Every year millions of people die and commit accidents due to unsafe driving behaviors such as overspeeding, irregular lane changing, not following the traffic signals and many more. Right now, there is no platform to monitor driving behaviors and we hope to fill the hole by providing ways to make the road safe for everybody.
What it does
Using True AR SDK we want to increase awareness of traffic rules and regulations for the drivers by analyzing their driving patterns on the roads and using their historic behavior to create their driving profile and scores. The violations would be penalized and the good behaviors would be rewarded thereby driving them towards good driving patterns.
Helping People to Drive Better The primary users for this would be general public who would be benefited the most as good driving behaviors can be embedded via this.
Automating Driving Tests Driving tests can be automated and the driver’s behavior can be monitored over a period of time to evaluate them.
Cab Driver performance Monitoring Monitoring Cab drivers to ensure the safety of the passengers and giving them appropriate training based on the scores.
Insurance companies These scores can be used by auto insurance companies to evaluate risk and charge monthly premiums accordingly. This could also be used to establish fault for accidents.
How we built it
It was mainly built using the Android Studio using the TrueAR SDK. We started looking at the documentation of the SDK , listed out all the capabilities which it currently provides and started leveraging those to communicate our ideas. We had 5 main features which utilizes the objects in the simulator and distributed tasks among ourselves. All the programming is done using c++ including features like exporting the simulator data to external text files. The 3D models and assets were built on Sketch and Blender and them imported into the android studio. The Admin dashboard is built using HTML5, CSS.
Challenges we ran into
SDK: The biggest challenge for the hackathon was the SDK as it is just 1 month old and still evolving, there is not really much you can accomplish with the features currently available.
Accomplishments that we are proud of
- We are really proud of making the drivers aware of the traffic rules and ensuring that they follow them via our experience.
- We were successfully able to export the driving behaviour scores from the simulator to a text document which is vital. 3.We were able to detect improper parking and alert the drivers accordingly and there is no technology right now which is capable of doing that.
What's next for AccuDrive
Implementing more complex features such as sharp turns, lane discipline, detection of crashes, one way roads, improper passing to name a few.
Leveraging machine learning to detect roads and locations where drivers are most likely to commit an accident based on previous data collected through the SDK.