Industrial IoT Systems with Android and Kotlin Multiplatform - DroidKaigi 2025

Explore how Android and Kotlin Multiplatform can power industrial-grade IoT systems. This talk covers hardware connectivity patterns (USB, Wireless, Web Serial API), KMP architecture for code sharing across platforms, and real-world implementation strategies for RFID systems and industrial applications. Learn about reliability patterns, offline sync, zero-trust security, and practical deployment considerations for enterprise environments.

Presented at DroidKaigi 2025 (September 11, 2025) in Tokyo, Japan.

Topics covered:
- Android hardware capabilities for industrial use
- Web Serial API as a bridge for hardware connectivity
- Kotlin Multiplatform architecture patterns
- Real-time data handling with Kotlin Flow
- Zero-trust security in industrial systems
- Deployment strategies for enterprise

GitHub: https://github.com/jsxs0
X/Twitter: https://x.com/SJasveen

Transcript

1. INDUSTRIAL IOT SYSTEMS with Android & Kotlin Multiplatform Jasveen Sandral • DroidKaigi 2025
2. ABOUT ME Cross-Platform Engineer Industrial IoT Specialist Technical Speaker
3. INDUSTRIAL CHALLENGE Legacy systems everywhere Vendor lock-in Need for modernization
4. Android's Potential Beyond consumer apps
5. WHY ANDROID? • Mature ecosystem • Hardware availability • Developer familiarity • Cost efficiency
6. TODAY'S TOPICS 1. Android's hardware capabilities 2. Connection approaches 3. Kotlin Multiplatform details 4. Implementation patterns 5. Practical considerations
7. CONNECTION OPTIONS USB - Direct connection Wireless - Bluetooth/WiFi Hybrid - Web technologies
8. USB CONNECTION PATTERN Direct hardware access // Conceptual example only class USBManager { fun findDevices(): List<device> { // Discovery logic } fun connect(device: Device) { // Connection logic } }</device>
9. WIRELESS PATTERN Flexible deployment // Conceptual example only class WirelessManager { suspend fun scan(): List<device> { // Scanning logic } suspend fun connect(address: String) { // Connection logic } }</device>
10. WEB TECHNOLOGY BRIDGE Experimental approach // Conceptual pattern class WebBridge { // JavaScript bridge concept // Requires WebView // Limited platform support }
11. WEB SERIAL ARCHITECTURE Android App ↓ WebView ↓ JavaScript ↓ Web Serial API ↓ RFID Reader
12. WHAT IS WEB SERIAL API? • Browser API for serial devices • Released in Chrome 89 (2021) • Requires user permission • Works with USB/serial ports
13. BROWSER SUPPORT ✓ Chrome/Edge (desktop) ✓ Chrome Android (via flag) ✗ Firefox (no plans) ✗ Safari (no support) Limited but growing support
14. WEB SERIAL USE CASES • Arduino programming • IoT device configuration • Industrial sensors • Educational hardware
15. WEB SERIAL LIMITATIONS • Requires HTTPS • User gesture required • No background access • Platform restrictions
16. WEB SERIAL VS NATIVE USB WEB SERIAL + Cross-platform + No app install - Browser only - Limited support NATIVE USB + Full control + Always available - Platform specific - App required
17. BASIC WEB SERIAL PATTERN // JavaScript example async function connectSerial() { // Request port access const port = await navigator.serial.requestPort(); // Open with standard settings await port.open({ baudRate: 9600 });
18. PERMISSION MODEL 1. User action required (click/tap) 2. Browser shows device picker 3. User selects device 4. Permission granted for session
19. THE BRIDGE: JS TO KOTLIN RFID Reader (Serial) ↓ Web Serial API (JavaScript) ↓ JavaScriptInterface ↓ Android App (Kotlin) ↓ KMP Shared Logic
20. BRIDGE IMPLEMENTATION // androidMain: Expose a Kotlin function to WebView's JS class WebAppInterface(private val onTagRead: (String) -> Unit) { @JavascriptInterface fun postTagData(hexData: String) { // Data from RFID reader is received here onTagRead(hexData) } } // In your Activity/Fragment
21. KOTLIN MULTIPLATFORM Share business logic Platform-specific UI Single codebase
22. KMP BENEFITS • Code reuse across platforms • Type safety • Native performance • Gradual adoption possible
23. KMP ARCHITECTURE // commonMain - shared code expect class DeviceManager { fun initialize() fun connect() } // androidMain actual class DeviceManager { actual fun initialize() { /* Android */ } actual fun connect() { /* Android */ }
24. KMP PROJECT STRUCTURE commonMain/ - Shared logic androidMain/ - Android specific iosMain/ - iOS specific jsMain/ - Web specific
25. SHARED RFID LOGIC // commonMain: Parse raw data into a typed object data class RFIDTag(val id: String, val type: String) class RFIDParser { // Shared logic to parse a tag from raw hex data fun parseTagData(rawHex: String): RFIDTag? { if (!isValid(rawHex)) return null // ...logic to parse SGTIN or other formats return RFIDTag(id = "...", type = "UHF")
26. PLATFORM SPECIFICS // androidMain actual fun getPlatformName(): String = "Android" // iosMain actual fun getPlatformName(): String = "iOS" // jsMain actual fun getPlatformName(): String = "Web"
27. SYSTEM ARCHITECTURE Application Layer ↓ Business Logic (KMP) ↓ Hardware Abstraction ↓ Device Layer (Web Serial)
28. RELIABILITY PATTERNS // Retry pattern concept suspend fun reliableOperation() { repeat(MAX_ATTEMPTS) { try { // Operation return } catch (e: Exception) { // Handle error } }
29. OFFLINE SYNC PATTERN [ Queued Locally ] ↓ [ Syncing ] ↓ [ Synced ] ✓ [ Failed ] ✗
30. STATE MANAGEMENT // Shared state management class StateManager { private val _state = MutableStateFlow<state>() val state: StateFlow<state> = _state fun updateState(newState: State) { _state.value = newState } }</state></state>
31. HANDLING REAL-TIME STREAMS A fast RFID reader can produce hundreds of reads per second. How do you prevent overwhelming the UI? // Use Flow operators to manage the stream without blocking rfidReader.dataFlow() .buffer(capacity = 64) // Absorb bursts of data on a background thread .conflate() // Drop intermediate reads if the UI is busy .onEach { tag -> // Only the latest tag is sent to the UI for rendering viewModel.updateLatestTag(tag) } .launchIn(viewModelScope)
32. UI WITH COMPOSE @Composable fun IndustrialDashboard() { Column { StatusIndicator() RFIDTagList() ControlButtons() } }
33. TESTING STRATEGY • Unit tests in commonTest • Platform tests in specific modules • Integration testing • UI testing per platform
34. ZERO TRUST SECURITY Principle: Never Trust, Always Verify Device Identity: Each device authenticates with a unique certificate. Per-Request Auth: Every API call is authenticated with a short-lived token. Assume Breach: Network is hostile; all traffic encrypted (TLS 1.3). Data Validation: Raw data is untrusted until validated by KMP logic.
35. DEPLOYMENT OPTIONS • MDM for enterprise • Private app stores • Kiosk mode devices • OTA updates
36. IMPLEMENTATION APPROACH Phase 1: Requirements analysis Phase 2: Architecture design Phase 3: Prototype development Phase 4: Testing & refinement
37. COMMON CHALLENGES • Device compatibility • Network reliability • Battery optimization • Update management
38. SOLUTIONS • Thorough testing matrix • Offline-first architecture • Power management APIs • Staged rollouts
39. RECOMMENDATIONS • Start with proven APIs • Focus on reliability • Plan for offline scenarios • Test thoroughly
40. INDUSTRIAL APPLICATIONS MANUFACTURING Process monitoring Quality control Asset management LOGISTICS Tracking systems Inventory control Fleet management
41. FUTURE CONSIDERATIONS • 5G connectivity • Edge computing • AI/ML integration • AR/VR interfaces
42. KEY POINTS • Android is capable for industrial use • KMP enables code sharing • Web Serial provides a hardware bridge • Reliability must be designed in
43. LEARN MORE • Android developer documentation • Kotlin Multiplatform resources • Industrial IoT best practices • Web Serial API on web.dev
44. THANK YOU Questions & Discussion Exploring Android's industrial potential