Introduction: Cloud Fit - Smart Devices for Smart Gyms(Intel IoT)
With CloudFit- smart Wearble we're trying to bring an interestingly new approach to monitor your fitness workouts in a Gym. Workout in a Gym is so boring and its been so from a long time. A part of it is because it cannot be measured in real time. Its all so manual. So here with our Cloud Fit platform we are putting some tiny cool but smart devices on your gym equipments ( dumb-bells, treadmills, weight machines ) which pushes all your workout details onto our cloud console. That's a bit cool. What's more cool is, in our cloud platform we analyze your workout plans and suggest you the recommended exercises . So from tomorrow you don't need those costly personal trainers. We'll monitor your every exercise move. Our motion sensors determine whether you are doing it properly or not, and if not we send you immediate push notification. Our app tells you the correct way to do the exercise.That was fairly cool,But the best part is, when you do it right and you do it daily, we appreciate you on your social media. But beware, one day you miss your workout we'll scream over there and ensure you come back to track the following day.If done we can think of gym equipments as things which talk to you. They get happy and praise you when you do the workouts. They shout on you if you try to be lazy. They help you to remain fit and healthy.
Step 1: Why Do We Need Smart Gym
Most of the people who joined gym won't do proper exercise because of absence of personal trainer which makes them serious injuries . Even if they pay for personal trainer ,personal trainer cost would be more and also you have to maintain whole workout logs manually.Also gym instructor have to manually prepare your workout and diet plans.That's quite mainstream right???
So here we are planning to bring some revolutionary change in gym industry.With the help of few sensors we track your activities in the gym ,can count on the number of sets you can done on the particular machine ,also we track your correctness of particular exercise. Isn't it sounds cool??
It will generate exercise logs for you,you don't need to write it down on pen and paper .It will generate the workout schedule based on past activity logs and your health conditions.It can suggest proper diet plan also.In short it is more than a personal trainer for a gym.
Here the target audience is fitness freaks who either go to Gyms or Use Gym equipment's at their home.Market size would be a huge :-Global gym health and fitness club industry have 78 Billion USD Fitness equipment industry and expected growth rate is 22-26 percent every year.
Step 2: What You Need?
Most challenging part of it to count the number of sets you did and correct of your exercise.
You have solved this problem using some data analysis and implemented a machine learning algorithm.
So what you need?
1) Intel Edison
2) Combination of gyroscope,accelerometer and magentometer sensors (I have used Wiced Sense which contains all) which acts as wearable.
4) Gym Equipment
5) Machine learning API.
6) A very good brain and good coding skills
Step 3: How to Implement?
Above is data plot of various sensors vs time for right and wrong motion.
1.) BarCode is attached with each workout machine , user will scan barcode with mobile which in result it will sense the type of exercise.We have used a Zxing barcode API for android. We can use NFC tag to detect workouts.
2.) WICED sense can be used as an arm band that will sense the motion to judge the workout correctness.
->Used Wiced sense device as a wearable to detect correct motion.
3.) To analyze the motion data first we see the live stream graph plotting using plotly.ly
-> Used Cylon node.js to extract the raw data and plotting live stream graph using node.js.
4.) After extracting gyroscope , accelerometer and magnetometer data from sensors ,we analyze the data and extract various features for our learning algorithm.
5.) On the Matlab tool ,we first implemented the algorithm to learn our hypothesis for binary classification(Correct/Incorrect) .
6.) After getting the data from sensors ,it will send to intel edison and after it will go to cloud for further analytic.
Step 4: Prototype
You can see the prototype on the video attached.Right now ,we have optimized our algorithm for dumb-bell exercise.We can easily enhance it for further exercise.
Step 5: Advantages of Using It
1.) App will reduce injuries that occurs while workouts.
2.) No need to pay to trainers ,app will act as a personal trainer.
3.) Help to learn various exercise in real time.
4.) App will alert on wrong motion axes and will help to learn correct motion.
5.) No more need to track your workout progress on paper , app will record workout history and track real time workout progress.
6.) App will schedule proper diet plan based on real time data .
Step 6: Future Vision
1.) Right now we have developed the prototype for dumbell exercise only .We can further enhance the algorithm to capture the different kind of exercise.
2.) This can be enhanced to provide real time audio coaching to the user about gym activity.
3.) For prototype we have used barcode ,Instead of barcode , we can use NFC tag.
4.) Also we can generate workout reports and send it to doctors for proper diagnostics ,likewise we can also sent it to nutritionist for proper diet plan.