RashCam
2021-2022
Abstract
Research by WHO[1] shows that a vast majority of road accidents are due to overspeeding or careless driving, moreover rash driving is a major cause of deaths among people of 5-29. Based on this simple yet thought-provoking research we propose a smart dash cam that can detect rash driving patterns and notify the owner. For rash driving detection, Rashcam uses the state of the art sensors to provide full insights into any useful incident. This dashcam when used properly can significantly reduce rash driving and other bad driving behaviors on-road and thus making road travel much safer.
Background
From the safety point of view, reckless driving is one of the most challenging and risky things that make headlines every year. It is also one of the things that parents tremble at when it comes to their children driving recklessly despite their advice not to do so. Drivers who are reckless when overtaking other cars turn into a danger to themselves and to others. There have been many road accidents recorded all over the world and even in some cases, fatal crashes occurred because of this kind of road behavior. When it comes to reckless driving, transportation companies, delivery service providers, and vehicle rental companies face financial expenses as well as the life of the driver if not the lives of the innocent people on the road.
Different companies use special tactics to overcome this issue which includes installing dash cameras or providing a public feedback number with the caption “how am I driving?”. These approaches come with their downsides as dashcam footage must be sought manually to detect driving events and public feedback numbers are not as trustworthy and cannot reflect the actual intensity of the reckless driving. It has been observed that most reckless drivers don’t own the vehicle. In which case, a simple automated reporting mechanism can drastically reduce the reckless driving patterns on the road.
We propose a smart dashcam equipped with state-of-the-art yet cheap and easily available sensors that can detect bad driving behaviors like speeding, aggressive driving, tailgating & hard cornering. With very little computational power we can easily detect these bad driving behaviors with astonishing accuracy. A huge amount of storage is required to store good quality videos which cost more storage space. To smartly manage, the storage RashCam has a built-in algorithm that can label different events with relevant importance scores. These scores can be used to remove less important footage in case of low storage, thus devices with a very small amount of storage can still accommodate important events for a long period of time.
Introduction
RashCam is a smart dashcam equipped with state-of-the-art yet cheap and easily available sensors that can detect bad driving behaviors like speeding, aggressive driving, tailgating & hard cornering. And notify the vehicle owner regarding these events along with video clip from the dashboard of the vehicle.
Hardware components
Raspberry Pi 4
Pi Camera Module
Inertial Sensors
Magnetometer
GPS Module
Features
Users can view live dashcam footage securely from any device.
The system provides detected reckless events with video evidence.
The user gets notified in case of any reckless driving detected.
Users can see the last known location in case the device is offline.
We can provide over-the-air updates and technical support for our devices.
A 3d model of prototype device we developed
Device Architecture
The device architecture is inspired by pipelines architecture which is shown in the diagram.
Overall System Architecture
Procedure
A new user will go through following steps to step up his/her newly brought rashcam device.
Constraints
The device should be strictly mounted in the vehicle
At least 5V 3A to power is required for the device
Although detection can happen even offline, a stable internet connection for live streaming
For now, we have very low-end GPS module which can only calibrate in open areas
Future Works
Specialized Hardware
Mobile Application
Store metadata along with incidents
Integrate computer vision
Fleet Management
Data export options
Integration with on-board diagnostics
A futuristic 3d render of our rashcam.
Team Members
Dr. Usama Ijaz Bajwa
Tenured Associate Professor
Department of Computer Science, COMSATS University Islamabad, Lahore Campus, Pakistan
Ameer Hamza Naveed
References
World Health Organization. “Road traffic injuries.” 21 June 2021, who.int/news-room/fact-sheets/detail/road-traffic-injuries.