Alliot – a Living Learning System

Made by Angela Wang

Found in Ecosystem Prototype Documentation · UNLISTED (SHOWN IN POOLS)

An IoT system that provides personalized experiences to Allegheny General Hospital visitors and real-time behavioral data for the hospital staff to deploy strategic physical and operational improvements.

0

Overview

The Allegheny General Hospital is challenged with various physical and operational constraints within the hospital – from navigation to the treatment experience – all stand as barriers to provide more humanized experiences for patients and their companions who suffer from cancer or other diseases. For the hospital staff, the lack of organized information makes it hard for them to identify and prioritize areas of improvement. AGH will need to seek innovative ways to quickly identify, predict, and respond to patients’ need during their visit.

Problem Statement

How might we leverage IoT technologies to enable AGH to make strategic decisions on hospital improvements using new and real-time information?

Our resulting IoT system, Alliot, addresses the three following question in particular:

  • How might an IoT system provide personalized navigation assistance to patients while identifying wayfinding pain points
  • How might an IoT system encourage patients to express their emotional responses to hospital visits and pinpoint the service and infrastructure shortcomings?
  • How might data gathered from an IoT system can provide insights to help AGH admins identify areas of improvement?
0

The Alliot Ecosystem

A fully implemented Alliot ecosystem will support patients and visitors throughout a full journey while allowing AGH to learn and improve on patient experience in real time. Figure 1 illustrates and overview of the Alliot system.

0

Technical Components

The Alliot system includes 5 main technical components that together support a robust living learning system.

0

For the purposes of this project, we opt for using existing purchasable products for the motion sensor and the fob. We focused on building out directional kiosk, directional signage, and emotional kiosk. We also proposed an admin dashboard where the collected information is displayed.

0

01 Directional kiosk

The directional kiosk provides basic map information of the hospital as well as personalized directional guidance based on individual patient’s itinerary information. The following video provides a detailed walk-through of the screen-based interaction.

0

Print-out

The print-out serves as an extended touch point for patients and visitors to continue explore in AGH. The print-out includes both the directional information and real-time and context-sensitive information about AGH facility. The latter provide patients and visitors activities to do during their wait time as some visitors and patients may not have readily accessible entertainment during wait time.

0

Navigation Algorithm Technology

Optimal routes through the hospital can be assigned algorithmically or by using predefined routes. While a correctly constrained algorithm is able to find the best route for every customer, there are many zones that should not or would preferred not to be accessed by patients or visitors due to safety and privacy concern. Therefore, all routes possibility should be determined by staff in the beginning of the development and then processed by the navigation algorithm technology.

0

02 Directional Signage

The directional signage will be mounted on walls, at major intersections within the hospital. The following video show a demo of how the directional signage would work.

0
The technology structure behind directional signage is simple –
  • a smart tag/fob with a passive RFID tag
  • an RFID reader (MFRC522) that converts the input into a digital output
  • a computer (Raspberry PI) that reads the digital input (unique identifier)
  • Wi-Fi connection to access the right direction information from cloud
  • the computer use the gathered information to light up the correct LED
0

Technical Development

Directional Signage

The following code is the basis for reading in the ID from the smart tag/fob.

0
import RPi.GPIO as GPIO
import SimpleMFRC522

reader = SimpleMFRC522.SimpleMFRC522()

try:
       id, text = reader.read()
       print(id)
       print(text)
finally:
       GPIO.cleanup()
Click to Expand
0

In order to light up the LED the code was adapted that based on the ID an LED lights up. To scale this solution the hardcoded look up of IDs stored in the memory has to be replaced with a request that accesses a database with the stored information.

0

Emotion Detection

For detecting emotions the Microsoft FACE API was used. An image was captured by using a webcam and processed via the API to receive the emotion output. Based on the output the mood screenshots were opened.

0
response = requests.post(face_api_url, params=params, headers=headers, data=image_data)
response.raise_for_status()
faces = response.json()

# Display the original image and overlay it with the face information
image_read = open(image_path, "rb"),read()
image = Image.open(BytesIO(image_read))

plt.figure(figsize=(8,8))
ax = plt.imshow(image, alpha=1)
for face in faces:
    fr = face["faceRectangle"]
    fa = face["faceAttributes"]
    origin = (fr["left"], fr["top"])
    p = patches.Rectangle(
         origin, fr["width"], fr["height"], fill=False, linewidth=2, color='b')
    ax.axes.add_patch(p)
    plt.text(origin[0], origin[1], "%s, %d"%(fa["gender"].capitalize(), fa["age"]), fontsize=20, weight="bold", va="bottom")
_ = plt.axis("off")
plt.show()

print(faces)
print(fa["emotion"])
Click to Expand
0

03 Emotional Kiosk

The emotional kiosk uses emotion detection as an entry point for the satisfaction survey. The following video provides a detailed walk-through of the screen-based interaction.

0

Print-out

Visitors can also print out a personalized postcard with their mood palette as a souvenir or for personal collection. AGH can use this for marketing purposes. 

0

Emotion Detection Technology

To simplify the solution of our application, we decided to use the existing FACE API from Microsoft. An API is an application programming interface that is used to communicate with Microsoft's computing resources. After calling the API the requested information are returned in the predefined way. In this application Microsoft would return the detected emotions of the patient / visitor. Based on the returned information the front-end programmed in Javascript is adapting its visible output.

0

04 Admin Dashboard

The admin dashboard is a platform for AGH admin to learn about the hospital operation, gather insights, and facilitate the making of operational decisions. The following video provides a detailed walk-through of the screen-based interaction.

0
As shown above, the admin dashboard displays the following information with accurate time-stamps:
  • visitor type (patient, visitor)
  • journey(arrive, engage, depart, etc)
  • location (area type, building)
  • feedback on services (medical, cleanliness of location friendliness of staffs, food services, etc)
  • journey

With the information, the accumulated insights include:

  • hourly foot traffic in specific locations
  • level of visitor engagement (# of input vs total traffic) with directional or emotional kiosks
  • level of visitor engagement at navigationally confusing areas (# of times directional guidance was provided)

The admin can also share critical information or schedule meeting with appropriate stakeholders within the platform and AI will auto suggest appropriate stakeholder based on the tags and the feedback content. Admin or other appropriate managers also receives weekly digest from Alliot system based on the monitoring preferences they made in their profile. 

0

Key Strategies and Rationales

During the concept development, we have established the following guiding principles we believe that would best suit the “living-learning” prompt, address feasibility, and provide a humanistic quality of IoT interaction to patients and companions.

01 Give visitors the opportunity to opt-in, not opt out.

We want to empower visitors and understand what services they would like to interact with.

02 Start with practical solutions that address existing urgent issues.

This led us to prioritize the implementation of directional kiosks and signages over emotional kiosks.

03 Use technology to reveal true needs that can be act upon.

When developing emotional kiosks, we wanted it to be more than an interesting experience for the visitors but one than can be most informative for AGH admins, hence including the following satisfaction survey.

04 Physical artifacts for low-fi implementations and familiar form factors

We opted for physical printouts of map as opposed to text messages/app-based solutions because we know that (1) it would take time for comprehensive Wi-Fi coverage to be implemented (2) it takes more money to iterate on interface and information digitally than physically (2) paper affords better glanceability. 

0

Guidance on Use

We hope the following implementation plan can provide as a framework for AGH to start considering implement part or all of the Alliot ecosystem.

0

Throughout the process we've used Sketch, Principles, Raspberry PI, Arduino, and a range of prototyping tools that can be easily adopted for further iterations. The native files are available upon request. 

x
Share this Project

This project is only listed in this pool. Be considerate and think twice before sharing.


Courses

Focused on
About

An IoT system that provides personalized experiences to Allegheny General Hospital visitors and real-time behavioral data for the hospital staff to deploy strategic physical and operational improvements.

Created

December 17th, 2018