HYDRATION ROBOT

A robot that provides beverages to nursing home residents

MY WORK:

Created the map for Fetch Robot to navigate around the environment.
Modify pick and place code and edit the parameters to pick up the can using OpenCV for object detection.
Created Context and Finite State Machine diagrams.
Collaborate with team members for documentation.

THE PURPOSE OF THE PROJECT:

The purpose of our project is to design a robot to help keep elderly people in a nursing home properly hydrated. Our project uses a Fetch Mobile Manipulator robot to serve drinks to the residents, freeing up caregiver time for other tasks. Our design goals include safe operation, useful functionality, and otherwise minimal interference with daily life for both residents and caregivers.

THE SCOPE OF THE PROJECT:

The goal of the project is to develop a robot to help keep nursing home residents hydrated. To facilitate achieving this goal, we were given access to a Fetch Mobile Manipulator robot and a workspace in GIX (Global Innovation Exchange), as well as a simulated version of the robot and workspace. Our project requirements mandate that we include manipulation (using the robot arm) and navigation (driving the robot base) tasks. The work for this project consists of designing the behavior of the robot, implementing that behavior in code, testing our code on the robot in simulation and in real life, and documenting our work.

THE PRESENTATION:

PROBLEM SUMMARY:

Elderly people in nursing homes often get dehydrated. They can have trouble or forget to get themselves a glass of water. Caregivers are often very busy and don’t have enough time to make sure that everybody has enough water. Mainly in nursing homes, but potentially in other places where elderly or disabled people live. Old or disabled people who are not able to/forget to get themselves water, busy caregivers who forget or do not have enough time to make sure residents are hydrated.

THE SOLUTION:

The solution will prioritize safety and functionality. The main function of the robot includes two tasks:

Autonomous navigation
Manipulation ability.

For the navigation ability, the robot should conduct the following tasks:

Navigate to the kitchen to pick up a can
Navigate to a table to give the can to a resident
Navigate to back to the kitchen with the empty can
Navigate back to the robot’s starting/home position

For the manipulation ability, the robot should conduct the following tasks:

Pick and tuck: pick up a can and tuck it so the robot can safely navigate
Untuck and place: place a can on a tabletop

STAKEHOLDERS:

Direct:
- Nursing home residents
- Nursing home caregivers

Indirect:
- Nursing homeowners/operators
- Nursing home residents' families

STAKEHOLDERS REQUIREMENTS:

Nursing home residents:
- minimum: medically adequate hydration, easy to use
- preferred: beverages delivered whenever they are desired, otherwise non-intrusive robot

Nursing home caregivers:
- minimum: medically adequate hydration of residents, reduced amount of time spent hydrating residents
- preferred: nonintrusive system, easy to use

ENTITIES AND ENTITY REQUIREMENTS:

Coke Can:

We chose a coke can because it is a beverage, it is red, it is easy and cheap to obtain, and it is similar in size and shape to a glass of water, so our algorithms will hopefully generalize to a glass of water as well.

One thing to note is that a closed can cannot spill, while a full glass of water can. We do not plan to make the robot work with a full glass of water in this project, because we do not want to accidentally spill water on the robot.
Furniture:

Our robot will be designed to pick and place coke cans on tables, which will be located in the GIX environment

We are assuming that we will know about the size and layout of the furniture in advance.
In the future, perhaps I could add the ability to automatically recognize furniture to the robot.

Doors:

We are assuming that all of the doorways are open for this project.

In the future, perhaps I could add the ability to open doors to the robot, but that is out of scope for this project.

CONTEXT DIAGRAM:

Figure 1: Context diagram

USE CASES:

John, an 80-year-old man, is moving into a nursing home. John sometimes forgets things, so his kids are wondering how the staff will take care of him. For example, they want to know who will be responsible for delivering drinks to John every day to keep him hydrated. One of the staff in the caring house has introduced a serving robot to John’s kids. The robot will fetch and deliver drinks to John as scheduled when he is in the dining room. The whole process of the robot is totally autonomous, and the robot will keep a certain distance from the elders to avoid collision, so they will be safe with a new partner. In doing so, the schedule of serving drinks will not be wrong as a robot staff takes care of the elders.

Top Level Use Case:

The top-level use case of our robot is to automatically deliver beverages to nursing home residents.

Locate the cup by camera and pick it up
Navigate to the area where residents are and avoid collision
Place the cup on the table and getting back for waiting

Figure 2: UML Robot, Kitchen & Residents

Diagram - Operational Use Cases:

Figure 3: UML Commander & Robot

Diagram – Goal-driven Use Cases:

Lower-level use-cases include avoiding interference with other humans in the scene, navigating cluttered environments, navigating through doorways, avoiding obstacles with the arm.

Avoiding interference with other humans in the scene:

John is waiting for his drinks while the robot is moving to him. Once the robot detects other people in front of it, it will slow down and stop to avoid interference with other people.

Navigating cluttered environments:

The robot is navigating to John, but it finds out the environment is pretty crowded. It will avoid collisions with other obstacles when making plans for moving.

Navigating through doorway:

The robot is coming through a doorway which is a little narrow for it, so it autonomously adjusts the direction and status to make it through the doorway.

Avoiding obstacles with the arm:

The robot stops in front of John's table and tries to place the cup on it. But the camera detects there is something on the table, so the robot tries to avoid collisions with the obstacles by planning the arm carefully and moving slowly.

Figure 4: UML Robot’s arm, kitchen & residents

Sequence Diagram:

High Level State Diagram:

The robot will go through this loop twice. The first time it picks up a cup in the kitchen and brings it to the resident, the second time it picks up the empty cup from the resident and returns it to the kitchen. This high-level system architecture represents the four main tasks that the robot will have to engage in, and the flow between them required by our use case. This system architecture represents the behavior flow of the robot, which is also a mock of normal behavior flow from a human being in real life. If a server wants to give drinks to people, he should first pick it up in the kitchen, then navigate to the area where the guests are, placing the drinks on the table next and waiting for other jobs as a last step. So, with this architecture applied to the robot, the serving process won’t be strange to other people.

More Detailed Finite State Machine Diagram:

Figure 5: Finite State Machine Diagram

CONCLUSION:

The purpose of this project was to learn about robotics development. Of course.

I learned how to use ROS and the Fetch Mobile Manipulator robot, but we learned much more than that.
I learned human-centered design skills, taking into account multiple stakeholders needs and specifications that aren’t purely technical.
I practiced the design-build-test cycle, where I made iterative incremental improvements to our design and code.
I learned new product evaluation skills, such as writing specifications, writing tests, and understanding precision and accuracy.
I learned about integration challenges, the problems that come up when you try to put all of your components together. In the end, I was able to present a slideshow with overview my teammates and a working live demo of our project, showcasing our work to the rest of the GIX community

TESTING AND MEDIA:

Page updated

Google Sites

Report abuse