Our cars have become more environmentally-friendly and low-cost. They have GPS systems that have every road mapped down. They have cameras that help you park (or can even park without needing your input). But in terms of safety, we have a long way to go.
Car-crashes currently cost about 3 percent of the world’s GDP and are the leading cause of death in children and young adults.
Yes, autonomous vehicles will greatly increase safety… but there are other things we can do until they go mainstream to protect our drivers.
This new deep learning model can help prevent traffic crashes by analysing crash history and flagging high-danger areas. It also finds similar roads and structures across different countries, to generalise the data and prevent future crashes. Such data can be shared with city-planners to redesign roads or add safety measures to these areas. It can also be shared with Google Maps or other applications so that they do not recommend high-danger routes to their drivers. And finally, it can be used to assist autonomous vehicles!
Click below to read more about this.