Anesthesia Robot Remote Platform
CASE STUDY

Anesthesia Robot Remote Platform

A web platform enabling anesthesiologists to remotely monitor and control intelligent anesthesia robots across multiple simultaneous surgeries in real time.

UIUX Human-Robot Interaction Product Design
Role
UX Designer
UX research, information architecture, interaction design, system thinking
Timeline
2023
Affiliation
Chinese Academy of Sciences
Tools & Stack
Figma UX Research Information Architecture Web Design

China’s anesthesiologist shortage
meets a robotic solution

2.4×

Ratio of surgical demand to available anesthesiologists, widening with aging demographics.

1 → N

Core proposition: one doctor supervises multiple robot-assisted surgeries remotely.

0ms

Missed command tolerance is effectively zero in remote infusion control.

The design challenge

The robot automates drug delivery, but the anesthesiologist must remain in control for critical decisions, anomaly response, and emergency interventions.

The UX challenge: provide full situational awareness and trusted remote control across cloud infrastructure, without latency ambiguity or cognitive overload.

The platform also acts as a nationwide patient record system with strict privacy, role-based access, and full audit traceability.

Understanding target users

👨‍⚕️

Dr. Zhang — Anesthesiologist

PRIMARY USER

“I monitor multiple surgeries at once. I need instant awareness if any case deviates.”

Core Needs
  • Real-time visibility of vital signs across concurrent sessions
  • Fast, trusted remote drug adjustment workflows
  • Clear normal-vs-anomalous visual states
  • Immediate connection loss visibility
Pain Points
  • Cognitive overload when monitoring in parallel
  • Latency anxiety between command and pump response
  • Record retrieval friction across systems
👩‍💼

Liu — Platform Administrator

ADMIN USER

“Account access has to be precise and revocable immediately, with full accountability.”

Core Needs
  • Permission governance with audit logs
  • Rapid access revocation and role reassignment
  • System-wide visibility of sensitive actions
Pain Points
  • Manual account workflows are brittle
  • Permission states are hard to inspect
  • Deletion/edit actions lack guardrails
🧑‍💻

Wang — Data Maintainer

SECONDARY USER

“I need to correct records safely without touching legally protected intraoperative data.”

Core Needs
  • Clear editable vs read-only field boundaries
  • Safe confirmation for data edits
  • Patient lookup with minimal friction
Pain Points
  • Editability ambiguity in existing systems
  • Weak per-record history visibility
  • Overly strict search field burden

Synthesized User Needs

Fast anomaly scanningColor and hierarchy should enable sub-2-second recognition.
🎛️
Confident remote controlClear command acknowledgements through the full pipeline.
🔐
Permission trustNo ambiguity on what each role can access or edit.
🗂️
Contextual navigationMove between records, logs, and vitals without losing context.

Critical Pain Points

🧠
Parallel surgery loadHigh chance of missing slow-forming anomalies.
📡
Network uncertaintyDoctors need explicit command-state feedback.
🔎
Lookup frictionPrivacy-required verification needs better UX explanation.
✏️
Protected-field confusionIntraoperative data must be visibly non-editable.

Understanding how the
anesthesia robot actually works

Interface decisions came from system behavior. I studied drug-delivery mechanics, command routing, and surgical protocol dependencies before finalizing interactions.

The robot’s role in surgery

The robot continuously monitors vitals and adjusts infusion rates to maintain target anesthesia depth. The doctor remains responsible for intervention and clinical decisions.

PropofolRemifentanilSufentanilEphedrineEsmololAtropine
Operating room robot system layout
Real-time data flow architecture
🏥

Local Machine

Collects vitals, runs dosing logic, controls pumps directly.

☁️

Cloud Server

Stores data, routes commands, enforces role-based access.

💻

Web Platform

Doctor-facing UI for live monitoring and remote intervention.

Critical constraint: Web UI → Cloud → Local machine → Pump must confirm each command stage.

Surgery phases and monitoring priorities
Phase 01

Induction

Rapid transition to anesthesia depth. Highest risk period.

Phase 02

Maintenance

Primary remote monitoring phase with AI adjustment.

Phase 03

Recovery

Drug tapering and post-op risk observation.

Mapping the system

Information Architecture — Frontend

Medical Data Management Platform
Auth
  • Login
  • Register
  • Password reset
Dashboard
  • Surgery metrics
  • Recent patients
  • Quick actions
Database
  • Search
  • Patient profile
  • Vitals and logs
Remote
  • Session hub
  • Live monitoring
  • Pump controls
User TypePatient DataEdit RecordsRemote ControlUser MgmtExport
AnesthesiologistOwn hospitalIf grantedIf grantedIf granted
Data MaintainerOwn hospitalNon-surgical onlyIf granted
AdministratorAll hospitalsAll records✓ Full

Three critical flows through
the system

Close recreation of the flow intent, presented in three parallel flow maps.

Authentication Flow

StartLoginValid creds?Role checkAdmin Panel / Dashboard

Database Lookup Flow

DashboardPatient SearchEnough fields?ResultsProfile + Tables

Remote Surgery Flow

Remote HubConnectServer connected?Active SessionAlert → Intervene

User journey

Six key screens across the full flow, with rationale aligned to critical UX constraints.

Screen 01 / 06

Login

Role-aware split entry with clear accountability.

Login screen
Screen 02 / 06

Home Dashboard

Workload and trend visibility with quick patient re-entry.

Home dashboard screen
Screen 03 / 06

Patient Search

Privacy-first retrieval with fast path and filters.

Patient search and data management screen
Screen 04 / 06

Remote Control Connecting

Connection setup and verification before entering active control.

Remote control connecting screen
Screen 05 / 06

Remote Control Hub

Grid-first parallel session management and notifications.

Remote control main hub screen
Screen 06 / 06

Active Session

Maximum-density command and monitoring environment.

Remote control in-session screen

Try the prototype

What I’d do differently,
and what comes next

Current gaps

Emergency flow completeness

Network drop, pump failure, and severe-state response flows need full coverage.

Key learning

Domain fluency shapes layout

Clinical understanding changed system structure, not just labels.