This project, developed by a team of two, focuses on creating a machine learning model for self-driving cars. The objective is to design a robust and reliable system capable of navigating various driving scenarios autonomously.
- Data Collection and Preprocessing:
Utilization of diverse datasets including camera images, LIDAR data, and GPS coordinates. Implementation of data augmentation techniques to enhance model robustness.
- Model Architecture:
Development of a convolutional neural network (CNN) for image recognition and object detection. Integration of recurrent neural networks (RNNs) for sequential data processing and decision making.
- Sensor Fusion:
Combining data from multiple sensors to create a comprehensive understanding of the vehicle's environment. Implementation of Kalman filters for precise tracking and estimation.
- Control Algorithms:
Development of algorithms for path planning and trajectory optimization. Implementation of feedback control systems to ensure smooth and safe vehicle maneuvers.
- Simulation and Testing:
Utilization of simulation environments such as CARLA and Gazebo for extensive testing and validation. Deployment of the model on hardware setups for real-world testing and evaluation.
- Safety and Redundancy:
Incorporation of fail-safe mechanisms and redundancy to handle unexpected scenarios. Continuous monitoring and adaptation to ensure passenger and pedestrian safety.
-
Technologies Used: Programming Languages: Python Frameworks and Libraries: TensorFlow, Keras, PyTorch, OpenCV, ROS (Robot Operating System) Tools: CARLA Simulator, Gazebo, Jupyter Notebooks
-
Road Signs Detection Accuracy of Detecting signs in 94%
A functional self-driving car model capable of navigating complex environments. Extensive documentation and analysis of model performance and limitations. Contributions to the field of autonomous vehicle technology through research and innovation. This project demonstrates a collaborative effort in advancing the capabilities of autonomous driving systems, pushing the boundaries of what is possible in the realm of machine learning and artificial intelligence.