search

aesculus / EVTO-App-Feedback

A project to track bugs and ideas for the EVTO App
MIT License
1 stars 0 forks source link

[Basic Tutorial] From the Frunk: The Origin Story #347

Closed EVGrokker closed 6 years ago

EVGrokker commented 7 years ago

This is a placeholder

EVGrokker commented 7 years ago

@aesculus:

aesculus commented 7 years ago

What motivated you to start developing EVTO?

I found the Tesla, and how you need to plan more for long range trips, to be very similar to how you need to plan for flying an airplane. You cannot just jump in and take off knowing you can always take any route or stop for fuel anywhere and when you want to. Forethought and planning are required, especially if you want to get to your destination in a reasonable amount of time.

When did you start development of EVTO?

May of 2016

When was the app first available in the App Store(s)?

March of 2017

What surprised you most about developing the app, once you were into it?

How difficult it was to deal with all the nuances of the car's energy consumption, weather, routing and the complexity of building an optimizing algorithm for all these factors.

How has your vision for EVTO changed since development began?

It was originally intended to be much more manual. Enter some waypoints and have the app show you if you could make it or not. Now it is much more automated.

What's your favorite feature?

Being able to see how a small difference can have an overall effect on trip times and where my energy is being consumed. I find the SoC and Consumption charts invaluable in understanding where and how the charging best takes place. But in reality, having all the data handy in one single app has made the biggest difference to me.

EVGrokker commented 7 years ago

I found the Tesla, and how you need to plan more for long range trips, to be very similar to how you need to plan for flying an airplane. You cannot just jump in and take off knowing you can always take any route or stop for fuel anywhere and when you want to. Forethought and planning are required, especially if you want to get to your destination in a reasonable amount of time.

What kind of planes have you flown? How many log book hours? What equipment are you rated for?

aesculus commented 7 years ago

My medical situation removed me from flying 25 years ago.

SEL, Instrument Rated. ~400 hours logged in one book.

Started in the usual Cessnas's (152, 170's, 177's). Then went on to the Pipers (PA52, Cherokee 6-300) for my instrument rating. The plane I flew the most at the end I really liked was s Bellanca Super Viking. It was an animal. Wood wings and fabric, tubular body, all engine.

Image result for bellanca viking

EVGrokker mailto:notifications@github.com Sunday, April 30, 2017 11:00 AM

What kind of planes have your flown? How many log book hours? What equipment are you rated for?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/aesculus/EVTO-App-Feedback/issues/347#issuecomment-298246950, or mute the thread https://github.com/notifications/unsubscribe-auth/AARS5NgHOU4HR3M9bY-dOy2c8O1a9BF1ks5r1Mw2gaJpZM4NMkAz.

EVGrokker mailto:notifications@github.com Sunday, April 30, 2017 6:00 AM

This is a placeholder

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/aesculus/EVTO-App-Feedback/issues/347, or mute the thread https://github.com/notifications/unsubscribe-auth/AARS5Km9S1gvmdvfQCJ1LDbp6WOU4-7bks5r1IYAgaJpZM4NMkAz.

-- Chris Couper

EVGrokker commented 7 years ago

Quick background on Digital Auto Guides?

A bunch of enthusiasts wanted me to create some apps for the restoration/maintenance. I collaborate (d) with the leading producer of paper automotive manuals to see if there was a market for moving to apps. Turns out everyone just steals their IP in the form of scanned PDF's so the answer is no. We were hoping adding content would drive people to the app concept. It appears that either people are too old to use apps and/or just prefer the paper or stolen PDF's.

The MG story

EVGrokker commented 7 years ago

How did you research and develop the 'secret sauce' algorithms for EVTO? Do you have any special training, background or knowledge that gives you a unique perspective on understanding the various factors (weight, speed, weather, elevation changes) that contribute to EVTO's prediction engine?

aesculus commented 7 years ago

Another interesting question. Check out this post for what got me thinking about this:

https://teslamotorsclub.com/tmc/posts/1084781/

It was good work by the author but he made a few mistakes that did not work out for him. It was very complex and of course I have added to the concept considerably. I asked him if he would join me in the pursuit but he was pretty much done with the concept and wished me well and encouraged me to take what he had if it was useful.

I had the app working pretty well as a manual tool. Diehards liked it. Then the ABRP came out and it was so simple to use and the optimization seemed robust, but lacking a number of features like weather etc). Some of my contributors bailed on me because they thought ABRP was just better. They liked the clean web interface and did not see an issue with its gaps (some fixed since then).

I had mentioned this to the author of ABRP and he encouraged me to continue and gave me some hints on how to do the optimization. In the end I did not use his suggestions but went another route (Using https://en.wikipedia.org/wiki/Dijkstra%27s_algorithm methods).

I have some training and experience with optimization systems but not at the model creation level so I was familiar but out of my league. After wandering around the Google space for weeks I was able to patch up enough capability to put the back ended piece against the prior manual front end. It's now a two pass system where the server finds the optimal charger path using some fairly crude algorithms (probably as good as or better than all the other route planners) and then the device 'refines' the models with nuanced details and weather.

EVGrokker commented 7 years ago

Imagine that you were being interviewed for an article about EVTO, and the question was asked "How did you come up with the algorithms?"

Would your response be the same as what you wrote above (crediting everybody), or "I researched it and combined several published approaches" (crediting nobody), or something in between?

aesculus commented 7 years ago

I am overly honest, so the former