SELF-DRIVING WHEELCHAIR

 

inspired by personal experiences, we've developed an autonomous wheelchair to bring greater independence to the disabled community.

Now, we're ready to take our protoype to the next level.

 

Our Goal

Delivering independence to more people.

Opensource software and wheelchair training data will create more freedom for more people: consumers will have access to the latest updates, and developers will have access to the latest research to inspire greater development and better opportunities for growth across the industry.

Self Driving wheelchair

The wheelchair below is being controlled by an Artificial Neural Network (or AI Agent). The AI accepts (“sees”) a video stream from a simple inexpensive web camera, which is mounted on top of the chair, and from that determines what joy stick command to send to the wheelchair to move, avoid obstacles, and stop if it detects a probable collision.


What We've Achieved

  • A working prototype that we are currently testing to establish a set of baseline tests, so we can measure improvements when they are made in the future versions of the product

  • This prototype uses free, opensource software; software developed by Blue Horizon AI; a low cost web camera; and simple inexpensive hardware computers (raspberry pi and arduino).

  • This Proof of Concept prototype confirms that our vision of applying a trained artificial neural network to the job of driving a wheelchair is possible.

  • When the wheelchair is told to "go", it currently travels on its own without any destination in mind. We have success with the first and foremost most important ability: to insure that the wheelchair will stop before striking something.

WHERE WE'RE HEADED next

  • Collect better and more specific training data for the wheel chair;

  • Make improvements to the chair's ability to drive on its own in order to follow a pathway or side walk (similar to lane assist in cars);

  • Add the ability to follow or accompany someone;

  • Add the ability to set a destination; and

  • Add the ability to make improvements as more data is collected by users

funding

At Blue Horizon AI, we're rapidly approaching the limits of what we can do as a small, non-funded team. Our hardware and computing power needs to be upgraded in order to experiment further with the systems in place. We require additional staff to assist with the development, collect data, and develop better learning algorithms, all in order to advance the release date of the product. Finally, we require a larger lab space in order to conduct additional experiments with the chair.

With your help, we can bring this chair to market within one to two years. 

 

COMMUNITY PLATFORM SUPPORTING AUTONOMOUS WHEELCHAIR DEVELOPMENT

Coming soon!