TAIPEI VETERANS GENERAL HOSPITALROBOTICS & AUTOMATION中文

CLINICAL ROBOTICS · PHYSICAL AI · SMART HOSPITAL

Robotics that take on risk.Clinicians who stay closer to patients.

Taipei Veterans General Hospital is moving from isolated robot deployments to integrated clinical automation—connecting AI, medical devices, autonomous mobile robots, and lifecycle governance.

8programs and governance pillars
7clinical and operational settings
5integrated system layers
Explore nuclear automation
OMNIVERSE DIGITAL TWINAutomated return of contaminated operating-room case cartsAccelerated simulation · 01:23

From technology demonstration to verifiable clinical service

STANDARDIZEINTEGRATEVALIDATEGOVERN
01 — NUCLEAR MEDICINE AUTOMATION

From system concept to integrated clinical workflow

Five project briefs document how TVGH advanced the automated radiopharmaceutical dispensing system from workflow definition in November 2025 to core automation and logistics integration by August 2026.

Diagram comparing manual radiopharmaceutical handling with the fully automated bidirectional closed-loop system
Blueprint: from manual handling to an automated bidirectional closed loop
NOV 2025WORKFLOW DEFINED

System blueprint and innovation showcase

Following the 2025 Taiwan Innotech Expo showcase, the team mapped the six manual operations—from preparation to delivery—and framed a four-part system combining cleanroom robots, precision dispensing, syringe transfer, and enclosed AMRs.

  • Mapped the baseline workflow across material preparation, quality control, dispensing, labeling, transport, and return.
  • Assigned four automation subsystems: R1 material handling, R2–R3 dispensing, R4 transfer and labeling, and enclosed AMR logistics.
  • Connected the cleanroom, pass box, corridor route, and injection-room exchange cabinet as one traceable service loop.
6 baseline processes4 integrated subsystems
Dynamic dose control equation showing patient target dose, measured concentration feedback, and radioactive decay compensation
Control logic: patient-specific target activity translated into draw volume
DEC 2025ALGORITHM DEFINED

Precision-control architecture

Specified the Xₙ control architecture: patient-specific target activity, concentration feedback, and radioactive-decay compensation, with ±0.5% defined as the system accuracy target.

  • Uses patient data and body-surface-area rules to calculate the prescribed target activity.
  • Feeds measured vial concentration back into the control loop instead of relying on a fixed nominal value.
  • Compensates for radioactive decay between reference time and dispensing time before calculating the required withdrawal volume.
Xₙ precision algorithm±0.5% design target
63-second system animation: robotic dispensing cell and material flow
Eighteen-step radiopharmaceutical automation workflow and equipment configuration
Open the complete 18-step workflow
MAR 2026ENGINEERING PLAN

Smart cell and 210-day execution plan

Converted the concept into an 18-step cell layout, robot roles R1–R4, digital-twin verification, secure closed-loop logistics, and a staged installation and acceptance plan.

  • Structured 18 steps across input, quality control, dispensing, labeling, autonomous delivery, return, and disinfection.
  • Defined robot roles: CRX-10iA for R1, LR Mate 200iD/4S for R2–R3, and CRX-5iA for R4.
  • Used digital-twin and physical proof-of-concept verification before staged installation, interface testing, and acceptance.
18 automated stepsR1–R4 robot roles
Enclosed AMR aligned with the injection-room radiopharmaceutical exchange cabinet
Milestone snapshot: the AMR-to-cabinet interface
MAY 2026LOGISTICS DESIGN

Closed-loop logistics interface defined

Finalized the physical interface and acceptance criteria needed to connect cleanroom dispensing with the injection room through one bidirectional AMR route.

  • Outcome: one sealed interface for outbound dose delivery and used-carrier return.
  • Acceptance focus: docking, payload, door interlocks, identity, and handover records.
3 enclosed AMRsBidirectional delivery & return
View the five-step workflow
August 2026 automation progress chart across six radiopharmaceutical processes
Integration status: completed core modules and interfaces still in progress
NSTC research diagram using LSTM, TSDC, precision dispensing, and activity-measurement feedback
NSTC research: AI-assisted precision dispensing, August 2026–July 2027
AUG 2026INTEGRATION

Core automation completed; integration advanced

Core quality-control and dispensing modules and the AMR delivery process were completed. Pretreatment vision and handling, activity-meter/API connectivity, and robotic labeling remained in integration; engineering review also covered load, power, and network interfaces.

  • Completed the quality-control and dispensing core and the autonomous AMR delivery process.
  • Continued pretreatment vision and handling plus activity-meter, API, and robotic-labeling integration.
  • Launched a NT$1.5 million NSTC study combining LSTM demand prediction, TSDC control, precision dispensing, and activity-measurement feedback.
Core modules completedNSTC research launched

Source: internal project progress briefs dated November 2025, December 2025, March 2026, May 2026, and August 2026. Stated performance values are design targets unless otherwise identified as completed work.

WORKFLOW DEEP DIVE

Five controlled handoffs, one closed loop

The roadmap above records the May 2026 design milestone. This section shows the operating sequence once, from AMR arrival through carrier return and disinfection.

37-SECOND DESIGN ANIMATIONDock · open · transfer · reset · depart
OUTBOUNDPrepared dose → injection room
RETURNUsed carrier → cleanroom disinfection
01
Dispatch and precision docking

The enclosed AMR reaches the injection-room exchange cabinet and confirms its docking position.

02
Door opening after alignment

After docking is confirmed, the AMR compartment and exchange cabinet open for a closed, low-contact handover.

03
Automatic carrier transfer

The exchange mechanism moves the lead-container carrier horizontally into or out of the AMR, reducing manual lifting.

04
Position confirmation and reset

The system confirms the carrier position, closes the doors, and records arrival and handover status.

05
Bidirectional return and disinfection

Used lead containers follow the reverse route to the cleanroom for automated 75% alcohol disinfection and reset.

The animation illustrates the mechanical and workflow concept. Routine operation remains subject to final safety interlocks, docking verification, infection-control requirements, and acceptance documentation.

02 — PROGRAM PORTFOLIO

Six highlights from the full Chinese site

A concise view of high-risk task automation, intelligent hospital logistics, human–robot collaboration, and evidence-based governance.

01SMART LOGISTICS

Operating Room Case-Cart AMR Logistics

Ten AMRs connect surgical schedules, dispatching, elevators, and visual control to automate cross-floor case-cart return and delivery.

10 AMRs · 200 carts · ≥95% task success target
02CLINICAL AUTOMATION

Nuclear Medicine Dispensing & Delivery

Four cleanroom robot arms, activity feedback, labeling, and three enclosed AMRs form a traceable dispensing-to-delivery loop designed to minimize occupational exposure.

18 steps · ≤90 sec per dose · 3 enclosed AMRs
03PHARMACY AUTOMATION

Automated Chemotherapy Compounding

Robotic compounding integrates prescription verification, precise preparation, closed-system handling, and digital records to strengthen safety and consistency.

Closed-loop verification · standardized workflow
04HUMAN–ROBOT COLLABORATION

Technology-Enabled Intensive Care Unit

Mobile robots support transport, retrieval, and routine logistics so critical-care professionals can spend more time on direct patient care.

TICU deployment · task-oriented collaboration
05CLINICAL EVIDENCE

Nutrition Delivery Study

A prospective, IRB-approved comparison evaluated robot-assisted meal delivery across twelve wards while documenting waiting time, temperature, integrity, satisfaction, and human intervention.

120 deliveries · 12 wards · IRB-approved study
06LIFECYCLE GOVERNANCE

Enterprise Robotics Governance

A hospital-wide governance model aligns clinical needs, infrastructure, cybersecurity, acceptance testing, maintenance, incident reporting, and continuous improvement.

Shared standards · traceable records · scalable deployment
03 — SYSTEM ARCHITECTURE

Five layers, one connected service

Automation becomes reliable when clinical events, AI decisions, sensing, device control, and governance are designed as one system.

01

Clinical Data

HIS/RIS, schedules, destinations, and patient-flow events

02

AI Decision

Prioritization, prediction, and dynamic compensation

03

Edge Sensing

Barcodes, vision, location, activity meters, and LiDAR

04

Device Control

Robot arms, AMRs, elevators, and automatic doors

05

Data Governance

Cybersecurity, audit trails, versions, and KPIs

04 — CLINICAL GOVERNANCE

Four gates before routine operation

Every deployment is evaluated as a clinical service—not only as a robot—so safety, integration, evidence, and long-term ownership remain visible.

01

Need & Value

Define the clinical problem, baseline workload, risk, and intended benefit.

02

Site & Integration

Verify routes, elevators, access control, networks, charging, fire safety, and clean/dirty zoning.

03

Safety & Acceptance

Test obstacle avoidance, emergency stops, payloads, cybersecurity, infection control, and exception handling.

04

Operation & Governance

Maintain device records, software versions, preventive maintenance, reporting, backup plans, and improvement cycles.

FULL PROJECT DETAILS

Continue to the complete Chinese website

View detailed project milestones, clinical evidence, patents, awards, technical specifications, videos, and downloadable reference materials.

瀏覽中文完整版