Robotics evidence guide · automated editorial preparation, not expert review

Keep simulation and hardware evidence separate

A simulation result supports a claim about the simulated model and conditions. Moving that claim to hardware requires new evidence about the physical task, not a change of label on the same result.

Evidence and interpretation

Maha’s proposed evidence packet identifies the execution domain and the assumptions that make the experiment tractable. Those assumptions can include perfect state information, simplified contact, discrete time or ideal actuators. State which are absent from the physical environment. A useful simulation can reveal software defects while leaving the hardest deployment questions open. Collect the hardware results under a separate run and connect the two through an explicit comparison, with divergences retained rather than explained away.

Proposed evidence workflow

  1. Mark each run synthetic, simulated or physical at collection time.
  2. List omitted physical effects and privileged information available to the policy.
  3. Require a separately reviewed protocol for any hardware transfer claim.

Worked illustration

The local grid simulation has perfect obstacle knowledge and no mass, torque, friction or people. Obstacle-aware planning can beat direct planning here without demonstrating that a humanoid can walk around an obstacle. The improvement is a fixture-level finding only.

Limits

No sim-to-real adapter or hardware experiment is delivered in this tranche. The source provides task decomposition; the transfer checklist is Maha’s original proposal.

Sources and review

NIST: Performance Assessment Framework for Robotic Systems

Locator: Summary; Description: Framework, Capability Models and Test Methods

The completed project decomposes assembly performance into perception, mobility, dexterity and safety-related components.

Boundary: A research-program description, not evidence of Maha performance or an operational safety approval.

Inspection and reuse basis

Selected sections inspected 2026-09-14. Link and original paraphrase only; no third-party images or publication copied. This is not independent expert review of this article.

Continue reading