Fight Red Lights

In the street cycling world, motors are no concerning to worry about, the traffic jam means nothing to a rider with two sleeky wheels. The only thing that a rider worries is a red light.

If only we can have all green-light traffic, I don't care about none other.

So here's the thought.

A rider can avoid the red light by konwing when it will come and how fast he should go. By using Deep Learning, a machine can estimate whether you'll hit the next red light by your current location and speed. Normally if you want to calculate that for cars, you need to take traffic into consideration, but for cyclist, things are purer and easier.

Considering that the data for traffic lights are not obtainable

1. If the lights are determistic with the daytime. The machine learning process goes like this: every time on my commute route, I will hit the red light button when I met one as the feedback. The speed / location should be the only input arguments, and with the accumulation of the experiments, this app should give a realtime feedback of whether I'll encounter the next red light.

2. But if the red light are not determistic with the clock, the solution should change to use the first red light that I hit as a starting hint and calculate the ramaining red lights. This mehods should be more useful in a tight traffic route with more lights

--

Idea inspired by Youtube Hotline Videos https://www.youtube.com/channel/UCHRUAMAzVUS_Szvxn55GXaQ/videos?view=0&sort=p&flow=grid