claimed_full_stack_repository_ownership_unresolved · hold_official_ownership_not_independently_verified
审核码：hold_official_ownership_not_independently_verified; 官方关系：claimed_method_repository_ownership_unresolved; 最近核验：2026-07-20代码：仓库包含实质 C++/ROS、KUAVO 全身控制、OCS2、Gazebo/MuJoCo 与 polygon_mapping 内容，但同时捆绑 installed/vendor 树，方法新增部分和第三方边界不清。
权重：未核验到论文使用的感知或控制模型权重、版本化 Release 或可比实验检查点。
数据：未核验到五次真机楼梯试验日志、标定文件、点云/深度记录或论文评测场景数据。
许可证：MIT；根 MIT 文件可见，但仓库包含大量 vendor、installed、OCS2、机器人模型和资产，分别可能受上游条款约束。许可证存在不能证明仓库官方归属，也不能自动覆盖捆绑依赖和机器人资产。
依赖：ROS and a bundled KUAVO humanoid control workspace、OCS2, Gazebo/MuJoCo and project-specific polygon mapping、RealSense RGB-D, LiDAR odometry and robot-specific calibration、dependency versions and clean-room installation steps are incomplete
硬件/传感器：paper hardware: 166 cm, 55 kg, 28-DOF KUAVO humanoid、Intel RealSense L515 RGB-D, Livox Mid-360 LiDAR, IMU and NVIDIA Orin NX、author-side stack runs at 20-30 Hz; repository-to-paper hardware identity is not independently verified
实机证据：The primary paper reports five indoor/outdoor KUAVO trials over four-to-six-step staircases, with maximum real foothold error 33.4 mm and a toe collision on a protruding edge during fast gait.；Those paper results are strong author evidence, but no released logs or independent evidence ties the wine3603 repository to the experiments.
独立复现：No issue, release, author statement or independent repository documents a clean deployment of wine3603/Polygmap on the paper hardware. BTFrontier/polygon_mapping is a separately licensed prior component cited as reference [15], not a reproduction of the full PolygMap locomotion system.
限制/反面证据：精确引用论文不等于作者官方发布；当前没有可复查的一手链接把 wine3603 账号与论文作者对应起来。；仓库体积大且混入 installed/vendor/完整控制栈，难以区分 PolygMap 新增实现、上游代码与预构建产物。；没有论文真机日志、标定、模型权重或版本化 Release，也没有独立复现。；BTFrontier/polygon_mapping 是论文引用的前置 RGB-D 多边形建图组件，不能误标为 PolygMap 官方整套代码。
当前建议：实机感知步态高价值线索，仓库归属待核；多边形地图到落足规划的真机路线具有明确实践增量，但来源真实性优先于仓库规模；只有作者或项目页确认所有权并补足可复现实验后才能晋级官方代码。
重点标签：recency:2025-paper、关联:主流感知运动技术2.0、关联:近期热门、关联:技术改进版、perception:polygonal-stair-mapping、evidence:paper-real-robot、artifact:ownership-unresolved、community:no-independent-reproduction
核验来源：来源 1 · 来源 2 · 来源 3 · 来源 4 · 来源 5
official_referenced_perception_planning_code · verified_official_code_only
审核码：verified_official_code_only; 官方关系：verified_primary_paper_direct_repository_reference; 最近核验：2026-08-14代码：已核验：论文正文直接引用仓库；固定提交含 ROS/catkin、KUAVO 控制与多边形建图、Gazebo 启动和阶梯测试入口。
权重：未核验到论文感知、规划或控制检查点；不能把仓库中的依赖或模型文件称为 PolygMap 权重。
数据：未核验到五次真机楼梯试验的点云、深度、标定、场景或失败日志。
许可证：MIT；根 MIT 文件覆盖该仓库发布代码；OCS2、ROS、KUAVO 相关硬件包、仿真器、模型、installed/vendor 树与机器人资产仍须分别核对上游条款。
依赖：Ubuntu 20.04, ROS Noetic and catkin workspace、OCS2, CuPy and the repository's KUAVO / polygon-mapping dependencies、Gazebo for the documented simulation launch; real hardware requires robot-specific drivers and calibration
硬件/传感器：Paper platform: 166 cm, 55 kg, 28-DoF KUAVO humanoid、Downward RealSense L515 RGB-D, Livox Mid-360 LiDAR, IMU and NVIDIA Orin-class onboard compute、The paper reports 20–30 Hz planning, but this review did not execute the stack or reproduce that rate
实机证据：The paper reports five author-run indoor/outdoor KUAVO stair trials, four to six steps per trial, 20–23 Hz perception/planning, and a maximum reported foothold error of 33.4 mm.；The outdoor single-step case reports a toe collision on a protruding stair edge; this is retained as a failure mode rather than omitted.
独立复现：None located. Code availability and direct paper linkage are not independent reproduction, and no released logs establish paper-comparable end-to-end hardware execution.
限制/反面证据：没有独立复现、版本化 release 或论文真机日志。；README 安装命令与大规模 bundled/installed/vendor 树仍需在干净环境逐项验证。；论文验证仅覆盖特定 KUAVO、楼梯形状、传感器配置与五次作者试验，不能外推至 G1、其他人形、未知楼梯或商用安全保证。；MIT 根许可证不自动覆盖上游依赖、资产、驱动或未发布数据/权重。
当前建议：主流感知运动技术 2.0 候选，特定本体实机可复核；把多传感器几何建图、落足可行域和全身规划连接到特定人形真机，且代码关系已由论文直接确认；下一步价值取决于干净环境复现、真实日志与跨本体/跨楼梯验证。
重点标签：关联:主流感知运动技术2.0、关联:技术改进版、关联:社区待验证、evidence:paper-real-robot、artifact:official-code-fixed-commit
核验来源：来源 1 · 来源 2 · 来源 3 · 来源 4