Distracted Driver Detection

For the last few years there is a continuous increase in the number of road accidents worldwide. According to world health organization (WHO) nearly 1.25 million deaths annually are due to road accidents and nearly one out of five accidents are mistakes of distracted drivers. We worked to develop an

Project Title

Distracted Driver Detection

Project Area of Specialization

Artificial Intelligence

Project Summary

For the last few years there is a continuous increase in the number of road accidents worldwide. According to world health organization (WHO) nearly 1.25 million deaths annually are due to road accidents and nearly one out of five accidents are mistakes of distracted drivers. We worked to develop an accurate system for detecting distracted driver and warn him against it. We used Convolutional Neural Network (CNN) to solve this problem. We designed a CNN based system that detects the likely hood of the driver and warns him. The system consists of genetically weighted entities of convolutional neural networks, we show that a weighted entity of classifiers using genetic algorithm yields a better classification result.

Project Objectives

To detect the likelyhood  of the driver while he is driving and to provide the helping hand by alerting the driver if he losses his primary focus from the road. This can be done using 2D dashboard camera. There are ten gestures included one of them is safe driving while the remaing nine gestures are distractions. This why we can reduce injuries on the road and major economic losses caused due to accidents. 

Project Implementation Method

Images of the dirvers are used to train the CNN model later on the trained model was load on the embedded board. Frame taken at any moment from live video stream can be manipulated and the scores are generated. These scores are further used to make prediction and alert the driver.

Benefits of the Project

Driver involved in activities e.g. (Calling, makeup) can be detected. These drivers can be saved  from major accident and economic loss. This is an intelligent system and a step toward advance cars.

Technical Details of Final Deliverable

There are ten classes 

Safe Driving

Makeup

Radio

Calling Right

Calling Left

Texting Right

Texting Left

Reaching behind

Talking Left

Drinking

Overall 80% accuracy is acheived at the moment and the system is in running state. CNN alogorithm was used to developed classifier in python and the embedded board used for the purpose is Raspberry pi . While Pi camera was used for streaming. 

Final Deliverable of the Project

HW/SW integrated system

Type of Industry

Manufacturing , Transportation

Technologies

Artificial Intelligence(AI), Internet of Things (IoT)

Sustainable Development Goals

Industry, Innovation and Infrastructure, Sustainable Cities and Communities

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com