有界保真度模拟即演示阶段：用于治理基准的动捕交接
Bounded-Fidelity Sim-as-Demo-Stage: Mocap Handoff for Governance Benchmarks
浏览论文内容
中文总结 AI 辅助
提出有界保真度模拟即演示阶段模式，通过抑制交接包络内接触物理，实现LLM驱动机器人治理基准的审计链字节级可重现，以近乎零成本提升基准可靠性。
中文摘要 AI 辅助
仿真到现实的研究将物理保真度作为首要目标：仿真器通过其再现真实接触动力学的程度来评判。对于由LLM驱动的机器人的治理基准测试，其中仿真器证明准入/策略/合同/审计流程行为正确，物体交接（抓取、携带、放置）时的接触保真度成为一种负担：接触力积分噪声引入了审计链分歧，该分歧与被测治理属性在结构上无关。我们提出有界保真度模拟即演示阶段，这是一种设计模式，在明确括起的交接包络内抑制接触物理，同时在其他地方保持完整动力学。该构造使用由220行Python适配器驱动的MuJoCo的动捕体原语，治理桥通过结构化意图调用该适配器。我们将审计链稳定性形式化为重放中哈希事件日志的字节相等性，并识别出两个隐含它的结构包络属性。在每种姿态N=1000次重放中，动捕变体产生一个不同的审计链哈希（1000/1000字节相同；Wilson 95% CI [0.997, 1.000]）；接触力基线产生584个不同哈希（993/1000分歧；CI [0.987, 0.998]）。时间步扫描（1、2、5、10毫秒）表明分歧是结构性的，而非调优伪影：在每个时间步它保持在0.985。包络边缘时间抖动（+/-10仿真步，1,400次重放）产生0分歧，并且审计链在K在{1, 2, 3}中顺序交接的物体（1,500次重放）中保持字节相等，每次拾放开销亚线性。该模式以近乎零的工程成本为基准设计者提供审计链可重现性；我们还绘制了其有害之处（仿真到现实验证、策略训练、接触丰富任务），以免误部署。
英文摘要
Sim-to-real research pursues physics fidelity as a primary objective: simulators are judged by how closely they reproduce real-world contact dynamics. For governance benchmarking of LLM-driven robots, where the simulator demonstrates that an admission/policy/contract/audit pipeline behaves correctly, contact fidelity at object handoffs (grasp, carry, place) becomes a liability: contact-force integration noise injects audit-chain divergence that is structurally unrelated to the governance property under test. We propose bounded-fidelity sim-as-demo-stage, a design pattern that suppresses contact physics within explicitly bracketed handoff envelopes while preserving full dynamics elsewhere. The construction uses MuJoCo's mocap-body primitive driven by a 220-line Python adapter that the governance bridge invokes via structured intents. We formalise audit-chain stability as byte-equality of the hashed event log across replays and identify two structural envelope properties that imply it. Across N=1000 replays per posture, the mocap variant produces one distinct audit-chain hash (1000/1000 byte-identical; Wilson 95% CI [0.997, 1.000]); the contact-force baseline produces 584 distinct hashes (993/1000 diverged; CI [0.987, 0.998]). A timestep sweep (1, 2, 5, 10 ms) shows the divergence is structural, not a tuning artefact: it stays at 0.985 at every timestep. Envelope-edge timing jitter (+/-10 simulation steps, 1,400 replays) produces 0 divergence, and audit chains remain byte-equal across K in {1, 2, 3} sequentially handed-off objects (1,500 replays) with sub-linear per-pick-and-place overhead. The pattern gives benchmark designers audit-chain reproducibility at near-zero engineering cost; we also map where it is harmful (sim-to-real validation, policy training, contact-rich tasks) so it is not mis-deployed.
发表机构
- Harbin Institute of Technology（哈尔滨工业大学）
- Soochow University（苏州大学）
机构由 AI 辅助整理，请以论文原文为准。
