Student Project for an Autonomous Bed Mover

Hello we’re a group of students from the University of Southern Denmark. Who has taken the task up to make an universal bed-mover, which will navigate in the hospital autonomous.
During the idea and concept phase we’re encountering some issue that we visualize could be a problem for the robot to solve, due to our lack of experts knowledge in autonomous robotic systems, we most reach out for a helping hand.

Basically the bed-mover will have to transport the beds from one room to another and also in different floor levels, sometimes with patients in it, so patient care is important and also without patients.

Some of the issues we’re thinking could arise is:

We image the robot will have to recognise the specific bed it has to pick up either by one of the following systems/methods:

Placing a beacon/button on the beds so the robot would know where to go and which bed to move.
Make a barcode for each bed so the robot would know which bed to take.
Place a button on the room doors so you could summon the robot, and then have a beacon/button placed on the bed so it would know which one to take.

But if a nurse would happen to manually move the bed, then how will the bed mover be able to recognise it and find the new position for the bed?

Also we need to make sure it doesn’t just take off with a patient when it was meant to pick up an empty bed - so some weight measuring or thermal inspection might also be needed in the system.

This is just what we’ve been visualizing for now, but probably 50 other scenarios could happen.

So what other issues are there to be aware off when implementing an autonomous system around other people?

Thanks, if something is unclear just, say it and we’ll explain further.

Best regards.
https://scontent-arn2-1.xx.fbcdn.net/v/t35.0-12/14600523_10154587388104419_1030320046_o.png?oh=04ac2318cabc3991b4a42c61fef5d297&oe=57FA8ED9

Absolutely put sensors to ensure that there is not a patient in the bed! It would be far too easy for someone to fall out or wake up and decide to get off while moving. With something that important, I suggest at least two independent checks; temperature and weight would do nicely.
You may want to look at RFIDs to help the robot identify the bed.
Consider the traffic pattern of the hospital. Observe how traffic flows, and especially how it works around obstructions and when paths would cross. How fast do beds move today? Interview an orderly or two, and ask about what can go wrong pushing an empty bed, and how they deal with it. Emulate those behaviors as much as possible.
There should also be a way to verbally stop the robot - something simple like calling out “STOP”.

Safety First!
Put bumpers on the robot/bed so it doesn’t run over patient’s in the hallway.
Use indirect and contact sensors to stop the bed.
(check out industrial robots)

https://www.bastiansolutions.com/solutions/technology/industrial-robotics/industrial-robotic-solutions/mobile-robotics/lynx

http://www.mmh.com/article/agvs_carts_and_robots

Is the robot going to tow or push the bed? (beds would need their own wheels)

If riding an elevator is needed; can the robot fit under the bed to save space?
Dave

This is a cool project.

It may be better to check for a change in weight or temperature. Picking up an empty bed seems to be one of the requirements.

Once you have the basics working, you may need to consider how to accommodate equipment such as IV’s or monitoring equipment that needs to stay with the patient as their bed is being moved.

Take a look at University of Michigan’s hospital system in Ann Arbor, Michigan. They have, or had, robotic delivery systems for food service, supply deliveries, and more. Last time I saw them was back around 1992 or so, but it basically did what you are looking for. If I recall they had a line following system that ran in a “robot carpool lane” of sorts. It was all pretty cool stuff when I was 9 years old.

I remember thinking they moved slow, but carried more than a person could.