code_weights_data_release_state · verified_hold
审核码:verified_hold; 官方关系:paper_and_author_video_only_no_public_method_artifact_found; 最近核验:2026-07-23代码:No official training, reward, simulator or Go1 deployment source was found.
权重:No LSTM, Transformer, Mamba or StratMamba policy checkpoint is released.
数据:No Isaac Lab rollout, Gazebo evaluation or 80-trial Go1 log is released.
许可证:not_disclosed;No method-code, model or trial-data license exists.
依赖:Isaac Lab training、Unitree SDK、RSLidar ROS2 package、VICON
硬件/传感器:Unitree Go1、RSLidar-16、offboard Intel i7-12700H and NVIDIA RTX 3050-Ti
实机证据:Four conditions use 20 trials each and report 19/20, 18/20, 19/20 and 17/20 success.
独立复现:No independent fixed-configuration Go1 deployment or policy reproduction was found.
限制/反面证据:Offboard GPU and VICON;single small indoor room and 5 m route;no code, weights, logs or license;simulation average success lower than vanilla Mamba
当前建议:高潜待持续验证;The design is worth tracking for legged navigation, but an onboard or localization-independent deployment and public policy closure are required for adoption.
重点标签:双时间尺度LiDAR导航主流迭代、artifact:not-released、real-robot:quantified-author-only、community:unverified
核验来源:来源 1 · 来源 2