TrackMeUp! - Smart Workout Log

automate the logging of your workout


Technology and mobile app for automatic recognition and logging of workouts and exercises, using sensors in smartphones and wearable devices.

Current approach

It is hard to overestimate the importance of logging your workout. And if you don't believe me, here are some links:

You can dramatically increase your long-term gym results if you record all exercises, during every workout, and assess this information from time to time.

There are currently some options for recording your exercises:

  • Obsolete - paper and pencil

  • Modern way - app in your smartphone

  • The most modern way - using your smartwatch as an additional device for logging


All these approaches will accomplish this goal, but they also have some disadvantages :

  • It takes time to create the habit of writing down all your sets and reps because it is boring and time-consuming, which is the reason that many people don't log their workouts at all
  • If you use intuitive training, even smartphone apps are not very convenient
  • Only sets and reps are usually logged, but there is also other useful data about your workout that is worth recording
  • Manually logging is prone to human mistakes - a person can forget or mix up something

TrackMeUp! Technology

We now live in a wonderful world full of smart devices, and have an opportunity to log our fitness activity almost effortlessly. It works well enough for walking, running, cycling and even swimming; but what about the gym, crossfit, and bodyweight workouts? Currently there is no an app or service which records your gym activity, which is why we decided to create "smart workout log" - an app that automatically logs all exercises that you perform during a workout, along with the amount of repetitions.


Obvious advantages of this approach:

  • Effortless and cool! You just go to the gym (or stay at home or go outside), do your routine without any distractions, and have a full log of your workout!
  • With this technology, it is feasible to record not only exercise type, sets, and reps; but also resting time and even time under load (which is perhaps more useful than reps)
  • You don't need to proactively set up your workout routine and then change it during a workout if you decide to do some other exercises. Just train. TrackMeUp! does all the work for you.
  • No need to purchase any special devices (like a special fitness tracker, for example) for auto-log functionality. TrackMeUp! works with any wearable device that has at least an accelerometer

Additionally, we are going to create an API which will be able to recognize exercises and repetitions, and will include the ability to connect with any existing workout logging app. So, maybe your favorite logging app will get this awesome functionality too.

There is only one disadvantage (but we should admit it is rather annoying):

  • If you work with weights, you should still enter the weight in kilos or pounds that you have lifted. You can do it using your smartphone after a workout, during it, or by using your smartwatch to enter only two or three digits after every weighted exercise. If you can suggest a better solution for this issue, feel free to share it with us and the community in any way.

How it works

Every wearable smart device has an accelerometer, some of them additionally have a gyroscope. People usually wear these devices on their wrist, and the arms almost always are involved in gym activities. 


Furthermore, the same exercises tend to have the same motion patterns, especially for the same person. Thus, using data from these sensors, we can recognize exercises, and the number of reps performed. Recognition will be done using neural network technology, which is common today and yields amazing results.

We are going to support different device positions:

  • Wrist-only device¬†
  • One wrist's device and one ankle's device
  • One wrist device and one device in a pocket

And different device sensor types:

  • Device with accelerometer and gyroscope (usually smartwatches)
  • Device with accelerometer only (usually fitness bracelets)

The accuracy of recognition will depend on the combination of these options, but the main variant will be the "smartwatch on wrist" combination.

Is it possible to do all this, or it is just fantasy?

Good question, and one which we ask ourselves every day :). There have been a few trials, which had very good results. For example, the accuracy of recognition when a device was on the wrist was about 95 percent, and on both the wrist and the ankle - about 99 percent. There are some links:

  1. Recognition and Repetition Counting for Complex Physical Exercises with Deep Learning
  2. MiLift: Efficient Smartwatch-Based Workout Tracking Using Automatic Segmentation
  3. Exercise motion classification from large-scale wearable sensor data using convolutional neural networks
  4. RecoFit: using a wearable sensor to find, recognize, and count repetitive exercises

Does something similar already exist?

We are sure that a similar service has not yet entered the market (at least not one that is well known). If you know about a similar product or startup, please share this information in the comments. If somebody else is working on this idea for commercial use, it will be very interesting and helpful.

The development road map for version 1.0

  1. Develop an app to gather marked up accelerometer and gyroscope data for different exercises <<--We are here now!
  2. Gather exercise data
  3. Train neural network using gathered data, and determine acceptable device positions
  4. Develop an app for workout auto logging
  5. Develop an API which provides auto logging functionality for currently existing apps

About us

My name is Yaroslav, I am the creator of the TrainMeUp! concept.


(I am on the left)

My experience in software development is about 8 years. Furthermore, my experience in gym training is about 12 years (I am sure this will be at least a little helpful for completing this project)

What is this all for

We need help from the fitness community to find out if this is idea useful for anybody. We crave to hear your advice and feedback, and we will need volunteers to collect initial exercise data of different types. So, if you like this idea and the opportunities that will emerge with it, please, subscribe to our mailing list and follow us on social media. Feel free to share your opinion and thoughts about this idea in the comments, this will be very helpful. And please, share this information with your fitness buddies, this will help us a lot.
We will have the next update very soon, we promise :)

Or send us an email!

Comments System WIDGET PACK