Goal-Driven Autonomous Optical Experimentation via a Self-Recursive Hardware-Software Harness
Organizations: Konica Minolta (China) Investment Ltd., Shanghai, China · Giant Testing Equipment Ltd., Shanghai, China
Abstract
We propose goal-driven self-recursive optical research in which agents use physical evidence to revise the hardware, software, models, and protocols used to pursue a scientific goal. The framework can be instantiated with different stimulus systems and measurement devices. We formalize its updates and derive a relation between physical attainability, command access, and model approximation error. We demonstrate the concept in an optical see-through color-matching case study using an iPad, RayNeo X3 glasses, and a TOP 300 probe coupled to a CAS140D spectroradiometer, both from Instrument Systems. A revised glasses application adds USB-controlled RGB holds and frame acknowledgements, keeping model-generated colors displayed during spectral acquisition. The interface passes 12 live control checks. Spectral feedback reduces one target color difference from 19.756 to 1.211 ΔE_76^*; all five targets improve in a separate comparison. These measurements demonstrate the concept in one configuration. The resulting software and protocols remain available for subsequent investigations.
Figures & tables
| Agent role | Harness contribution | Goal-directed development |
|---|---|---|
| Scientific planning | Experimental questions, models, and test plans | Select informative tests and revise the investigation using new evidence |
| iPad application development | Background generation and stimulus-state observation | Add background patterns, sequences, or control features required by new questions |
| RayNeo application development | Glasses rendering and display control | Extend stimulus representation and rendering behavior for subsequent tests |
| Observation application development | Camera-spectrometer acquisition and analysis interfaces | Relate spatial and spectral observations and develop new acquisition protocols |
| Progress monitoring | Interpretation of progress against the persistent goal | Identify unresolved objectives and direct testing, analysis, or capability development |
| Scientific question | Measured outcome | Role in the research process |
|---|---|---|
| Can a new command interface enable the optical test? | Arbitrary RGB holds, 64 checked commands, and 12 passed live control tests | Software and protocol development make model-generated settings executable and observable. |
| Can feedback improve one displayed color? | Color difference after two updates; 93.9% reduction | An observation changes the next physical action while the target remains fixed. |
| Can the approach support several desired colors? | For five colors, median difference ; all five improve | A shared control and measurement environment supports a family of optical targets. |
| Can a revised model improve later tests? | The updated model gives closer matches in seven of nine new physical tests. | Evidence changes the reusable relationship that generates subsequent display settings. |
Autonomous Experimentation
Autonomous experimentation uses AI, active learning, and robotics to design, run, and interpret experiments with less human intervention.
See all 167 papers