Accident Warning System
Safe Drive is an accident warning system that generates a warning to the user if the particular vehicle or object is close to our vehicle and it is in our trajectory. Safe Drive is an application-based product run on a dev board or in mobile that can be used easily on a mid-range phone having
2025-06-28 16:30:07 - Adil Khan
Accident Warning System
Project Area of Specialization Artificial IntelligenceProject SummarySafe Drive is an accident warning system that generates a warning to the user if the particular vehicle or object is close to our vehicle and it is in our trajectory.
Safe Drive is an application-based product run on a dev board or in mobile that can be used easily on a mid-range phone having a camera of at least 12 (or more) megapixels or a Coral Dev board with the camera. The mobile or dev board is mounted on the dash of the car or bike, the application detects any objects which come in front of the car or bike (through the rear camera of the phone) and generates a warning alert to the driver about the possible collision threat.
Project ObjectivesSafe Drive is a product that is used by people who ride a bike or drive a car. Safe Drive is used in any type of car or bike all you need is one mid-range phone and good connectivity(for higher accuracy in collision prediction ). It provides a warning to the user when a collision is possible to happen. The Safe Drive also tells the user if there are some hurdles are coming on the road it warns the user early so they take action according to that.
Safe Drive Application is trained on a custom dataset so it’s familiar with the environment and condition in Pakistan. So what's the most unique thing in our Application right? You might be thinking Tesla Motors or some other automobile sector giant‘s also do the same thing but they are limited to their manufactured vehicles only and they also charge an additional amount for this feature. Safe Drive’s most significant and relevant thing is that it is easily used with any brand vehicle because it’s an application that provides you all safety parameters while driving a car or riding a bike.
Safe Drive main objective is as follow:
- detect and track particular objects which are in our camera frame
- create bounding boxes around that object
- calculate the relative distance from our vehicle to the particular object.
- check that particular object is our trajectory or not
- decide based on the relative distance, trajectory, and speed of our vehicle that either is a generate warning or not.
The most important part of our application is that how we calculate the relative distance between our vehicle and the potential object. We calculate the distance that either is the particular object is far or near on the basis of the amount of pixel taken by the object on screen for trajectory purposes we used the midpoint of the particular object to check that is in the trajectory or not.
For image classification and localization we used Single Shot Detector which is base on Convolution Neural Network.
We deploy this model on mobile and google coral dev board for the user end.
Safe Drive main objective is to provide a cost-effective solution many automobile companies can provide a warning alert mechanism to prevent a collision but Safe Drive main advantage is that is not any company dependant automobile companies which provide these feature can charge a huge amount on the basis of this feature safe drive cut this amount and used in any brand automobile vehicle.
Technical Details of Final DeliverableOn technical aspects of we used TensorFlow Object Detection API, for the object detection model we used SSD Model on the backbone of SSD we used VGG-16 In the backbone and the header (Dense Layer) we used FPN to increase accuracy. For deployment purposes, we used TensorFlow lite on android. For third parties libraries, we used Google Map and Google Coral dev board.
Final Deliverable of the Project HW/SW integrated systemCore Industry ITOther Industries Education , Transportation Core Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Industry, Innovation and InfrastructureRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 70000 | |||
| Google Carol Dev Board | Equipment | 1 | 50000 | 50000 |
| Google Carol Camera | Equipment | 1 | 15000 | 15000 |
| SD CARD 8 GB | Equipment | 1 | 1000 | 1000 |
| C type USB Cable | Equipment | 1 | 2000 | 2000 |
| HDMI to USB | Equipment | 1 | 2000 | 2000 |