基于遗传算法的双足机器人足踝蹬地参数优化

Parameters Optimization of Ankle Push-off of Planar Bipedal Robot Based on Genetic Algorithm

为提高双足机器人步态的经济性,研究了足踝蹬地扭矩和蹬地时机对行走速度和能耗的影响。基于Matlab/Simulink建立了双足机器人仿真模型并搭建了控制程序,以行走速度为目标函数,利用遗传算法对不同步长下足踝蹬地扭矩和蹬地时机进行优化,以两腿间夹角确定步长。利用无量纲速度和无量纲能耗作为评价双足机器人运动性能指标,仿真结果表明,随着步长由40°增加到60°,机器人行走速度和能耗均随之增加。当步长为60°,蹬地扭矩为41 N·m,蹬地时机为整个步态周期的43.82%时,双足机器人行走速度最大为0.48,对应能耗为2.97。以速度能耗比作为评价机器人步态经济性指标,仿真结果表明,双足机器人在步长50°,蹬地扭矩为35 N·m,蹬地时机为步态周期的45.18%时,机器人步态的经济性更高,对应速度和能耗分别为0.43和2.26。机器人足底地反力出现3个波峰,当步长为50°和60°时,随着足踝蹬地阶段的出现,足底地反力出现第3个波峰,峰值分别为245.45 N和281.23 N。本研究可为双足机器人在特定步长下选取合适的蹬地参数提供重要参考依据。

Ankle push-off is defined as the generation stage of burst of positive power by muscle-tendon units about the ankle joint during the step-to-step transition in human walking.It primarily contributes to both leg swing and to center of mass(COM)acceleration during human walking.However,the influence of ankle push-off on the walking speed and mechanical cost of transport of biped robots has been paid less attention.In order to improve the motion economy of the biped robot,the influence of the amount and timing of ankle push-off on the walking speed was studied.The simulation model of planar biped robot was established in Matlab/Simulink and the corresponding control program was built.In trajectory planning module,the quintic polynomial function was used to generate the motion trajectory of hip and knee joints during stance and swing phases.The PD controller was used for both hip and knee joints of the simulated robot.The application process of ankle torque can be divided into four stages:early stance stage,push-off stage,early swing stage and late swing stage.The push-off height was corresponding to the push-off timing.When the height of the ankle joint of the leading leg was less than the push-off height and the hip joint angle was within the range of-0.6θ^-0.4θ,the ankle joint of the trailing leg entered the push-off stage and the amount of ankle torque was the push-off torque.When the ankle angle was greater than 20°,the push-off phase ended and the trailing leg would start to swing.In order to prevent the foot of swing leg from scuffing with the ground in the early swing stage,the ankle joint was controlled within a certain angle range using the state machine of ankle joint torque.The ankle joint was in passive mode in the late swing stage and the early stance stage.With walking speed as the objective function,genetic algorithm was used to optimize the amount and timing of ankle push-off under different step lengths.Taking the dimensionless speed and dimensionless mechanical cost of transport as indicators to evaluate the performance of biped robots,the results showed that as the step length was increased from 40°to 60°,the walking speed and mechanical cost of transport of the robot was increased accordingly.When the step length was 60°,the torque of ankle push-off was 41 N·m,and the push-off timing was 43.82%of the one gait cycle,the maximum walking speed of the simulated robot was 0.48,and the corresponding mechanical cost of transport was 2.97.Taking the ratio of speed to mechanical cost of transport as an index to evaluate the economy of robot gait,the results showed that the simulation robot obtained the economy gait when the step length was 50°,the torque of push-off was 35 N·m,and the push-off timing was 45.18%of the gait period.The corresponding speed and mechanical cost of transport were 0.43 and 2.26,respectively.In addition,the ground reaction forces of the robot had three wave peaks.When the step length was 50°and 60°,with the appearance of the ankle push-off stage(43%~60%),the ground reaction force appeared the third wave peak and the peak values were 245.45 N and 281.23 N,respectively.The results provided a reference for the biped robot to select appropriate ankle push-off parameters under a specific step length.