Spatial Critic - Evidence-Based Spatial Interaction

A spatial prototype becomes research when its claims are testable.

Many architectural prototypes stop at representation. They show how a space might look, how a user might move, or how a design concept might be experienced. Spatial Critic begins from a different question: can a spatial scene become a controlled instrument for studying movement, ambiguity, and decision-making?

The project does not claim that a simulation can replace human experience. It also does not claim that AI can predict how people will feel inside architecture. Its value is more careful than that. Spatial Critic creates a baseline environment where spatial conditions can be adjusted, navigation behavior can be recorded, and future participant data can be compared against explicit design variables.

How do spatial conditions influence navigation behavior?

The central research question is: how do changes in threshold, visibility, orientation, and route structure affect how people choose paths through a spatial environment?

This question matters because architectural experience is often evaluated through subjective language: intuitive, confusing, open, compressed, engaging, disorienting. These words are useful, but they need a method for being tested. Spatial Critic translates selected spatial qualities into observable behavior: wrong turns, hesitation, path variation, task completion, repeated routes, and decision time.

The prototype therefore sits between architectural design and HCI. It treats space not only as form, but as an interface that organizes choices.

Participant mode and researcher mode are deliberately separated.

Spatial Critic uses two modes because the participant and the researcher need different relationships to the same environment. The participant should encounter the scene as a navigation task, without being overloaded by analytic labels. The researcher needs to inspect conditions, compare traces, and understand how the scene is structured.

This separation is an ethical and methodological choice. If the participant sees every variable, the task becomes self-conscious. If the researcher sees only the final path, the design logic becomes opaque. The prototype keeps experience and analysis connected, but not collapsed into one view.

Anonymous trace export extends this logic. The goal is to study movement patterns without attaching unnecessary identity to the data. For a future pilot study, this makes it easier to treat movement traces as research material while keeping participant privacy at the center of the workflow.

The first measures are simple on purpose.

Spatial Critic should not begin with complex prediction. It should begin with observable, explainable metrics. A useful first study can measure time to complete a task, number of wrong turns, pauses before decision points, route length, repeated backtracking, and whether the participant reaches the intended destination.

These measures do not explain the whole experience. A pause might mean confusion, curiosity, or aesthetic attention. A longer route might be inefficient, but it might also reflect exploration. For this reason, behavioral traces should be paired with short post-task questions and interviews. Quantitative traces show what happened. Qualitative responses help explain why it may have happened.

The purpose of the Trace Analyzer is therefore not to produce a final verdict. It creates a structured way to look at spatial behavior, compare sessions, and identify where design conditions produce uncertainty, delay, or discovery.

The next research step is human validation.

The immediate next step should be a pilot study with 12 to 20 participants. The goal of the pilot is not publication by itself. It is to test whether the task instructions are clear, whether the logging system captures useful data, whether participants understand the environment, and whether the measures are sensitive enough to show differences between spatial conditions.

After the pilot, the workflow can be revised before a larger formal study with around 40 to 60 participants. That study could compare two or three spatial conditions: for example, a clear route, an ambiguous route, and a route with misleading visual affordances. Participants would complete navigation tasks while the system records movement. After each task, they could report perceived difficulty, confidence, and cognitive load.

This sequence matters for graduate study because it demonstrates research maturity. The project does not jump from prototype to conclusion. It moves from prototype, to pilot, to revised method, to formal study.

Spatial Critic is the first instrument in a larger research path.

Within ALTN.VISION, Spatial Critic should be understood as the first testbed for Human-AI Spatial Interaction. It is the project that converts the phrase "space as interface" into something that can be examined.

Its future value is not limited to the current scene. The same method can later move into museum wayfinding, exhibition design, commercial interiors, hospitals, transit spaces, or responsive environments. In each case, the question remains consistent: how do people read spatial cues, make decisions, and preserve agency inside increasingly intelligent environments?

The strongest claim Spatial Critic can make is also its most careful claim: before AI-mediated spaces can adapt to people, designers need better ways to observe how people already respond to spatial conditions.

1. Add screenshots from participant mode, researcher mode, and Trace Analyzer.
2. Define a precise pilot protocol with task script, consent language, and measures.
3. Add a table connecting spatial primitive, design condition, behavior, and metric.
4. Write limitations around simulation, sample size, and ecological validity.