A powerful, emotion-aware mobile assistant robot built for real-world intelligence, research, interaction, and multi-domain AI applications.
ASTRA (Autonomous Smart Tech Robotic Assistant) is designed to be a modular, scalable robotic system combining computer vision, real-time ML, intelligent mobility, and emotional intelligence to interact with the physical world in an intuitive, human-centered way.
The MVP serves as a proof-of-concept platform for robotics + AI/ML + embedded systems + HCI research and innovation.
| Component | Description |
|---|---|
| Main Controller | NVIDIA Jetson Orin Nano Super (8-core ARM Cortex + 20 TOPS AI) |
| Support Boards | Raspberry Pi 5 (Sensor I/O & lightweight ML), Arduino Mega (actuators), ESP32 (wireless, BLE, sensor fusion) |
| Camera | Intel RealSense D435i (Depth + RGB) |
| Lidar | RPLidar A1/A2 (for mapping & navigation) |
| Display | 64x64 LED Matrix or OLED screen for Emotion UI |
| Mobility | 2WD or 4 Omni-wheel base (with geared motors + motor drivers) |
| Sensors | Ultrasonic, IMU (MPU6050), IR, ToF, GPS, Compass |
| Connectivity | GSM module (SIM800L), WiFi (ESP32 + Jetson), GPS (Neo6M) |
| Power System | 12V Li-ion battery (8000–10000mAh), 5V and 3.3V regulators |
| Chassis | 3D printed + lightweight aluminum/ABS custom frame |
| Optional | 6-DOF Robotic Arm (MG996R / Dynamixel motors) |
| Item | Approx. Cost |
|---|---|
| Jetson Orin Nano Super | ₹25,000–28,000 |
| Raspberry Pi 5 | ₹7,000 |
| Arduino Mega + ESP32 | ₹1,500 |
| Intel RealSense D435i | ₹15,000–20,000 |
| RPLidar A1 | ₹10,000 |
| GSM + GPS Modules | ₹1,500 |
| Emotion Display (LED/OLED) | ₹2,000 |
| Motors + Wheels + Drivers | ₹4,000–6,000 |
| Battery + BMS + Power Mgmt | ₹3,500–4,500 |
| Chassis & Structure | ₹8,000–10,000 |
| Misc Sensors, Wiring, PCB | ₹3,000 |
| Total | ₹90,000–1,30,000 |
About : 1,50,000 INR
| Layer | Tools / Languages / Frameworks |
|---|---|
| OS & Middleware | Ubuntu 20.04 LTS (Jetson), Raspberry Pi OS |
| Languages | Python, C++, Bash, Arduino (C), JavaScript (web UI) |
| AI/ML | PyTorch, TensorFlow Lite, OpenCV, YOLOv8, ONNX |
| Robotics | ROS2 Foxy, RPLidar SDK, Jetson GPIO |
| CV & Perception | RealSense SDK, DepthAI, MediaPipe |
| Voice Control | Vosk / Whisper / Google STT |
| Emotion Engine | Custom expression-to-LED rendering logic |
| Cloud/Comms | MQTT, WebSocket, GSM API, GPS tracker |
| Web UI (Optional) | Flask / Node.js dashboard for remote monitoring |
- ✅ Real-time Object Detection & Tracking (YOLOv8)
- ✅ Emotion Display System (LED Matrix)
- ✅ Voice Command Execution (offline-capable)
- ✅ Mobility & Navigation with Lidar & Depth Cam
- ✅ Environment Mapping & Obstacle Avoidance
- ✅ Sensor Fusion with IMU + GPS + GSM
- ✅ ROS2 Integration for modular system design
- ✅ Data logging, remote diagnostics (via web dashboard)
- ✅ Modular design (sensor/arm additions possible)
- 🟢 Wall-E’s compact & expressive form
- 🔵 NVIDIA's humanoid research robot (GTC 2025) for futuristic, emotion-aware presence
- ♻️ Designed for modularity, extensibility & research-grade experimentation
| Part | Size (approx) |
|---|---|
| Height | 60 cm |
| Width | 35 cm |
| Depth | 30 cm |
| Wheel Base | 25–28 cm |
| Emotion Display | 10x10 cm |
| Arm Length (opt) | ~25–30 cm |
- 🤖 Robotics & Mechatronics
- 🧠 AI/ML & Real-time Inference
- 📍 SLAM & Autonomous Navigation
- 👁️ Computer Vision & Emotion Recognition
- 🗣️ Human-Robot Interaction (HRI)
- 🌍 Edge Computing & IoT Fusion
- 🧩 Multi-Agent Coordination (future expansion)
- Integrate GPT-4 Vision or LLaVA-style perception models
- Add autonomous path planning & swarm logic
- Upgrade to custom-built PCB for cleaner internals
- Publish white paper on ASTRA’s system design
- Open-source ROS2 packages for education/research
- Professors & Researchers in AI/Robotics
- Labs like MIT CSAIL, Stanford AI Lab, ETH Zurich ASL
- Hackathons, Expo Showcases
- Open-source robotics & edge-AI community
M4YH3M-DEV
🔗 GitHub
🧪 BTech '28 | AI/Robotics/OS/LLMs | Building for the Future