组合行为学习优化离线动作库数据集
收藏资源简介:
本数据集包含四足机器人动捕数据重定位的全套数据,涵盖以下内容: 1)原始动捕数据:bvh 格式动作捕捉文件 2)Blender动画文件 3)中间处理文件:三维位置数据(.csv/.txt)、骨骼ID调整文件 4)仿真结果视频:质心位置、关节角度、轨迹速度文件 验证动画: PyBullet仿真视频:机器人动作验证视频(.mp4) Blender动画视频:足端绑定效果演示(.mp4) 数据生成方法: Retarget Motion方法:基于逆运动学的仿真映射 Blender足端绑定:通过控制骨与IK骨实现动作优化
This dataset provides a comprehensive collection of data for quadruped robot motion capture retargeting, covering the following contents: 1) Raw motion capture data: BVH format motion capture files 2) Blender animation files 3) Intermediate processing files: 3D position data (.csv/.txt), skeleton ID adjustment files 4) Simulation result data files: centroid position, joint angle, and trajectory velocity files Validation animations: PyBullet simulation video: Robot motion validation video (.mp4) Blender animation video: Foot-end binding effect demonstration video (.mp4) Data generation methods: Retarget Motion approach: Simulation mapping based on inverse kinematics Blender foot-end binding: Action optimization achieved by controlling bones and IK bones



