Quadruped Robot Body Posture Estimation Methods Research Based On GPS/SINS Integrated Navigation

With the strong maneuverability and carrying capacity, and strong ability to adapt to complex road environment, Quadruped robot has a promising future and widely used in reconnaissance, anti-terrorism anti-riot, engineering exploration, and many other special circumstances, which means the research on quadruped robot body posture has important practical value and significance for the robotic path planning, gait generation, balance control and so on.In this paper, through the analysis of the mechanical parts structure and motion mechanism of quadruped robot, the robot body posture is estimated with quadruped robot as research object. On the basis of the analysis on motion characteristics and drive mode of control system, the ontology attitude algorithm of quadruped robot based on strapdown inertial navigation and the body posture estimation method of quadruped robot that based on the integrated navigation are studied.First, the geometric structure model of quadruped robot ontology is established, then according to the relationship of the carrier coordinate system, the northeast geographic coordinate system and the GPS coordinate system, the strapdown transform matrix between the carrier coordinates and navigation coordinates is deduced.Then, according to the requirement from coordinate planning layer of the quadruped robot state sensing system, combine with robot body geometry model, this paper present a research on quadruped robot attitude algorithm based on strapdown inertial navigation. By quaternions real-time calculating, a robot ontology calculation model of the real-time speed, position and pose is drawn, thus, to enable the calculation of robot ontology speed, position information, and pose information such as heading angle, pitching angle and roll angle.Further more, on the basis of strapdown inertial navigation calculation model, this paper based on GPS/SINS integrated navigation mode, using extended kalman filtering algorithm, estimate the posture of quadruped robot body. Through the establishment of state equation and measurement equation in integrated navigation estimation, the integrated navigation estimation model of robot ontology is established.In order to verify the validity of the robot ontology posture estimate method based on integrated navigation this paper proposed, a simulation experiment has been done with inertial navigation equipment to collect data, which testified the validity and feasibility of the method presented in this paper.