Skip to main content

Final Presentation

The Transition milestone (MS9) closes the project. This page summarizes what we presented in the final presentation (MS2.6), where we walked through the journey of the Automotive App from its objective to its final results.

Objective​

Develop an app for Android Automotive OS (AAOS) capable of leveraging the sensor data deployed around Aveiro to provide an Automotive Unified Road Assistant - a single in-vehicle assistant that warns drivers about road events and helps them drive more safely.

Opportunity​

The Telecommunications Institute (IT) has deployed a wide range of equipment across the Aveiro region - radars, cameras, weather stations, and other sensors. On top of this, the Digital Twins for Mobility (DT4MOB) infrastructure, built on Eclipse Cloud2Edge and orchestrated with Kubernetes, provides:

  • Faster communication and interaction between objects.
  • Rapid and efficient event detection - a perfect fit for a road assistance app.

This combination of real sensor coverage and a digital-twin platform was the opportunity our project set out to exploit.

Architecture​

The final architecture is split into clear layers and a set of independent backend microservices, all decoupled through a Message Broker (MQTT):

  • Data Gathering - real data ingested from IT through Hono and modeled as digital twins in Ditto.
  • Data Processing - the Position Processor and Proximity Filter clean and route incoming data.
  • Event Detection - dedicated detectors for Overtaking, Speeding, Traffic Jam, Lane Merge, Accidents, Emergency Vehicles, and Weather.
  • Data Storage - a Preference Database accessed through the Database Manager.
  • Authentication - user authentication on top of the DT4MOB authentication service.
  • Event Transmission & UI - alerts published over MQTT and rendered by the Android Automotive OS app (built in Kotlin).

Components were clearly classified during the presentation as developed for the project (app/backend), recreated DT4MOB infrastructure used for validation, or out of scope.

Development Methods​

Throughout the project the team relied on:

  • Constant communication (Slack, WhatsApp) and weekly meetings with the advisers.
  • Task management in Jira, organized by sprints and releases.
  • Pull requests with mandatory review (at least two approvals) and CI checks before merging to main.
  • A shared environment on GitHub across the four repositories.

Main Obstacles​

  • No Google services - being on AAOS without GMS, we replaced Firebase/Google Maps style stacks with open alternatives (e.g. Ditto for real-time data and OSRM for routing).
  • Complex, still-in-development infrastructure - Hono, Ditto, and the surrounding digital-twin platform were evolving as we built on top of them.
  • Strict environment - Android Automotive OS imposes distraction-optimization and safety constraints on every screen.

Demonstration​

We demonstrated the assistant end-to-end, including a real-car field test at the IT camp for overtaking and maneuver coordination, plus the remaining use cases:

  • Provide knowledge of: Speeding, Overtakes, Accidents, Traffic Congestions, Emergency Vehicles, Weather Conditions.
  • Assist with: Maneuver Coordination and Navigation.

Results​

Functional Requirements​

Most functional requirements were met, including overtaking detection, speeding alerts, accident and traffic-congestion identification, emergency-vehicle detection, weather conditions, navigation and route generation, safe highway entry/exit, and enabling/disabling alerts. A few were left out of scope, namely roadworks identification, pavement condition alerts, and recording data for later analysis.

Non-Functional Requirements​

We satisfied safety (encrypted communication), accuracy (GPS error ≤ 5 m), performance (alerts displayed ≤ 2 s), and usability (no intrusive pop-ups). Scalability (100,000 users) and reliability under long disconnections were partially met and are addressed in our future work.

Tutors:

Group:

Institution: Telecommunications Institute of Aveiro (ITAv)