Portfolio

HospiGuide Mobile Application

Project Overview:

Key Features and Technologies:

• Real-Time Navigation: The app offers real-time, step-by-step directions from users’ homes to their appointments within the hospital using GPS for outdoor guidance and our IPS technology for indoor guidance.
• Two Navigation Modes:
  • Advanced Mode: A real-time, step-by-step indoor guidance with the help of our Indoor Positioning System.
  • Assisted Mode: Offers adjustable navigation speed with on-screen instructions.
• Different Navigation Options: Users can choose from various navigation modes. For outdoor routes, options include traveling by car, bike, or on foot. For indoor navigation, users can select preferences such as using only elevators, only stairs, or a combination of both.
• QR Code Integration: Users can scan QR codes within the hospital to establish their starting point and begin navigation. If the application is installed on the device, the scan opens it directly, if not, the QR code redirects to the web version of the indoor map.
• Appointment Management: Users can sync their hospital appointments with the app, receive real-time updates, and notifications. This feature is helpful if they are unsure of the exact name of the department or service they need to visit.
• Testing and Quality Assurance: Conducted extensive on-site testing to guarantee reliability and accuracy of the navigation system, focusing on user experience and operational efficiency.

Indoor Positioning System back-end Server

Project Overview:

Technologies:

• Python: The primary programming language for the IPS back-end server, chosen for its simplicity and flexibility.
• Django: This web framework was used to build APIs that allow seamless communication between the mobile applications and the server.
• Postgres Database: A reliable and scalable database that manages the large datasets required for indoor mapping and wayfinding.
• pgRouting: A Postgres extension leveraged to calculate fast, efficient routes in real-time, ensuring quick response times in mobile apps.
• Custom OpenStreetMap integration: We used a custom fork of OpenStreetMap to populate the database with detailed indoor maps, which are then optimized for use within our system.
• Tensorflow: Employed for machine learning tasks, allowing the server to train models that improve the accuracy of indoor positioning.
• Rest Framework: A powerful library integrated with Django to create API endpoints that enable mobile applications to request and receive real-time data.

Key Features:

• Mapping Data Collecting: The system collects mapping data from our custom OpenStreetMap server and converts it into Python objects, making the data usable for our indoor positioning algorithms.
• Indoor Wayfinding: Using pgRouting, the server computes optimal indoor routes that are sent to mobile applications, enabling users to navigate complex environments efficiently.
• Outdoor-Indoor Wayfinding: By integrating with Mapbox Directions, we create routes from any location to the chosen indoor destination. The system determines the best entry point or parking area among multiple options. It can also divide the journey into segments, such as outdoor driving routes, outdoor walking routes, and indoor navigation.
• Offline versioned data: The server supports the delivery of offline data to mobile apps, allowing them to function without an active network connection, a crucial feature for areas with poor or no reception.
• Machine Learning Training: The back-end server trains machine learning models using TensorFlow, which are then deployed to mobile apps to perform different tasks for our indoor positioning system.
• On-site Data Gathering: The server stores data collected on-site, transmitted by our internal data-gathering application. This data is used to enhance the accuracy of our Indoor Positioning System.