0. Wiki Socio-Cognitive Engineering

Principles

The development of interactive, human-centered automation should be built on theory and empirical research. To support the research & development processes systematically, a Socio-Cognitive Engineering (SCE) method was constructed for building, maintaining and re-using design knowledge based on the following principles:

•     Creating human-centered automation is a multi-disciplinary collaborative activity
•     Functional modules are defined and tested incrementally in an iterative refinement process
•     Design decisions are explicitly based on claims analyses, explicating the up-downside tradeoffs
•     Keeping and sharing the design rationale is key for progress and coherence in automation development
   

Origin

In an international project, the European Space Agency asked to establish a sound requirements baseline for a "Mission Execution Crew Assistant" (MECA) for future manned deep space missions (e.g. to Mars). As a concise method was lacking for the research & development of the envisioned human-automation system, the first version of the SCE methodology was constructed and applied. This methodology combines approaches from user-centered design, cognitive engineering and requirements analyses to establish a coherent "self-explaining" requirements baseline consisting of:

1.     The foundation that captures the relevant domain, human factors and technological knowledge.
2.     The specification of the objectives, use cases, functions (requirements) and the (expected) effects (claims).
3.     The evaluation validates these claims and advances the foundation knowledge.

The SCE activities that provide these outcomes can be performed in parallel. At "some time" they will be integrated into an evaluation (i.e., a prototype or simulation). For this we distinguish development cycles. Each development cycle provides a next version of a prototype. Milestones are specified for the SCE outcomes that need to be finished for such an evaluation (note: a demonstration can be viewed as a very minimal evaluation).

For agile R&D, SCE defines the Minimal Viable Product (MVP) as a coherent and concise set of (interim) SCE outcomes, i.e. a coherent set of milestones that lead to the envisioned prototype or simulation.

WiSCE is the successor to the Socio-Cognitive Engineering Tool (SCET) that was hosted on Atlassian Confluence. WiSCE provides design rationale templates and links design concepts to each other (see the SCE Guide). The top menu of WiSCE shows the SCE components (i.e., the "boxes" of the Figure: Foundation, Specification and Evaluation), the "meta-models" (i.e., Ontology and Design Patterns, and reference items. General information about the Socio-Cognitive Engineering methodology can be found at http://scetool.ewi.tudelft.nl/; an example application is provided by Neerincx et al. ^[1] (https://doi.org/10.3389/frobt.2019.00118).

Method

• The Foundation describes the
  • Operational Demands (e.g., stakeholders' values and needs, problem scenarios, work context),
  • Technology that will be used and/or (further) developed (e.g., cloud computing, AI frameworks) and
  • Human Factors knowledge that should be addressed in the design and evaluation of the technology to meet the operational demands.
• The Specification defines the
  • Objectives: the target outcomes
  • Use cases: how the human-machine collaboration takes place, i.e., the structure and flow of actors' actions with the task allocations (who, when, where),
  • Function (requirement), i.e., what the machine shall do to serve the objectives in the corresponding use cases,
  • Claim, specifying the expected Effect of the situated Function (i.e., situated in the use case) to provide the justification (why).
• The Evaluation provides the outcomes of the tests with the Prototype and/or Simulation.

The SCE method is iterative in nature, which means that usually several cycles of designing and testing are required to eventually arrive at a prototype or simulation. The generated behavioral and declarative design knowledge is formalized and maintained for re-use and sharing via, respectively, Design Patterns and a corresponding Ontology.

Detailed information on the methodology can be found in the publications section of this site.

References