Thoughts on a new Cycling App I'm developing? (could replace head unit?)
 

I’m reaching out because I’m looking for some validation for this idea I’ve been building a proof of concept of.

A little about me to set some context, I, along with my roommates, make free cycling related tools in our spare time, including a routing tool, https://sherpa-map.com, a physics simulator tool for cycling, GPX Route Speed Estimator for Cyclists: Multi-Surface, Weather, and Nutrition Strategy, and a GPX activity racer tool GPX Activity Racer - Race Against Past Performances or your competition.

I’ve made a few others as well, but this post isn’t about that, I just wanted to showcase the fact that I can build the upcoming idea, but I’m wondering if it would be worth it to develop it for anyone else.

The idea:

A phone app that lets users import a route GPX file and/or start an activity. The app is designed to work entirely offline, in your pocket, without much of a GUI, providing voiced turn-by-turn directions if following a route.

While you are riding, it allows you to ask questions or give commands into your Bluetooth headset, like, “What’s the distance until the next aid station?”, “Route me back to the beginning on the flattest possible route.”, “What’s my best 20-minute normalized power so far?”, “What’s my current distance?”, “What’s my net elevation gain?”, “What mile has the beginning of the biggest climb on the route?”.

Utilizing the physics simulator code and your riding so far, it could even determine when you’ll be where, so you can ask, “How much longer until I’m back on pavement?”, “When should I start heading back to avoid the sun going down when cycling?”.

I also plan on having it continuously monitor and analyze your ride, so if you finish a climb it can recognize this and state “Nice job! You managed 340 watts for 3 minutes up the last 300ft climb!”.

Basically, the idea is, that this could replace your head unit (or accompany it nicely).

I even intend it to have rerouting capabilities, i.e. “take me to the nearest gas station”.

So far, I’ve managed to develop a functioning Android app that works entirely offline, recording with real-time AI background noise removal, monitoring for speech using silero-vad, recording speech to memory, transcribing speech using a specially trained Mozilla Deepspeech model, on the phone, tokenizing the transcription with a WordPeice algorithm, and feeding it to a specially trained BERT Tensorflow model which maps your intent to a particular function.

I plan on developing this app no matter what for my purposes as I feel having practically limitless knowledge of the entire route, the activity thus far, as well as voiced turn-by-turn directions that function entirely offline could help me with upcoming gravel races. However, if people find this concept interesting, I’d buy a MacBook to make an IOS version and flesh out the interface to be more user-friendly and have things like, a login, a connect with Strava feature, etc.

Also, second idea:

I found that the background noise cancellation using AI/ DTLN, is extremely effective, I could additionally make a peer-to-peer Bluetooth phone-to-phone mesh network app that allows for multiperson real-time voice communication, offline, so you could easily talk to your buddies, say, when mountain biking, with background sound/wind eliminated. If anyone finds that to be an interesting concept as well, just let me know.

Also, I'd make all of these apps free. Don't get me wrong, I love money, but I love seeing people use my apps/sites even more.

Until there's some way to do a low-power display on a phone, I just don't think phone apps will ever replace dedicated cycle computers. The problem with phone-based cycling apps is that the phone's screen consumes a lot of power and thus drains the phone's battery very quickly (especially if WiFi and Bluetooth are enabled). I use my Garmin 830 primarily for navigation on my rides, which are typically several hours long. Using my phone for navigation doesn't work as the battery doesn't last long enough and if my phone goes dead, I have no way to call anyone in an emergency. :(

Interesting concept, but I really don’t get on with voice controlled apps and it’s about the last thing I would want while riding a bike. I prefer visual data on my Garmin, along with simple audio alerts when appropriate eg Varia radar alerts or upcoming turns.

It's great that you are doing all this, however I second the two concerns re 1. Visual is better than spoken, especially with wind/road noise to compete with, and 2. Phone battery life.

Right now I am in the process of manually mapping out three 1400-3000 mile loops I plan on tackling over three different upcoming summers and have been seriously considering using a GPS unit for assist while riding, but don't want to have anything more than my phone with me. I like the ideas you have presented and I think it is worth pursuing. Voiced directions are a great idea, however having a map display is essential.
How about picking up where Pedalcell left off and produce a rim powered power unit to keep the phone running while in use? I bet you know someone who has the engineering skills for that project. Then I can use your phone app alongside my handmade maps!

I think this could be an interesting app for TT'ing or other racing. It could be like having a pacing coach in your ear.

Don't let the UCI hear that.

The AI interface seems very promising however very few serious cyclists who would benefit from the features you have outlined use a phone as their cycle computer thus severely limiting potential users. If the features you have outlined do become desirable its a very short leap for the traditional manufacturers such as Garmin to integrate and make the app redundant.

‘Had a running app for some time that gave me splits, pace, stride length, total time and time to finish, etc For running it was great for not having to look at a watch, but have the nice lady talk into my ear buds. I agree that for a TT it would be invaluable. For anyone out for longer than a couple of hours, battery life of the phone would be the limiter. And most people out for less than that time are probably not data junkies.

Doesn't the Ridewith GPS phone app do all of this?

As was said above, the phone will burn through battery charge very quickly. Last year's Bike Virginia went all gps last year and those without Wahoo/Garmin had to use their phones. Some (my brother included) waited until ride day to figure that out. There is a setting...somewhere...to turn off the screen except when a turn or cue is needed, but lots of folks had a hard time figuring it all out.

Here is how I attempted to solve the problem: I later that year upgraded my Wahoo and gave my old one to Luddite Brother. So far, it has not left his mantle, but it's only been a year.

I would absolutely LOVE a peer-to-peer Bluetooth phone-to-phone mesh network app. We ride a tandem and have not found a good intercom system. We have a wired one, but the sound quality is pretty bad. I would guess many tandem riders would enjoy this. As for all the comments about power, I ride with a small battery on 40+ mile rides, so that's not a problem. https://www.amazon.com/gp/product/B0...e?ie=UTF8&th=1

The navigation is pre computed unless you get off track, further, with the screen off it consumes as much power as spotify.

That was the initial idea, all audio alerts/cues, then I started exploring other possibilities, it seems like it would be a good idea to pare it back a bit though, you make good points.

They don't let you voice communicate with RideWithGPS app, it essentially just gives voiced navigation. Additionally, my app would be screen off first, and completely free :D

As mentioned, this would take less power than spotify, especially if I pare down the app to specifically give the running statistics you're interested in, with the screen off and in your pocket.

Good to know there's a running version that has some of the capabilities, I'll have to look it up.

Do you have an alpha or beta version of this available? Would be very pleased to be a tester for you.

Garmin units are extending “legacy” architecture. I doubt they’d be able to add these features very easily. The only manufacturer that would be able to do this somewhat easily is Hammerhead (and they aren’t a “traditional” manufacturer at all). There’s a small chance he could shop it to Hammerhead.

In any case, [MENTION=471973]firebird854[/MENTION] doesn’t appear to be much interested in monetizing it (strange dude).

I will soon and will definitely make a post about it then since it seems to have enough validity to make it reasonable to spend the time to develop it into a more production ready state. Frankly though, I'm working on a couple of other creations simultaneously, so it might not be alpha/beta quality for a bit.