Home - Tauvic/DriverAwareness GitHub Wiki

Note: Work on this project started in november 2020 and is incomplete and under construction

Of all car accidents 94 percent are caused by driver error and 57 percent by distracted driving. A system that helps the driver to focus its attention on driving can reduce injuries and damages.

The driver coach system will check for 3D aspects:

• Distractions (eating, drinking, phone usage)
• Drowsiness (tired driver warning)
• Driving Maneuvers (errors, traffic rule violations, situational awareness, mirror checking)

The objective is to make driving safer by developing an affordable personal driving coach. Therefore I open sourced this plan and its code and hope others including companies and organisations will step in and adopt it.

On this page you will find lots of information brought together from various sources that will help you to understand the problem and whats required to develop a working solution. I have started to work on this project in november, intend to make a proof of concept device that i can demonstrate and at the same time make contact with several organisations to get them interested in adopting this project.

Inspirational video of some guys who developed their first proof of concept video do not take this very serious!! But use it as an example to learn that it does not have to be very complicated. source

Table of contents:

• Problem description
  • Statistics
  • Effectiveness of proposed solution
• Problem analysis
  • Distractions
  • Drowsiness
  • Driving Maneuvers
  • Evaluation and Scoring
• Requirements
  • Functional requirements
  • Non functional requirements
  • Security, Privacy and Legal
• Data collection
• Development environment
  • Software stack
  • Hardware
  • Development tools
  • Deployment
  • Test
• Design
• Model Training
• Proof of concept

Secret strategy:

• Learning by doing
• Copying others ideas when required Nokia Driver behaviour analytics
• Finding a solid software stack
  • Quality of software/hardware
  • Well supported / documented
  • Ease of implementation
  • Will last for several years

Examples:

• https://www.youtube.com/user/embeddedvision
• https://github.com/yinguobing/facial-landmark-dataset
• https://github.com/shashwat623/Computer-Pointer-Controller-With-Face-Authentication
• https://github.com/nullbyte91/Driver-Monitoring-System-Gaze_Detection
• https://github.com/incluit/OpenVino-Driver-Behaviour

EU Road Safety