Project Details
Autonomous Shuttle Service Platform
A production-grade backend platform combining secure APIs and real-time fleet streaming to power autonomous shuttle operations.
completedLast updated Feb 8, 2026
What It Does
- Delivers authenticated APIs for mobile and web clients plus admin tooling for operational management.
- Aggregates live vehicle, telemetry, passenger-seat, and announcement data and broadcasts updates in real time.
- Combines queue-driven batch workflows with event-driven streaming for reliable large-scale transport operations.
Architecture Highlights
- Modular CMS plugin boundaries keep backend domain concerns isolated and maintainable.
- Single real-time streaming service is designed for horizontal scalability and reliable fan-out.
- Snapshot + delta delivery model enables low-latency dashboards while preserving consistency.
- Queue-backed batch imports and idempotent processors improve resilience under load.
Key Features
Core Backend API
- JWT authentication with token login, refresh, and invalidation flows
- Multiple login methods including credential-based and card-based identity
- User profile retrieval, updates with media uploads, account removal, and verification handling
- Passenger provisioning with identity-based creation/removal and time-bound verification
- Public endpoints for vehicles and schedules, plus multilingual admin announcements
- Role-based admin access with restricted editor permissions
Real-Time Stream Service
- Live WebSocket API with full initial snapshot and incremental delta updates
- Real-time ingestion of location, speed, battery, temperature, charging, and seat-state events
- Database change streaming for announcements without polling
- Normalized relational persistence with auto-broadcast hooks after upserts
Operational Reliability
- Event-driven processing across WebSockets, queues, and database events
- Idempotent ingestion and synchronization to prevent duplicate or stale state
- Threshold-based critical battery alerts with email notifications
Architecture Flow
- Ingest: listeners consume real-time feeds, queues, and database events
- Process & Store: data is normalized and upserted into relational storage
- Broadcast: clients receive instant updates plus a full state snapshot on connect
Tech Stack
Backend APIs
- PHP 8
- Laravel 9
- WinterCMS
- JWT authentication
- MySQL
- Laravel queues and caching
- PHPUnit, PHPCS, CI workflows
Real-Time Streaming
- Node.js
- Express
- Socket.io
- MQTT
- RabbitMQ
- MySQL / MariaDB
- Sequelize
- MySQL binlog event streaming
- Email alerting system
Performance & Product Notes
- Built for high-frequency ingestion and continuous real-time client updates.
- Event-driven pipelines minimize polling overhead and reduce end-to-end update latency.
- Operational safeguards via alerts, queues, and deterministic processing improve reliability.