Welcome to the knynet wiki!

This project aspires to develop a network of intelligent agents that is capable of giving people with disabilities more independence in their daily lives. The ultimate goal is to have the software embedded in agents as well as robotics, and all of the entities communicating with each other on a network. They will work together to provide care and assistance to individuals.

Reason for the project

My son Kenny, hence the name of the project, has multiple severe disabilities. One of the most severe impairments is his lack of muscle control. The only part of his body that he can control somewhat consistently is his head. He therefore cannot move independently in his wheelchair and cannot use the controls for a power chair, if he had one. He also speaks only a few words, so he uses a Tobii device (http://www.tobiidynavox.com/) with eye tracking to speak for him.

So I started thinking about how he and people like him could gain more independence when their options for manipulating any kind of device are severely limited. Here are some of the questions and answers I ran through to ultimately arrive at the design for this project:

• Could I develop a software power chair controller for the Tobii and a hardware interface for the power chair that would allow him to move the chair by interacting with the screen?
  • Not exactly. Even his movements while looking at the screen are usually erratic. While communicating which activity he wants to do next can allow for errors, navigating a power chair near hazards cannot.
• Could the Tobii software then just present him with destinations which he could choose from, and the chair take him safely there?
  • This is a step in the right direction. We are starting to feel the need for some intelligence now as the chair cannot deliver him safely to a destination without awareness about the immediate environment, concerns for hazards and other variables we take for granted.
• What if he tells his chair to take him outside, but it is time for his medications to be given to him?
  • Now we are getting to a real need for a system of intelligent agents that can give him independence but also make sure he is taken care of, as he is unable to make those decisions himself.

Let's take that final point and give an example, high-level overview of how this network would react to the scenario (technical details of how this works will be explained further in this wiki):

1. Kenny uses his Tobii screen to say "Take me to the backyard," to his chair.
2. ChairBot hears the order.
3. ChairBot cannot conflict with any orders given to it or up its chain of command, which includes GuardianBot. Both bots are owned by Randy, who has given subowner status of the ChairBot to Kenny (which is why he can give it orders).
4. GuardianBot was ordered by Randy to give Kenny his medication in the kitchen every day at 2:00pm.
5. ChairBot calculates that bringing Kenny to the backyard would conflict with delivering him to the kitchen by 2:00pm.
6. The Tobii announces to Kenny that it his bringing him to the kitchen first so he can take his medication.
7. ChairBot safely brings him to the kitchen.
8. After medications are administered, ChairBot brings him safely to the backyard.

I want to point out that the end result of this project is not allow caregivers to ignore their dependents, but to give people with special needs more independence and allow them to get cared for with their various needs in a more timely fashion when other responsibilities delay their human caregivers.