Objective Quadruped robots gain widespread attention in hazardous and complex tasks such as search and rescue, mine clearance, and steep slope climbing due to their superior terrain adaptability compared to wheeled and tracked robots. However, existing trajectory planning algorithms predominantly focus on slip control and kinematic optimization, often neglecting the biological motion characteristics and walking stability observed in natural quadruped mammals. After millions of years of natural selection, quadruped mammals evolve optimal locomotion patterns, such as leg-lifting backswing and leg-lowering retraction, which minimize ground impact and ensure motion continuity. This study aims to develop a zero-impact foot-end trajectory that integrates biological motion mechanisms to address the limitations of traditional trajectories, including insufficient bionic adaptability, abrupt acceleration at leg lifting and lowering instants, and poor stability in complex environments. The goal is to enhance walking stability, reduce ground impact, and improve the bionic performance of quadruped robots, providing a reliable solution for their practical application in complex terrains. Methods Firstly, a bionic quadruped robot was designed with reference to the physical characteristics of German Shepherds, a breed known for robust muscles, strong bones, and agile movements. Key structural parameters were determined based on standard German Shepherd dimensions, including a body length of 790 mm, a body width of 440 mm, a body height of 418 mm, thigh and calf lengths of 200 mm each, and an overall weight of 37.5 kg. The robot adopted a front-elbow and rear-knee joint configuration, with four legs symmetrically arranged and motion transmitted through linkages. A two-stage leg structure was simplified for kinematic analysis, and a coordinate system was established at the hip joint to derive the inverse kinematic models of the hip and knee joints. Using geometric relationships and the sine theorem, mapping equations between foot-end position coordinates and joint driving angles were deduced, which laid the foundation for trajectory tracking control. Secondly, a quintic polynomial zero-impact foot-end trajectory was proposed by mimicking the biological motion characteristics of quadruped mammals. The trajectory was divided into a swing phase and a support phase, with separate planning for the x (forward) and y (vertical) directions. In the x-direction, the swing phase was divided into three segments: backswing (0 to T/8), forward movement (T/8 to 3T/8), and retraction (3T/8 to T/2), where T represented the gait cycle (1 s). This design replicated natural mammalian locomotion and reduced impact during leg lifting and lowering. In the y-direction, the swing phase included a lifting segment (0 to T/4) and a lowering segment (T/4 to T/2), which ensured smooth vertical motion. Boundary conditions were defined to achieve zero velocity and acceleration at the beginning and end of the swing phase, eliminating impact between the foot-end and the ground. For comparison, traditional cubic polynomial and composite cycloid trajectories were selected as benchmarks, with identical gait parameters, including a step length of 50 mm and a lifting height of 20 mm, applied to all three trajectories. Finally, comprehensive simulations and physical experiments were conducted to validate the proposed trajectory. ADAMS and Simulink were used for co-simulation: the mechanical model of the robot was developed in ADAMS, while the control system, including inverse kinematic solutions and trajectory generation, was constructed in Simulink. The simulations analyzed joint angles, angular velocities, and acceleration curves for the three trajectories. Physical experiments were performed on the prototype robot equipped with a ZMC432 motion controller. A laser tracker (API R-20 Radian) was utilized to capture the actual foot-end trajectory, and trot gait control experiments were conducted to evaluate walking stability, slip resistance, and motion continuity. The performance of the three trajectories was compared in terms of bionic characteristics, impact reduction, and stability. Results and Discussions Kinematic analysis verified the correctness of the inverse kinematic models, and smooth and continuous joint angle curves were obtained for all three trajectories. Simulation results showed that the quintic polynomial trajectory exhibited distinct bionic characteristics. In the x-direction, the backswing (15 mm) and retraction (15 mm) segments were clearly observed, whereas the cubic polynomial and composite cycloid trajectories lacked these biological features. Velocity and acceleration curves in both x and y directions confirmed that the quintic polynomial trajectory achieved zero impact at the instants of leg lifting and lowering, with no abrupt changes. In contrast, the cubic polynomial trajectory exhibited constant acceleration slopes that led to potential impact. The composite cycloid trajectory also achieved zero velocity and acceleration at key points but failed to reproduce the backswing and retraction behaviors. Physical experiment results demonstrated that the prototype robot stably walked using the proposed quintic polynomial trajectory. The forward displacement over 10 seconds reached approximately 500 mm, which was consistent with the theoretical speed of 50 mm/s. Laser tracking data indicated that the actual foot-end trajectory closely matched the planned trajectory, with a maximum position deviation of less than 5 mm. This deviation was primarily caused by assembly gaps, frictional contact between links, and slight motor torque insufficiency. Compared to the cubic polynomial and composite cycloid trajectories, the quintic polynomial trajectory significantly reduced slip phenomena, with the slip rate reduced by approximately 30% and 15%, respectively. The robot's center of mass remained stable during locomotion, with vertical fluctuations of less than 10 mm, which confirmed improved walking stability. In addition, the bionic backswing and retraction behaviors effectively reduced ground impact, as evidenced by smoother joint torque curves and lower operational noise. Conclusions This study successfully integrates biological motion characteristics into quadruped robot trajectory planning, addressing the limitations of traditional trajectories in terms of bionic adaptability and impact control. The proposed quintic polynomial zero-impact foot-end trajectory, which mimics the leg-lifting backswing and leg-lowering retraction of quadruped mammals, achieves smooth motion with zero impact at critical instants. The designed bionic quadruped robot, developed based on German Shepherd parameters, validates the effectiveness of the trajectory through rigorous simulations and physical experiments. The results demonstrate that the proposed trajectory outperforms traditional cubic polynomial and composite cycloid trajectories in terms of stability, slip resistance, and bionic performance. This research provides a solid theoretical and technical foundation for the development of high-performance quadruped robots and raises their practical application in complex terrains, such as mountainous areas, disaster sites, and unstructured environments. Future work will focus on optimizing the trajectory for dynamic locomotion, such as trotting and galloping, and on enhancing adaptability to variable terrains through real-time sensor feedback and adaptive control algorithms.
KangRu, MengFei, ChenXuechao,et al.Structural design and crawling pattern generator of a planar quadruped robot for high-payload locomotion[J].Sensors,2020,20(22):6543. doi:10.3390/s20226543
[2]
ChenJiupeng, HongjunSan, WuXing,et al.Structural design and gait research of a new bionic quadruped robot[J].Proceedings of the Institution of Mechanical Engineers,Part B:Journal of Engineering Manufacture,2022,236(14):1912‒1922. doi:10.1177/0954405421995663
[3]
ShiHaojie, ZhouBo, ZengHongsheng,et al.Reinforcement learning with evolutionary trajectory generator:A general approach for quadrupedal locomotion[J].IEEE Robotics and Automation Letters,2022,7(2):3085‒3092. doi:10.1109/lra.2022.3145495
[4]
MengXin, LiuWenfei, TangLeijie,et al.Trot gait stability control of small quadruped robot based on MPC and ZMP methods[J].Processes,2023,11(1):252. doi:10.3390/pr11010252
[5]
ZhangYihui, HuCong, QiuBinbin,et al.Responsive CPG-based locomotion control for Quadruped robots[C]//Neural Information Processing.Singapore:Springer,2024:279‒290. doi:10.1007/978-981-99-8073-4_22
[6]
XiaoLe, HongjunSan, SunHaijie,et al.Motion control of upward jumping step for SLIP model of legged robot[J].Control Theory & Applications,2024,41(4):631‒637. doi:10.7641/CTA.2023.20591
HuangSenwei, ZhangXiuli.Biologically inspired planning and optimization of foot trajectory of a quadruped robot[C]//Intelligent Robotics and Applications.Cham:Springer,2021:192‒203. doi:10.1007/978-3-030-89092-6_18
[9]
LeiJingtao, WangFeng, YuHuangying,et al.Energy efficiency analysis of quadruped robot with trot gait and combined cycloid foot trajectory[J].Chinese Journal of Mechanical Engineering,2014,27(1):138‒145. doi:10.3901/CJME.2014.01.138
[10]
YangChaofeng.Research on gait planning for a hydraulic quadruped robot[D].Beijing:Beijing Institute of Technology,2015.
[11]
杨超峰.液压四足机器人的步态规划研究[D].北京:北京理工大学,2015.
[12]
WangLipeng, WangJunzheng, WangShoukun,et al.Strategy of foot trajectory generation for hydraulic quadruped robots gait planning[J].Journal of Mechanical Engineering,2013,49(1):39‒44. doi:10.3901/JME.2013.01.039
KimK Y, KwonO, YeonJ S,et al.Elliptic trajectory generation for galloping quadruped robots[C]//Proceedings of the 2006 IEEE International Conference on Robotics and Biomimetics.Kunming:IEEE,2007:103‒108. doi:10.1109/robio.2006.340304
[15]
LuoHongyan.Mechanical characteristics and gait planning of bionic robot dogs[D].Wuhan:Wuhan University of Technology,2018.
[16]
罗红艳.仿生机器狗机构特性与步态规划[D].武汉:武汉理工大学,2018.
[17]
ChenJiupeng, LiChunlei, HongjunSan,et al.Model based gait transition control for quadruped robots[J].Transactions of the Chinese Society for Agricultural Machinery,2024,55(3):431‒440. doi:10.6041/j.issn.1000-1298.2024.03.043
WeiShunxiang, WuHaibo, ChenJiupeng,et al.Optimization and performance analysis of foot compound trajectory of quadruped robot[J].Acta Scientiarum Naturalium Universitatis Sunyatseni,2022,61(4):142‒150. doi:10.13471/j.cnki.acta.snus.2020B149
HuaZisen, RongXuewen, LiYibin,et al.Active compliance control on the hydraulic quadruped robot with passive compliant servo actuator[J].IEEE Access,2019,7:163449‒163460. doi:10.1109/access.2019.2951830
[22]
MaHuishu, LiuYanxia, FangJianjun,et al.Research on bionic foot-end trajectory planning for bionic quadruped robot[J].Journal of Northeast Normal University (Natural Science Edition),2019,51(3):66‒71.
LeiJingtao, WangFeng, YuHuangying.Analysis on trajectory planning and energy consumption of quadruped robot[J].Machine Design & Research,2014,30(1):29‒34. doi:10.3969/j.issn.1006-2343.2014.01.008
ZhangQiushi.Jumping motion control of quadruped robot based on trajectory optimization[D].Qinhuangdao:Yanshan University,2024.
[27]
张秋实.基于轨迹优化的四足机器人跳跃运动控制[D].秦皇岛:燕山大学,2024.
[28]
ChenMingfang, ChenHao, WangXuejun,et al.Design and control of a novel single leg structure of electrically driven quadruped robot[J].Mathematical Problems in Engineering,2020,2020:3943867. doi:10.1155/2020/3943867
[29]
YuanShuaidong, ZhouYijun, LuoChen.Crawling gait planning based on foot trajectory optimization for quadruped robot[C]//Proceedings of the 2019 IEEE International Conference on Mechatronics and Automation(ICMA).Tianjin:IEEE,2019:1490‒1495. doi:10.1109/ICMA.2019.8816477
[30]
ZhouKun, LiChuan, LiChao,et al.Motion planning method for quadruped robots walking on unknown rough terrain[J].Journal of Mechanical Engineering,2020,56(2):210‒219. doi:10.3901/JME.2020.02.210
ChenMingfang, ZhangKaixiang, WangSen,et al.Analysis and optimization of interpolation points for quadruped robots joint trajectory[J].Complexity,2020,2020:3507679. doi:10.1155/2020/3507679
[33]
UsherwoodJ R.An extension to the collisional model of the energetic cost of support qualitatively explains trotting and the trot-canter transition[J].Journal of Experimental Zoology Part A:Ecological and Integrative Physiology,2020,333(1):9‒19. doi:10.1002/jez.2268
[34]
KimK Y, ParkJ H.Ellipse-based leg-trajectory generation for galloping quadruped robots[J].Journal of Mechanical Science and Technology,2008,22(11):2099‒2106. doi:10.1007/s12206-008-0705-1
[35]
LiuYongchao.Research on structure design and foot trajectory planning of imitating cheetah quadruped robots[D].Jinan:Shandong University,2017.
[36]
刘永超.仿猎豹四足机器人结构设计与足端轨迹规划研究[D].济南:山东大学,2017.
[37]
Ben CheikhS, HammadiS.Multi-hop ridematching optimization problem:Intelligent chromosome agent-driven approach[J].Expert Systems with Applications,2016,62:161‒176. doi:10.1016/j.eswa.2016.06.001
[38]
ChenQiang, ShiHuihui, SunMingxuan.Echo state network-based backstepping adaptive iterative learning control for strict-feedback systems:An error-tracking approach[J].IEEE Transactions on Cybernetics,2020,50(7):3009‒3022. doi:10.1109/TCYB.2019.2931877
[39]
NaJing, JingBaorui, HuangYingbo,et al.Unknown system dynamics estimator for motion control of nonlinear robotic systems[J].IEEE Transactions on Industrial Electronics,2020,67(5):3850‒3859. doi:10.1109/TIE.2019.2920604
[40]
ChenJiupeng.Research on gait and motion control of quadruped robot[D].Kunming:Kunming University of Science and Technology,2021.
[41]
陈久朋.四足机器人步态及运动控制研究[D].昆明:昆明理工大学,2021.
[42]
LiHuashi.Biomimetic locomotion control theories and methods of quadruped robot[D].Beijing:Beijing Institute of Technology,2014.