Final project

Week 01

The last couple of months I collected a lot of ideas.

Based on video's / work we saw during bootcamp, the local session, I got inspired to try to make it big but not too big.

Swim monitor

Every Saturday morning I go to the indoor pool to improve my swimming technique / speed. I always swim in the 2nd lane (whereas 6 being the ones for the masters). To improve I need to focus on technique and then speed. But swmming around I've no idea of how my body is floating in the water. And speed drills? Actually I hate those.

Of course I followed a course and was once filmed in the water using a camera on a stick; whereby a coach walks along side to film, you. But that has been a long time ago. And it is both costly and time consuming to do that again.

Regarding the speed drills. Yes I have a watch but Ihave to manualle activate and deactivate it. So I'm always starting with one hand on the other, give myself some speed and the end I have to find the right button again. Depending on how good I do this, the speed measurement is kind of inaccurate.

What if there was a way to (more accurately) measure my speed and make get frequently inside into the way I swim?

Solution

The [swim monitor] is a system that enables swimmers to:

Track their speed, lap time, distance while swimming
Be challenged/paced by a pre-programmed exercise
Compete with another swimmer (gamefication)
Afterwards see how they moved through the water

Visuals/ mockups

The [swim monitor] is a system that consists of multiple parts:

Start/ stop (on both sides of the pool) that the swimmer can use to start/stop the monitor
Lap monitor that keeps track of time, speed
Pacer that shows how fast a swimmer should swim to meet his/her criteria
Camera that films a swimmer from the front, side and behind
Storage that combines timely data with visual data

Non-functionals

It has to be safe especially around/ in water
It is not allowed to pollute the water in any way (pool water is clean water)
It should be possible to install/ deinstall in 10 minutes time (as other people might need the pool)

Open water buddy

I personally like to swim in open water whether it is to improve my distance or just to cool down. Usually I do this in a rive ("de Vecht") which is very near by. This river is very popular and, especially in the summer, populated with boats.

So when I swim a longer distance I'm constantly looking around to navigate but also to look around whether boats actually see me. And as I swim alone there is nobody who can take a picture of me (just for fun later on).

And when I'm in the water to cool down I usually get out quit quickly as I have no fun just being in the water; I like to do something. Even if I'm with one of my kids throwing a ball, it gets boring after a while.

What if I could have buddy that guides me and plays with me in the water?

Solution

The [Open water buddy] is a device that:

Shows you the way (pre programmed route/ geo-fence)
Tries to catch you or any other person that is close by (like a shark)
Follows you around while you swim
Monitors you while you swim (distance, photo, video)
Randomly goes any direction so that you have to catch it
Can be set to stealth mode to film wildlife
Comes to you when needed
Can be controlled remotely

Visuals/ mockups

The [Open water buddy] is a device that consists of multiple parts:

Simple interface to activate a function / user feedback
Camera to film and take photo's
GPS tracker for loaction\
Light(s) for guidance / protection
Intelligence to recognize people/ objects / birds
Storage that combines location data with visual data

Non-functionals

It has to be safe especially around/ in water
It is not allowed to pollute the water in any way (although it is open water)
It should be small/ strong enough to be thrown away (like a ball)
It should be quiet
It should be possible to put a shark vin in top
It should be possible to put a yellow rubber duck on top of it

Week 2

I thought about creating a new page but just updated this one for now. Biggest decision is that my final project will be, what I now call, the Open Water Buddy. The main reason: I have no idea how to make it . Sounds silly but what the heck; I only live once. First change is the working title.

UWO

Udentified Water Object. Which in a sense does not do justice to what it will be able to do. To better help me design I will decribe it's functions from a users perspective. I will therefore call this person "User".

The object supports 2 main type of modes: normal and game mode.

NORMAL MODE

Scenario 01: GUIDE ME

User wants to go for a swim
User enters the water and turns on UWO
User selects "GUIDE ME"
User sets total distance in meters
User sets interval distance in meters
User selects "ACTIVATE"
UWO starts moving away from the user and keeps a safe distance
User start swimming
UWO signals when the total/ interval distance has been reached

Optional:

User selects photo on/off at interval
User selects video on/off
- User selects continuous or interval/duration in seconds

Scenario 02: FOLLOW ME

Basically the same as GUIDE ME only now the user is followed by UWO.

Optional: - User puts a shark vin on top of the object or a rubber duck

GAME MODE

Scenario 03: CATH ME IF YOU CAN

Looks similar to "GUIDE ME". Object moves away from the user. Difference is that the movements of the object are random but always meant to get away from the swimmer. Of course you can set the maximum radius in which the oject is allowed to move. Also the maximum speed while moving.

Scenario 04: JAWS

Looks similar to "FOLLOW ME". Object tracks the user but tries to catch it. There is a multi-player mode which means that it will try to catch any swimmer it can. Again here you can set a max radius and max speed while moving.

WILDLIFE MODE

User wants to film wildlife on the water from nearby
User turns on UWO
User selects "WILDLIFE"
User sets total radius in meters
User sets total time in minutes
User selects "ACTIVATE"
User puts UWO in the water
UWO starts moving away from the user away from the shore
UWO goes into stealth mode
UWO moves towards wildlife
UWO films wildlife
UWO returns to the starting point once time is up

Non-functionals

I adjusted my non-functionals accordingly

It has to be safe. This means
- no sharp edges that can cut you
- soft impact if you get hit or you hit the object yourself
- no change/ low risk of an electric shok when you touch the object (especially in water)
It is not allowed to pollute the water in any way (although it is open water)
It should be small/ strong enough to be thrown away (like a disc/ frisbee)
It should be quiet when in stealth mode
- so animals are not disturbed or spooked by it
It should be self-supporting regarding energy
- at least 30 minutes when in game mode
- at least 2 hours when in guidance mode
- at least 4 hous when in stealth mode
It should be difficult to detect it
- only external sensors are above water
- overal height above water should be kept to a minimum
It is able to put itself in an upright position
It should be possible to attach special accessoiries on it. For example
- a shark vin (game mode)
- a rubber duck (stealth mode)
It's direction should be clear for other vessels on the water when it is dark

System integration

UWO as a system will consist of several subsystems.

UX

Imagine being the water; maybe the user already has goggles on. This means that the interface should be easy to use and it's information clearly visible; even when it is dark(er). Rotating knobs? LED display in numbers. On the other hand may be the interface should be outside UWO. This would make it lighter? But what if you want to change some stuff while you are swimming. If the UX is outside how do you operate it then. Hmm..

Based on the directions from the user this subsystem is able to navigate in the right direction, the right speed. It's inputs are:

GPS location
Object detection (distance in meters, direction)
Swimmer detection (distance in meters, direction)
GeoFence
Mode/ scenario

It uses this information to pick a path at the right speed. For instance depending on the scenario it will I go toward or away from a swimmer.

While it is navigating it will output traveled distance in regular intervals.

Awareness

This subsystem is all about detecting objects; surroundings. It's input are:

GPS location
Map data (water side, buoys)
Camera/Image feed (maybe 360?)

Whether the camera feed is continous or just images I don't know yey

It uses the information to detect and output:

if UWO will hit something (if continuing in the direction)
what that something could actually be (could be wildlife, buoy, shore, swimmer)
at what distance

Imagery

This subsystem is responsible for detecting/ recording UWO's surrounding and storing it. It's input are:

GPS Location
Time/date
Continous camera feed (maybe 360?)

There is no output at this time but this could be cleaned images that are needed for the awareness subsystem

Navigational lights

This subsystem is responsible for all lights above (ships, vessels) and under water (to follow UWO). It's input are:

Navigation lights on/off
Guidance lights on/off
System warning lights on/off (this is to show what the status is of UWO)
- low battery
- stuck/ don't know what to do
- ...

Central unit

This part is responsible for the overall command structure; for receiving and providing information to the different subsystems.

Visuals/ mockups

Must stuff I can visualize myself but my biggest concern at the moment is the propulsion. I would like to use some type of jet engine (the ones that create forward motion based on pumping water; so no propellor). I think this would be fast and safe. So before I start creating a newer version of my 2D/ 3D sketch I asked for inspiration from GenAI. Short summary it also came up with a waterjet propulsion idea. For now I will assume I can create this.

GenAI also mentioned the term holonomic robot In short if I could accomplish this UWO would be able to "move in any direction from any orientation, creating an incredibly mobile robot. So, it's easy to move the robot in a congested area".

First 3D model in Solvespace

See week 2 for a description of the process. Just again the result for now.

Week 03

During our local session on Thursday, Henk suggested to discuss our final projects with Bas Pijls; just as an independent person. Regarding my idea he had some thoughts/ tips:

Student Mkhitar (2025) made a submarine; look into that
Water pressure from a "peralstaltic movement" is not fast enough
By the pump; it is very tricky to make yourself
Water tight is possible using epoxy.
Maybe I can create my model from "piepschuim" and cut out what I want (I will learn this during CNC week), then put epoxy in it and suc it vacuum (sounds nice )
Overall do-able but choose what you make yourself

Week 09

So it has been a while since I've updated this page but assignments took over . Anyway this week I again thought about my final project. I think that RoboQuack 3000 is a more suitabble name for it.

RoboQuack 3000 is an autonomous swimming fetch companion cleverly disguised as a classic yellow rubber duck.

alt text — created with Grok Imagine using the prompt `a cute yellow rubber duck floating in open water, actively trying to catch swimming people, dynamic action scene, realistic water splashes, duck with determined expression pursuing swimmers`

At first glance it looks like an innocent bath toy, but inside it hides custom-designed electronics, AI vision, navigation, powerfull thrust engine and a couple distinct operating modes (also see week09):

Game mode — it actively chases and tags swimmers with a cheerful quack
Swim mode — it swims steadily in front, maintaining perfect distance to lead the road
Stealth mode — it goes quiet and observes wildlife/ takes photos without disturbing the scene

Week 10

I decided to use the current duck I have. It's no too big, it's nice looking, it's already there. So now I only have to attach a motor in it, give it eyes and some kind of GPS sensor to know where it is.

I therefore took time to 3D scan the duck for the single purpose of using it later on in 3D programs like Blender or FreeCad or whatever. Why? Because I can then take it's measurements and use it to attach a floating platform on the outside and a plaform inside the duck to hold electronics; etc. It's nice for me to do that with a real duck in front of me.

Anyway please see [Scanning RoboDuck]](../finalproject/scanning_roboquack3000.md). And yes I changed it's name again to RoboDuck 3000 (a name says it all).

Week 11

So at the end of global class yesterday (1st of April) Neil mentioned again that midterm review is coming up and I should have finalised my plan by then. Also make sure to explain choices; they should answer a question. Specifically important are:

system diagram
tasks to complete
schedule

I'll therefore use this page as a diary and keep track of questions/thoughts and decisions and use sub pages for more in depth detail (like system diagram, tasks, design, bill of material, etc). I already started with this last week but now I'm sure that's the way I want to go forward.

Thoughts/questions

Based on input devices week and this week were I could try the Grove Vision AI v2 module, I still think that detecting heads/faces is able using an AI sensor. What's going to be a challenge is that it's also able to detect upon a range of 10 meters and under shaking circumstances.
How to prevent the duck from flipping over in water with waves and wind? It is very light.
How to attach the water jet engine to the duck?
How to control the device in early stages; like having a remote kill switch? If it goes off into the distance then I should be able to stop it remotely. Maybe I should just make it completely remotely?
What about water inside the duck? Were do I put the electronics?

Answers?

Maybe I can train my own model on these objects? Using Sensecraft it's not that hard I think; doable. I can then train first with object A and extend that with more objects.
I could make some kind of platform under the duck. I can then also use this platform to attach/ store the batteries. Maybe a catmaran design were the batteries are on both sides? I could attach the water jet in between and just stick the duck on top of it.

image>?

From Irja I got a link to chemicals to make 3D prints water tight. But I'm thinking of making the underwater hull during the molding and casting week; epoxy is water proof. So then the electronics could go in the duck it self. Probably as high as possible; maybe in the head? Also nice place to receive GPS signals and near the camera that I want to use.

image

Regarding remote control ... maybe I should just make it completely remote control. That way I can keep the duck like it is as much as possible. But of course the remote control should then be water tight as well as it will be used near the water.
Maybe a general switch on the duck (on/off)?
Remote control means a way to communicate through air. So not only receiving but also sending (to let the remote controller know everything is ok). What kind of technique can I use for this? From the global class I learned that it depends on power, required functionality and distance.

Trying out

My idea was first to squeeze this pump into an already floating device ...

... but the ships plastic was very hard and I'm not sure how to attach it; maybe first watch a video or something.

Week 12/13

Still in middle of machine week but I also want to work on some ideas for my RoboDuck 3000.

During my midterm review with Henk and Take (global evalator) Henk mentioned that floating would actually be the first priority; and that is true. Once a base platform with my duck on top floats I can add other things. I did also mention that I bought the product Dichtol based on the YouTube video to make any 3D print water resistance (that was a tip from Irja (one of the interns)). But it was their shared opinion hat the coming molding & casting week would be the answer to creating my platform because the base needs to be water resistance; Take also did a lot of experimenting with 3D and water and found it to be very difficult.

Inspiration from others

I mentioned that I'm especially fearfull for my 3D skills with FreeCad/ Blender. So I like to see if/how others made objects that float using molding/casting/3D printing. What kind of shapes and how did thy make it water tight?

I found a couple of websites that I found to be interesting ...

Radio Controlled Duck

%DIV

DIV

Kit-bashing a duck

%DIV

DIV

How to Build an RC Duck - With Arduino and Wi-Fi

%DIV

DIV

Rene the rubber duck boat with a nice video

%DIV

DIV

How I built the fastest RC Duck boat

An extension by the same maker to his radio controlled duck. He does this by using a waterjet propulsion. Good tip is to use velcro (klittenband) inside to attach parts so it can easily be removed/ attached again later on.

Molding and casting

There's a lot of info to take from this. But first I notice that this is mainly 3D printed. So why would I need molding/ casting? What is this anyway? So I looked up a video using a search phrase like what is molding and casting.

I watched about half of it and my take on this:

molding is the process of creating a container or "negative" impression of an object
casting is the process of filling a mold cavity/ container with a liquid material (for instance silicone, plaster, epoxy, or steel). The end result is the "positive" object again.

And yes I understand that I could use it to make a more water resistance platform on which the duck sits. I mean using silicon would even make it a bit soft so that when somebody hits the platform of the duck they are not hurt that much.

From my 3rd week I wrote down that the advise then was to make a mold from foam, use epoxy and then vacuum it. I don't know yet how that works and why that would be usefull but it sounds fun

So for now I'll assume that the platform will be created using molding/casting.

alt text

To access all the inner parts it would be wise to create a hedge on top of it that can be closed with some screws. To make sure no water comes in I can use a sealing.

%DIV

alt text

DIV

Other parts

The camera/ gps sensor will be mounted on the outside of the duck as high as possible. One way I can do that is by making it really look like a Robo duck; for instance by putting on a helmet like RoboCop.

%DIV

DIV

As I have a scan of the duck I could 3D print the helmet and paint it later on with a silver metallic spray. Behind the visor I could make sure the camera can see outside. The GPS sensor could be just under the top of the helmet!?