计算机工程与应用2026，Vol.62Issue(10)：89-98,10.DOI:10.3778/j.issn.1002-8331.2509-0409
四足机器人越障姿态估计算法研究
Research on Obstacle-Crossing State Estimation Algorithm for Quadruped Robots
摘要
Abstract
To address the issues of insufficient motion stability and excessive energy consumption during obstacle cross-ing by quadruped robots in complex and rugged terrains,this paper proposes an adaptive posture estimation method based on body perception,combined with model predictive control and whole-body control algorithms,to enhance the motion stability of quadruped robots on uneven terrains.Firstly,by optimizing the foot-end trajectories,the robot can adapt to various terrains,achieving smoother and faster foot-end movements,and enhancing the obstacle-crossing stability in dif-ferent terrain environments.Secondly,a finite state machine is designed to dynamically switch the obstacle-crossing gaits based on the terrain posture estimation results,enabling the robot to consider terrain information during obstacle crossing,optimize its foot-end trajectories,and further improve the obstacle-crossing performance and stability.Finally,to verify the effectiveness of the proposed method,the algorithm is tested on a stair model with obvious ruggedness and unstruc-tured terrains.The results show that the robot's motion energy consumption is reduced by 9.78%during the stair climbing process and by 18.27%during the stair descending process,demonstrating that the proposed method can effectively adapt to rugged terrain environments and perform posture estimation,achieving low-energy and high-stability obstacle-crossing behaviors in such environments.关键词
四足机器人/足端轨迹规划/模型预测控制/状态估计/本体感知Key words
quadruped robot/foot-end trajectory planning/model predictive control/state estimation/self-perception分类
信息技术与安全科学引用本文复制引用
吕明昱,魏霞,蔡鑫,李云飞,罗相涵..四足机器人越障姿态估计算法研究[J].计算机工程与应用,2026,62(10):89-98,10.基金项目
天山英才-青年拔尖人才项目(2024TSYCCX0011) （2024TSYCCX0011）
国家自然科学基金(62303394). （62303394）
