poutnikl
commented
2 years ago Speaking just for myself, not the developers:
I am not sure if the gain is worthy of the development effort. I guess more practical could be to go the same way as for human powered bicycles, where most bikers do not know their sustainable power either.
There can be used tuning of the effective power to obtained the predicted ETA close to the real ETA for well known routes and driving scenarios. Then using it as a statistical average for such a combination, which would include also typical driving breaks for whatever reasons.
I use such an approach personally for long routes on loaded bicycle with economic pedalling. It gives ETA looking quite pessimistic to my optimistic brain, but realizing at the end the estimation of my own profile was much more accurate than from my brain in the field mode.
In Germany simple assisted bikes called Pedelec are allowed to help cyclists up to 25 km/h. Looking at code of some profile it looks like the power is a global value. But what would be required is something like: Depending on elevation profile sum human power and electric power but if speed is beyond 25 km/h limit to human power again which means when elevating limit speed to 25 km/h. Some bikes have a more complicated profile cause power they can output depends on speed. Some front wheel engines have most power at round about 12 km/h ?
But brouter is really well done