Feasibility study demonstrates an active–reactive digital twin architecture in a robotic cell, indicating viable low-cost asynchronous coordination for custom manufacturing.
Mass-customized production introduces variants faster than certified asset models can be rebuilt. This feasibility study presents an active–reactive architecture for type-3 Asset Administration Shells (AASs) in which two submodel templates separate the reactive layer from independently deployable decision services. On the requester side, ActiveSteps represents a variant’s process sequence and references the capability each step needs, so a product’s AAS becomes a type-3 on demand. On the provider side, ActiveSkills maps offered skills to the capabilities that realize them, so one provider-core image serves every resource and only its domain services are scenario-specific. The type-2 to type-3 upgrade adds these submodels to both sides and costs one restart of the provider alone; subsequent rollouts needed none, so consumers were not interrupted. Evaluation used a university robotic screwdriving demonstrator, not a live production line. Of 104 calls for proposal, 102 were answered at a median of 4 s. Median requester activation was 8 s over 100 activations. Capability checking over 454 balanced cases with five repetitions reached an F1 of 89.6% and a 10.8% false-positive rate. An evaluation costs USD 7.0×10−4. These results support feasibility for asynchronous coordination in the evaluated cell; scalability and cross-vendor interoperability were untested.
No takes yet. Share an insight, caveat, or question.
Silva et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: