PISPersonal Informatics System Lab
We study the person inside the system—then build interfaces, environments, and embodied agents that can sense, explain, adapt, and remain under human control.
Explore our projects
Our research loop
Sense.
Understand.
Adapt.
From human variation to intelligent systems that respond with care.
Personal Informatics begins with the individual.
PIS Lab studies how people perceive, understand, anticipate, and direct intelligent systems. Through interaction design, we turn computational capability into human capability—creating products, environments, intelligent hardware, and agents that respond to individual needs and support a harmonious human–AI relationship.
Critical change reaches the right person at the right moment.
Evidence, boundaries, and uncertainty form a usable mental model.
People can inspect possible consequences before they commit.
Questioning, steering, overriding, and recovery remain explicit.
We do not begin with a technology category. We begin with a human situation—then combine psychology, AI, design, and engineering until the whole loop works.
Human-in-the-loop XR
A lightweight XR system combines gaze, physiology and movement to estimate individual state, adjust embodied training intensity, and keep every intervention inside a visible safety boundary.
Human–AI co-creation
Instead of polishing a document, the system uses reflective dialogue to help students surface meaningful experience, construct a coherent narrative, and turn emerging self-knowledge into the next action.
Human factors + safety
We connect gaze, physiology, behavior and task context to model workload, attention and situational awareness—then translate those states into training, interface and safety decisions.
Interaction methodology
We develop interaction-design methods and prototypes that externalize system states, reasoning, boundaries and consequences, so people can inspect, question and steer intelligent behavior.
Tangible informatics: Bridging the cognitive disconnect in the intelligentization of industrial systems
Journal of Industrial Information Integration, 54, 101216, 2026.
Human-centered physical intelligence
We study embodied and physical AI as systems that enter human situations: interpreting intent, negotiating shared space and action, and keeping capability, uncertainty and control legible to people. VLA and world models connect perception, language, prediction and action around human goals.
Every project begins by inventing an interaction, puts it in front of people, builds a model from what emerges, and returns that knowledge to a product that can serve individuals better.
Invent a visual, auditory, tangible, or embodied interaction, then build enough of it to test the idea.
Put people inside the prototype and study the behavior, strategies, breakdowns, and new questions that emerge.
Turn observations into a working model that explains the new problem and guides the next design decision.
Refine the prototype into a product, validate it in use, and adapt the service to individual needs.
Everyone in the lab moves from a research question to a working prototype, then uses live interaction to make assumptions, system behavior, and human response testable.
Psychology, human factors, AI, interaction design, and robotics meet through shared projects—not parallel silos.
Lab administrator
Collaborator
Collaborator
Research assistant
Research assistant
Research assistant
Research assistant
Research assistant
Research assistant
Research assistant
Research assistant
Research assistant
Research assistant
Join the lab
We welcome students, collaborators, and partners who want to turn complex human conditions in the age of superintelligence into research that can be built, tested, and used.
poly.zhsun@outlook.comProspective students: please read How PIS Lab approaches research training (in Chinese) before contacting us.
Lingang Campus
