top of page

Smart wheelchair
with IoT solution

Skateboarding for wheelchair users

In this article, we discuss the design of a smart WCMX wheelchair. Our WCMX wheelchair has a set of smart capabilities which can help the user learn new tricks and share completed tricks with friends. 

Overview of features

Trick Detection

/

The wheelchair will learn which tricks you perform and will show a list of statistics after each session.

Trick Learning

/

The wheelchair will learn the user during the sessions to improve their tricks by giving tips through the light in the wheel handles.

Trick Recording

/

The wheelchair can be connected to a camera that will automatically record your tricks and name them appropriately.

Fall Detection

/

The wheelchair will response to the user if a fall has occurred during training by attracting nearby people's attention for help.

User Experience

Shock Sensor

wc pic.png

The wheelchair is outfitted with 2-speed indicators on both of the sides of the wheelchair. These will light up green and forward to show you have to increase speed. And the other way around they will light up red and move backwards to tell you you have to slow down. This works independently per side so it also shows you if you need to make a turn.

Pressure sensor

Gyroscope

Accelerometer

Other Electronics

Speed Indicator

Speed sensor

App Interface

The app is complementary to the wheelchair and it shows the collected data of the wheelchair.

iPhone XS.png

In the first section, you can view your high scores and statistics, this is all created from the data that is gathered by the wheelchair.

iPhone XS Copy.png

In the second section, you can view and share your recordings of tricks.

iPhone XS Copy 2.png

In the third section, you can see overviews of your sessions and start a new one.

iPhone XS Copy 3.png

In the fourth section, you can see an overview of all the tricks, the ones you still have to learn and the ones that you completed. Also, it shows a percentage that represents how well the system thinks you can do the trick.

iPhone XS Copy 4.png

In the final part of the app, you can see an overview of a trick. This will show the predicted chance of you doing the trick, how much per cent of the trick you did perfect and it shows what you need to improve to complete the trick.

Data Model

wcmx-personal.png

/

Personal

Data elements

User name

Password

IP Address

Location

Health Profile (Weight, Age, Height)

/

Description

The wheelchair collects basic user data in order to function. By having the weight of the user the pressure on the stair is better understood.

wcmx-collective.png

/

Collective

wcmx-wheelchair.png

/

Wheelchair

The amount of time spent on each tricks

Highscores on certain locations

The range of distance

/

GPS: location visited

Motion data over time

  • Angular velocity

  • Signal of shock or impact

  • Acceleration force

  • Speed value

Pressure on seat over time (weight)

Camera output (Video record & History)

/

Most visited WCMX places.

GPS: Is used to tie it to a location and can help with determining speed to provide the best recommendation of tricks to the user in this court.

 

Motion data: It is necessary to recommend video clips and analyse by the system to predict the recommended route before next practice.

​

Pressure detecting: It checks the presence of the user and detects the change of pressure point to know and predict how the user does their weight shifting.

​

Camera: It is for recording the video clips and be analysed by the system for providing the recommendation and social interaction with friends.

wcmx-contextual.png

/

Phone Contacts

Contextual

/

The user can build the social network in this special sports via share the video, find friends with similar predicted trick success, or offer the challenge to friends.

to view our data architecture

Click

to view our hardware architecture

Click

Machine Learning

wcmx-super1.png

supervised learning

Trick Labelling

wcmx-unsuper1.png

unsupervised learning

Motion Pattern Detection

Anomaly Detection

Know more about our Metrics, Machine Learning Algorithms and Ethics

Many Thanks to IoT6:

Jesse Nijdam

Tim Smits

Ying-Ju Chen

bottom of page