双足机器人因其仿人特性，在家政医疗、工业生产、军事领域均有广阔的应用前景。然而，由于其存在高维、非线性、天然难稳定等特点，其运动控制方法一直是研究的热点。传统的控制方法存在可扩展性差、动态特性低等不足。本文采用状态机的控制框架，针对IP、LIP和SLIP模型下的双足机器人分别设计控制律，实现了稳定行走、跳跃、奔跑等高动态运动，且具有一定的扩展性和鲁棒性。此外，本文还通过实验的方式探究了双足机器人的运动规律，包括行走过程中质心位置和速度的变化规律、不同步态下的关节轨迹以及足端形状对行走性能的影响等，为进一步了解人类和腿足运动的物理规律提供了参考。首先，根据对IP和LIP模型进行开环分析，提出行走轨道和相位图的概念，并绘制了两模型的行走相位图，描述其在行走过程中质心速度的变化关系。在Coppeliasim环境下建立IP和LIP模型下的简化双足机器人模型，并分别设计两模型的行走状态机，实现稳定行走和通过楼梯、斜坡、不连续地面等非结构化地形。通过实验验证了行走相位图，并得出同等条件下，LIP模型具有更高的行走速度。其次，在Webots环境下建立八自由度的仿人机器人模型。通过对人类行走模式的观察，设计了基于行走姿态图的状态机，为不同姿态下机器人设置相应的关节控制律，采用身体质心速度作为反馈，实现了机器人的稳定行走。利用八个线性、非耦合的控制参数，将机器人行走控制参数化，大大简化了机器人的控制难度。同时，实现了直行、转弯、原地踏步等步态，并对不同步态下关节的运动规律进行了分析。另外，本文还利用状态机的可扩展性实现了足端形状改变下的稳定行走，并对足端形状对行走性能的影响进行了分析。最后，在Webots环境下建立了SLIP模型的双足机器人，并设计了SLIP模型的状态机。对前文中的固定状态机进行改进，设计了可变状态机，此状态机允许双足机器人在运动过程中改变其关节控制律和控制参数，并调整每一步的时长和步长。借此，本文实现了SLIP模型下双足机器人多种高动态特性的运动模式，并实现了丰富的步态。关键词：双足机器人；运动控制；状态机；倒立摆模型

Biped robots have wide-ranging potential applications in areas such as domestic medical care, industrial production, and military fields, owing to their humanoid characteristics. Nonetheless, biped robots are characterized by high dimensionality, nonlinearity, and inherent instability, making motion control a persistent research hotspot. The conventional control approach suffers from drawbacks such as limited scalability and poor dynamic characteristics. The control framework of the state machine is employed in this thesis to develop control laws for biped robots using IP, LIP, and SLIP models, enabling stable locomotion such as walking, jumping, and running., etc. Dynamic locomotion and rich gaits are achieved, and at the same time, when the biped robot mechanism changes, the control method is still valid and has certain scalability. And it can achieve anti-interference, pass through structured terrain, and has certain robustness. Additionally, this thesis conducts experimental studies to investigate the motion behavior of biped robots. This involves analyzing the walking center of mass position and velocity laws, examining joint trajectories for various stances, and studying the impact of foot shape on walking performance. Provides a reference for the physical laws of human and leg-foot movement.Firstly, based on the open-loop analysis of the IP and LIP models, the concept of walking trajectory and phase diagram is proposed, and the walking phase diagram of the two models is drawn to describe the relationship between their speed changes during walking. And it is concluded that the walking speed of the LIP model has no upper limit in theory. Simplified biped robot models under the IP and LIP models are established in the Coppeliasim environment, and the walking state machines of the two models are designed respectively to realize stable walking and pass unstructured terrain such as ladders, slopes, and discontinuous ground. The walking phase diagram is verified by experiments, and it can be concluded from the phase diagram that the LIP model has a higher walking speed, but due to the limitation of the joints, the maximum speed is 2.5 times that of the IP model.Secondly, an eight-degree-of-freedom humanoid robot model is established in the Webots environment. The design of the state machine is founded upon the walking posture diagram. By observing and mimicking human walking patterns, this study establishes corresponding control laws for the robot‘s joints in various postures and uses the velocity of the body center of mass as feedback to achieve stable walking of the robot. Using eight linear, non-coupling control parameters, the walking control of the robot is parameterized, which greatly simplifies the difficulty of robot control. At the same time, the limitations of the step length and direction in the previous chapter are solved, and gaits such as straight walking, turning, and inplace walking are realized, and the joint motion laws in different gaits are analyzed. Secondly, using the scalability of the state machine, the stable walking under the change of the shape of the foot end is realized, and the walking performance of the foot end shape is analyzed.Finally, the biped robot of SLIP model is established in Webots environment and the state machine of SLIP model is designed. In the preceding chapters, we enhanced the fixed state machine and developed a flexible state machine. This new state machine has the ability to modify its joint control law and control parameters while the biped robot is in motion. Additionally, it can adjust the duration and step length of each step. It realizes a variety of motion modes with high dynamic characteristics such as walking, running, jumping, and crawling of the biped robot of the SLIP model, and realizes a variety of gaits.Keywords: bipedal robot; motion control; state machine; inverted pendulum model