized for trajectory generation.Thanks to linearity properties,the proposed method can obtain the biped locomotion computationally fast without the need for numerical time integration.Herein,the trajectory generation algorithm for different aperiodic gaits including from standing to walking,stopping walking,as well as speed switch is proposed.A full-dynamics 3D humanoid robot is used for the tests of the developed reference trajectory through simulation.The results are promising for implementation.4.A high anthropomorphic gait planning method.A trajectory generation algorithm with which a 3D biped robot can perform biped locomotion with the variable center of mass height is presented.Herein,a new abstract model is proposed and the NMPC is used to generate the Co M trajectory,torso rotational angle and adaptive footholds.The new model fully considers the robot’s movement and rotation in a total of six dimensions as well as the footholds.As a result,a QCQP with goal function,feasible constraints,etc,is formed.Thanks to the SQP,the QCQP can be solved to generate various biped gaits in the complex environment.A fulldynamics 3D humanoid robot is conducted for the tests of the proposed method in steering and walking underneath a low door through simulation.The results are promising for implementation and show walking with variable Co M height can save about 15% energy per meter.5.Exploration of the effect of the arm swing motion on biped locomotion.The upper body motion is often neglected in biped locomotion.An effective trajectory planning approach for biped walking with arms is proposed.Another issue addressed in this paper is the effect of arms on biped locomotion,including the mass and length characteristics on energy consumption and stability.Herein,the humanoid robot is modeled as a 5-link or 7-link robot depending on with the arms or not in 2D.Virtual constraints are applied to reduce the dimension of dynamics,and the coefficients of the virtual constraints are optimized by genetic algorithm and SQP.The optimal walking shows that the arms swing motion expands the region of attraction of the nonlinear locomotion system and obtains a faster gait.