Research On Parameter Design And Gait Control Algorithm Of Biped Robot

Robotics as an interdisciplinary research area brings together the findings of related fields such as machinery,electronics,automation control,and computers.At present it is at the cutting edge of scientific and technological research in the world.Compared with traditional wheeled robots,biped robots with their kinematic structures are able to adapt to human life and working environments and coexist with humans.However,in order to use biped robots to assist humans in everyday life,there are still technical challenges to be addressed,including poor mobility and slow walking speeds caused by insufficient robot performance,as well as poor anti-interference,unstable walking at high speeds,little improvement in walking speeds incurred by non-robust path planning and control methods.The above problems have restricted the promotion and applications of biped robots.Relative to foreign countries like the United States and Japan that take the lead in the research and development of biped walking robots,our country is lagging behind with regards to robot prototype design,motion system construction,gait trajectory generation and optimized control.Therefore,in this paper,with an aim to prototype a high-performance biped robot that can achieve stable walking and demonstrate the capability to resist disturbances,and to further narrow the gap with foreign robot research level,the author explores how to improve the motion ability,gait control,and gait stability of biped robots to gain continuous improvement in relevant parameter design methods and control theory systems.Given the above problems,in order to improve the motion ability and walking stability of biped robots,this paper conducts research as follows.Firstly,as the biped robot switches between a fully actuated mode and an underactuated mode in complex walking environments,traditional modeling methods are only able to maintain stability when the soles of the feet come into complete contact with the ground.In response to this problem,first of all,this paper analyzes the dynamic model of the biped robot,examines the traditional seven-link dynamic model,and introduces the floating base method to the dynamic modeling of the biped robot,with the purpose of maintaining body balance even when the soles of the feet slide relative to the ground,and thus heightening gait stability.Meanwhile,based on the foot rotation indicator(FRI)point under the foot rotation state,this paper expands the scope of application of the traditional zero-moment point(ZMP).Then combined with influencing factors like dynamic analysis,walking stability criterion and walking environments of the biped robot,this paper gives a variety of constraints to ensure the stable robotic walking as the basis for gait trajectory planning.Finally,based on dynamics,the falling process is analyzed,and the damage to the robot when it falls over to the ground is reduced through force control.Secondly,concerning the low motion ability of existing biped humanoid robots,the multi-objective parameter optimization method is utilized to design a high-performance biped robot prototype.First,by analyzing the design requirements of the robot,the key parameters of the robot prototype are proposed.The parameter design process can be viewed as a multi-objective optimization problem.Then,the functional relationship between walking speed,walking stability and performance requirements during the walking process of the robot is established through the optimization analysis of the walking of the biped robot.Taking walking speed,stability and performance requirements as the optimization goals,the optimization objective function is presented.A method to obtain the optimal set of target parameters through nonlinear multi-objective optimization is proposed,and the target parameter set is obtained through this method.Further,in combination with gait simulation based on physics engines,the degree of freedom,structural design,motion performance and walking effect of the robot are evaluated,and the effectiveness of the parameter design in the whole machine is verified.At the same time,structure optimization is carried out by reasonably selecting the robot leg driving mechanism,so as to obtain better driving efficiency of the robot joint under the same quality.Last but not least,the parameter design of the biped robot prototype is completed.Compared with foreign robots of the same specification,the robot prototype has a lighter weight and a larger output torque.Thirdly,in order to solve the problems of low trajectory tracking accuracy and poor motion robustness in traditional robot control methods,this paper proposes a robot motion trajectory control framework based on the full-body dynamics model and quadratic programming.First,the biped robot gait is planned grounded upon the three-dimensional inverted pendulum model,and the trajectory planning method in landing phase and swing phase is described.Aiming at the problem of poor tracking accuracy of the centroid trajectory of the existing model,the DCM tracking control algorithm is introduced for feedback control of the centroid trajectory,and the tracking accuracy of the DCM algorithm is analyzed.Then,in view of the poor robot motion robustness,the friction cone model is used to describe the contact state between the foot and the ground,solve the plantar contact force and the contact force vector at each contact point,and put forward a joint force control trajectory tracking method based on inverse dynamics.Finally,an optimal controller based on PD control and quadratic programming is designed to realize joint force and plantar force control and improve the robustness of motion trajectory tracking.The experiments have shown that the above-mentioned control framework can realize more compliant robot motion and stronger robustness compared to the existing position control methods.Forthly,in order to further enhance the robustness of the gait algorithm,the antiinterference methods for internal and external disturbances during the walking process are studied respectively.To cope with internal disturbances in the walking process,firstly,the influence of angular momentum is reduced from smooth connection of the biped robot walking.Then,by analyzing the simple particle model and the whole-body angular momentum model,an optimized trajectory generation scheme based on wholebody angular momentum is proposed,and the scheme is used to generate a reference CMP trajectory.According to the reference CMP trajectory,the DCM trajectory is solved and a centroid trajectory solving algorithm is raised,which optimizes the walking trajectories.Experiments show that,compared to the simple particle model,the trajectory optimization method based on the whole-body angular momentum model improves the robot's ability to resist internal disturbances.For external disturbances in the walking process,the method for detecting pushes and the method and restriction of maintaining balance and recovering from disturbance are studied.Based on the orbital energy analysis,a method to solve the position of the dynamic capture point is proposed,which finds out the robot's stance when being pushed,constructs a floating base stability anti-push strategy after being pushed,and controls the robot to keep balance after being disturbed in the walking state.Taken together,this paper focuses on improving the performance and motion robustness of the biped robot,with the purpose of designing a high-performance biped robot prototype at the theoretical level.The robot prototype,based on a complete theoretical model and an optimal trajectory control framework,can improve the robustness of robot gait walking both theoretically and methodically,and therefore,to some extent,can promote the research and development of humanoid robots in our country.