Project Video
Robotics
[Jetank AI Obstacle Avoidance Robot]
An AI-powered autonomous Jetank robot trained using computer vision and deep learning to detect obstacles and navigate intelligently using a camera-based obstacle avoidance system.
Overview
This project is an autonomous obstacle avoidance robot developed using the Jetank robotic platform, computer vision, and deep learning techniques. The robot uses its onboard camera to capture environmental data and make intelligent navigation decisions in real time.
A custom image dataset was collected directly from the Jetank camera during different navigation scenarios. The dataset included images representing safe forward movement as well as obstacle conditions requiring directional changes such as turning left or right.
Using the collected dataset, a convolutional neural network (CNN) model was trained with TensorFlow to classify navigation actions based on live camera input. The trained model enabled the robot to autonomously avoid obstacles and navigate indoor environments without requiring manual control.
Features
- AI-based obstacle avoidance
- Real-time camera processing
- Autonomous navigation
- Deep learning image classification
- CNN-based movement prediction
- Dataset collection and preprocessing
- Real-time decision making
- Intelligent directional control
How It Works
The Jetank robot continuously captures images using its onboard camera while moving through an environment.
A custom dataset was created by manually collecting and labeling images into different navigation categories such as:
- Move Forward
- Turn Left
- Turn Right
- Obstacle Detected
The dataset was then preprocessed and used to train a deep learning model using TensorFlow and CNN architectures.
During runtime, the trained model processes live camera frames and predicts the appropriate movement action. The robot then executes navigation commands based on the prediction output, allowing it to avoid obstacles autonomously.
Technologies Used
- Python
- TensorFlow
- Convolutional Neural Networks (CNN)
- OpenCV
- Jetank Robot Platform
- NVIDIA Jetson Nano
- Ubuntu Linux
- Camera-Based Vision System
- Deep Learning
- NumPy
Applications
- Autonomous robotics
- Smart navigation systems
- AI-powered mobile robots
- Indoor autonomous vehicles
- Educational robotics
- Computer vision research
- Intelligent obstacle avoidance systems
Learning Outcomes
This project strengthened skills in:
