If any of the info on this website was useful for your projects or made your head spin with creative ideas, and you would like to share a token of your appreciation- a donation would be massively appreciated!

Your donation will go to my Robotics Fund, straight towards more sensors, actuators, books, and journeys. It takes a lot to continue building these robots, and I’m really thankful of everyone who helps encourage me along the way.



“A wise robot once said to me through Serial.println- that robots teach us about ourselves.”

All posts in Projects

Nutrition Monitor for the Elderly


This is a nutrition monitor for the elderly. It calculates a malnutrition score based off of body mass index and the daily food intake. The score is then uploaded to a remote dashboard that a guardian can view, and be alerted if it exceeds a threshold.

With the nutrition monitor, a problem can be addressed in a few days, rather than many months.

It is a step in the direction of being able to be more aware of ourselves by using sensors placed in our homes.

Watch it in action:

You can get started making this yourself! Check out the Instructable, NutriModule Arduino library, example Arduino sketch, and the Processing sketch.

MAKE Connected Home Contest

I entered this project into the MAKE Connected Home Internet of Things contest. The prize was a $500 gift certificate at the Maker Shed. We won! Read the article here.


Thank you to everyone who watched the video, commented and shared it with friends, and to the MAKE staff.

It was quite a shock to get a call from Stett at MAKE on Friday evening saying that the Nutrition Monitor won! (My hands were shaking well into Saturday morning.) I will never forget this!

Right now, I’m considering the Printrbot Jr V2. After there are some more purchases in the RoboBrrd Store, we can probably get it!


Just as some background, I saw a tweet about the contest, and noticed that the deadline was really quick. If you read this blog frequently, you might have realised finishing a project very quickly is a struggle for me.

Knowing a gist of what I wanted to make, I researched around about malnutrition. On Sunday (Feb. 2), I dived more into it and learned all about it. There were two references in particular that helped:

Nutrition Support for Adults (accessed Feb. 2, 2014)
Malnutrition Universal Screening Tool (accessed Feb. 2, 2014)

Monday (Feb. 3) it was time to start building. I finished the Arduino library before the hardware was done, then finished the hardware, and created the application plus the remote dashboard. The reason why it’s so tricky to make IoT projects is because of the numerous ‘components’ that have to talk with each other. There are many points of failure. I really lucked out that I didn’t hit many snags.


The one snag I hit was with the conductive foam in the DIY force sensitive resistors not un-compressing fast enough. Instead of the weight being recorded after it automatically detects removal/putting back of the food on the module, a button is pressed instead. The sensors are still a little glitchy, however they can be improved.

Reading the feedback from comments, with some heads nodding thinking it’s an interesting idea, is very motivating. This project is only the first try at the idea. With the future modifications in mind, it will get better.

Future Modifications

- Adding wireless radios to broadcast the data to the internet
This way, there will be no computer needed (except to access the remote dashboard).

- Improving the weight sensors
The conductive foam is very DIY, and this can be done much better to result in more accuracy and reliability.

- Shape detecting mat
Instead of modules, have one ‘mat’ (or plate), that can detect the shapes of many foods placed on it. And detect the weight within that shape area. It will be easier to clean than individual modules.

- Better remote dashboard
Currently the BMI is hard-coded into the desktop application. This should be able to be input via the remote dashboard. Also, ensuring privacy and security of the data is important to be improved.

There are also other interesting ideas suggested, to increase battery life, when to send the data, and maybe switching the place of the module (on the food, instead of the food going onto the module).

One of the obstacles will be the elderly using it / not dismantling it. They may still believe they do not need help, or they may believe it is a spy device by the CIA.

It would be great to run a beta test with some of the future prototypes. I wonder what the feedback would be?

Our future will be full of connected devices. It will be so interesting to have everything synced, communicating with each other, and helping our lives.

Now, I am looking for people who are experienced with health/medical/connected devices. If you know of anyone (or are someone), please email me!

Also, if you are interested in potentially beta testing this (or know of someone who would be able to beta test it), let me know.

Until next time, make something weird or connected, maybe both- weirdly connected!

Progress Update: Nintendo DS + 3DP Game Controller


The goal of this hack is to make an interface to the Nintendo DS. We use Caleb Kraft‘s 3D printed game controller pieces (D-Pad, 4 Buttons) for the interface.

Been working on this for a little while now, and finally hit a milestone with it.

Here is a progress update video!

Took a while to figure out how the buttons on the DS worked. Finally figured it out, and use a level shifter to trigger the buttons from the Arduino. The Arduino is also checking the button presses from the external 3D printed controllers, then sends the corresponding press to the DS.

I want to document this very well so that other people can do this too. Been taking lots of photos throughout. Caleb Kraft’s talk at the Open Hardware Summit was really eye-opening, and I think it’s cool how the ‘copy & paste’ empowerment can have such a huge effect on people that really need it. Games are fun, and everyone should be able to play them. Why should physical ability prohibit what happens in the gaming worlds? So unfair.

Will also be trying out a speech recognition interface, wonder how well it will work. If you have any other crazy ideas for interfaces, let me know.


A cardboard prototype sled with flexible frog feet!


Continuing on from last update, it took a while to brainstorm a new design for the feet. Eventually, we figured out that the outside toes have to be sturdy enough to push the leg, then the inside toes are the ones that can grip, then flex once the outside toes are no longer touching the ground.

This realization resulted in a few changes:
- The inside toes don’t need a bone in them, as it prohibits their flexibility
- The outside toes have to be angled properly so that the toes hit the ground at the same point as the end of the ankle
- The inside toes need a cut to increase their flexibility in the middle of the foot, exactly where we would want it in order to have more grip

With this in mind, it was time to design a new model for frog foot v4!





If you look closely enough to the above screenshots, you will notice that there are construction lines representing the important places mentioned in the list above. The cuts on the bottom foot were accomplished by revolving a rectangle (making it become a cylinder), cutting away from the existing shape, then adding a fillet to smooth it out.

Here is what the bone looked like:


The mould prints:


The 2nd piece of the mould didn’t print properly, however it was only missing the last couple layers so we used it anyway. It was trickier to get the cast out of the mould once it was set, because it was in all of the infill shapes.

After cleaning up the mould, here was the foot that emerged:


The cut on the back of the foot appears to be working as planned:


It was time to put the foot to the test, and it works!

With the addition of the ‘sled’ to the cardboard body, it was able to climb an obstacle that was greater than its body height!


See it in action: (also it survived the fall from the table)

We weren’t sure if there should be cuts on both sides of the foot. It wouldn’t hurt to try, and also make a few quick fixes. These fixes included: removing the fillet from the top of the foot bone, extending the length of the middle toes a little bit further (only a few mm).

After printing out the new mould, foot bone, and adding a new “special” ingredient, here is frog foot v4.5!


Apart from the green food colouring, another aspect that was different from v4 to v4.5, was that for v4, we placed the moulds in the freezer for 20 minutes before adding in the silicone. It helped extend the set time, thank you to Gav for suggesting it. For v4.5 we didn’t, and it was a little trickier to get all of the material in the mould before it became too sticky.

Already, there is something noticeable that has to be fixed:


This can either be fixed by making a better rig to hold the foot bone in place when adding it to the mould, or making those toes a bit larger in diameter, and adding a flap on the exterior of the toe side.

To attach the foot to the froggymobile, the same technique from last time was used: failed prints cut up + hot glue, to attach it to the motor.


It started to move around like this:

But, that is a little slow…

We mentioned last time that our multimeter was broken. It turned out that the ‘Hold’ button was on, so that is why nothing was working. Oops. Still not sure why the other L293D blew up, but oh well.

We added 11.1V to the froggymobile, and it started flopping around frantically!

Here is underneath the froggymobile. Also, you can see some dirt that the white foot has picked up… heh!


We were tweeting out photos and videos of the frog feet whenever an interesting development happened, and it was really great to see the replies, rt’s, favs. Thanks for coming along in the adventure with me, haha!

There are a few things to finish off with the frog feet: small improvements, as well as flipping the design so that there are right and left versions of the foot.

Still not sure if there should be four feet on the robot, or only two. Maybe there can be four, but only the front or back will be active at one time. If the back feet are backwards, then when they are activated, the robot could reverse its direction.

… Obviously we will need to fix the L293D for this task, though! :P

Our next big design task is for the chassis!
- Coupler from the motor to the leg
- The leg
- Motor mounts
- The chassis (this is a large, daunting step)

Although the chassis looks like a sled, what if it would be made out of a flexible back bone or something? Would that even work? Much more brainstorming to do…

After the chassis, would be the sensors and feedback for the feet- but that is a little further in the future from now.

We want to get this done by Dec. 15th. Think it’s possible? Eeek!