AR Gym Trainer
Identify problems faced by users in Gym and design a solution for them.
Year:
2023
Duration:
4 months
Deliverables and Tools:
UX Research : Miro, Interviews, User Testing
UI Design : Figma
Prototype : Spark AR, Unity
Target Audience
For gym beginners who feel uncertain about where to start or are uncomfortable with exercise routines
The Problem
​Activation (number of people buying Gym Subscriptions) in Gyms is 50 lakhs yearly and Retention (number of people showing up to the Gym) is 12 lakhs only. The retention is only 25% of the activation.
Reasons for Low Retention of Users
The reason was identified by conducting a survey with 112 participants
Busy Schedule
People are unable to manage time, and dedicate a certain amount of time to fitness
Intimidated by Environment
People are unable to manage time, and dedicate a certain amount of time to fitness
High Gym Expenses
Even if members get a gym subscription, getting a trainer, dietitian, protein, etc is expensive​
Insights from Interviews
-
93% of users exercise in the wrong posture
-
Most injuries are caused while lifting free weights because machines are made to restrict joint movements
-
The first emotion most users feel while entering a gym is “confusion “
-
9/10 users have ego-lifted and injured themselves Gyms are not beginner-friendly
User Journey of a New User
Brainstorming ideas to correct posture
Concepts for Body Tracking
Solution: Body Tracking with AR
What will an AR Trainer do?
1
It counts the repetitions and sets for the user
2
It detects wrong posture
3
It helps users to identify which joints need to be used while doing the exercise
4
It places the right momentum against the users momentum to help users identify what is the accurate momentum while exercising to get the best result
Where will Cult.fit integrate AR Gym Trainer ?
Open Cult.fit / Check-into gym
Cult shows your target muscles
Scan QR Code at the Gym
Instead of cult sample videos open AI Gym Trainer
The app shows the Workout
Break Screen
Scope of the Project - 1
A screen post the exercise asking users what weight did they lift with today Impact : Research says after 480 repetitions regularly with the same weight your muscles are safely capable of shifting to an increased weight
Scope of the Project - 2
This will help the gym identify which weights and machines are being used the most in the gyms so that the gym can increase those weights and machine This will cater to the problem of high waiting time of users in the gym
Development of the prototype
The prototype has been created on Unity by using mediapipe library.
The logic behind the code is that the phone camera detects the human body and maps down all the joints. Each joint has been assigned a specific number in mediapipe.
For each exercise, joints are noted and the angle to be maintained while exercising is noted.
While exercising if the user is not at that angle the joints turn red, and if they're at the right angle the joints are green.
Limitations :
1. The prototype has been developed only for doing biceps
2. To identify the momentum correctly and for it to work on each body type the solution needs to be made by using machine learning