Md Nazmos Sakib

An Android app for jogging a five-axis robot arm using touch controls or body gestures. Built with Kotlin and Jetpack Compose, it connects over BLE to an ESP32-S3-ETH bridge, which exchanges commands and status with a B&R X20 PLC over Modbus TCP.

What I learned

• Connecting Android, an ESP32 microcontroller, and a PLC into a complete robot-control system, with live status flowing back to the app.
• Working with BLE GATT notifications and buffering newline-delimited commands across fragmented Bluetooth packets.
• Bridging BLE commands to Modbus TCP registers over Ethernet and coordinating power, homing, and jogging with a PLC state machine in B&R Structured Text.
• Using CameraX, ONNX Runtime, and a YOLO pose model to turn body keypoints into axis selection and directional jog commands.
• Handling stop-on-release, communication timeouts, heartbeat watchdogs, and readiness checks across the app, firmware, and PLC.

Watch the two robot-control demos · Project details

Architecture

• Kotlin and Jetpack Compose UI with ViewModels, domain repository interfaces, and Hilt dependency injection.
• Room stores notes and content blocks locally; Kotlin Flow exposes changes through the repository layer.
• Optional Parse synchronization uses an outbox to track local changes while keeping notes available offline.

Development takeaways

• Separating UI state, domain models, and storage responsibilities in a growing Android app.
• Modeling text, checklists, tables, images, audio, and drawings as note content blocks.
• Managing local transactions, account-scoped data, and queued synchronization.
• Testing repository behavior and Room database migrations.
• Account authentication and cloud synchronization.
• Backend development with Parse Server and self-hosting.
• Docker and Microservice: hosting backend in a Raspberry PI with Docker
• Data protection, backups, and recovery.
• Setting up tunnel behind CGNAT

Checkout github

Development takeaways

• On-device AI: Whisper transcription and offline translation.
• Audio processing: Recording, decoding, playback, and resampling.
• App architecture: ViewModels, StateFlow, SQLite, and background services.
• Reliability: Pause/resume, crash recovery, backups, and secure downloads.
• Quality assurance: Unit tests, UI tests, debugging, and performance checks.

Checkout github

An Android app that receives pressure readings from two sensors connected to an ESP32 over Classic Bluetooth (SPP). It displays head and body readings during a monitoring session and stores sensor data locally with Room.

Architecture

• Kotlin and Jetpack Compose UI, organized with MVVM and Clean Architecture, with Hilt for dependency injection.
• Android Bluetooth APIs handle device discovery and connections; Coroutines and StateFlow coordinate asynchronous data and UI updates.
• ESP32 firmware reads two analog inputs and sends comma-separated values over Bluetooth serial; Room persists readings on the phone.

What I learned

• Connecting pressure-sensing hardware to an Android app through an ESP32 and Classic Bluetooth.
• Managing Bluetooth permissions, device discovery, and serial data transfer.
• Keeping the Compose interface responsive while receiving sensor readings asynchronously.
• Separating presentation, domain, and data responsibilities with MVVM, repositories, and dependency injection.
• Storing sensor data locally with Room and working with analog readings in ESP32 firmware.

View source code and project documentation on GitHub

What I learned

• Android permissions, UI state management, Navigation and activity lifecycles.
• Bluetooth communication and Asynchronous programming.
• Local storage with Room and cloud storage with Firebase.
• Local and cloud synchronization.
• Application architecture and Dependency injection with Hilt

Checkout github

What I have learned

What I have learned

What I have learned

What I have learned

What I have learned

What I have learned

What I have learned

What I have learned

What I have learned

What I have learned

What I have learned

What I have learned

See the Android app, ESP32 bridge, and B&R PLC working together to control the five-axis robot arm.