Tuesday, 15 September 2009

FLL 2009/2010

First Lego League holds a competition yearly, with the same question being used for the whole world. What I love about the FLL is that, everyone competes using the same questions, same play field, as well as same lego sets. Where all remains the same, only their creativity/ingenuity of each team to solve the problem will make them the winning team!

You can view the full FLL rules for FLL Smart Move 2009/2010 here.

For a quick summary of what you will expect from the FLL Smart Move 2009/2010, you can view the video below

Sunday, 13 September 2009

I suck at writing blog.
But still, here's my point of view on the robots.

Small and agile with risks are what our competitors did during the K. Terengganu NRC.
They used the aim and shoot method.
This method basically is to make the robot move straight and grab ( other methods also apply) the object and put it to the destination.
They risk:
  1. The position of the robot at start point must be EXACTLY where they have programmed.
  2. Battery voltage must be EXACTLY the same as what they have programmed using.
  3. Luck. No luck = GG
  4. They must programmed it correctly in a sense that the programming must be accurate by the millimeter. And that's just too back-breaking for us?
  5. Even the condition of the mat will AFFECT the performance AKA determine their fate.
THE SLIGHTEST MISCALCULATION, MISTAKES, EVEN BY A MM, WILL RESULT IN FAILURE.




But if, and if, they are:
  1. Lucky
  2. Able to program well
  3. Makes no mistake.
They get lots of points as their robot is fast.


Next, is what we did.

Use Light Sensors to move with the black lines on the mat.

This method is:

  1. More accurate.
  2. May auto-recover.
  3. Minimises human-error as the robot will move itself without requiring us to aim the robot.
  4. It's now a fully-automated robot. No a car that picks up objects blindly.
Nothing is perfect. Same goes with this method.

It's :
  1. SLOWER if compared with aim and shoot method. But if it's programmed well it may go faster ? ( unconfirmed, no chance for us to see others use this method )
Method of getting object.

This is gives us a headache till the very end of the competition.
We thought of all kinds of methods. Some fail. Some shines.
At last, we decided to go with the box method.
The robot's arm is a square that extends out of the robot.

The arm

We got frustrated thinking of other methods that we go ahead with a very straight forward box that just swoops in the object.

Well, this method brings a lot of problem.
Especially with the rubber object.
Rubber + mat = Friction and loads of it.

With the robot pulling the object, it's like telling it a mammoth task for the robot. It squeaks, jerks and even jumped over the object with the object still stationary.


There's also the grab method, grab-n-lift method, scoop method and so on.




It's late now. gtg. So, anything missing? Wrong? Bullshits? Please dump it in the comments.
And hopefully, I'll make up to it in my next post.




P.S. This is my first time writing and posting a blog post.


Friday, 11 September 2009

Point of View: Robot Massacre

Hello world ! This is my first post. Thanks to Eric for the invitation.
To make sure our robot will keep moving forward, I came up with these few questions and opinions that I believe is essential for our future robot construction.

First of all, the size of the robot. Our previous robot moves slow and steady although a bit bulky.
The mass of our previous robot is a pro and also a con. The weight is important to decide because it will eventually affect our turning, movements, power consumptions and of course the type/number of wheels. How can we give a balance "diet" for our future robot from being fat and slow? We can build it small and compact, mild and light or just with simple mechanisms.

I believe the robot should be small, simple and not-too-light.
Small, it is easier to build and does not interfere with the tracks or turnings.
Simple, just in case anything screws up, rebuilding it would be easier.
Not too light, a little bit of weight is recommended to prevent the robot from slipping, jerking and possibly gets off track.

Another thing to take note of is the sensors. It is always light sensors. Our previous robot is causing a lot of inaccuracy and failures on turning.
The obvious reason is the distance between the sensors and the wheels. The sensors shouldn't be too far from the wheels nor too close.
Our future robot should have a close enough distance from the wheels. So when the robot stops, the wheels are in the exact location as we want.

Arm mechanism. So called the pain in the ass.
Is there an arm design that could pick up almost every shape and sizes ?

Is grab-n-lift a good option ?
Do we really need an arm (guess so) ?

The following is the battery. Which battery should we use on competition day, rechargeable or new AA batteries? It would be best if we get some sponsorship from the school for batteries. Program using rechargeable then optimize it to a higher voltage, sounds good ? I believe the old design is a power hogging machine, therefore the new design must be small. It failed on the design and programming. Better design is a must. Until now, I'm still wondering how did we won third prize in the competition with this rotund golem.

Next, the wheels. Should we consider more than 2 wheels ? Do we need small wheels or big wheels? I think we are better off with big wheels (previous wheels).
The small wheels isn't giving much "grip" from what i see, and it may slip if we use a faster speed, it's just too light. The wheels are depending on the size of our robot. It's not like we put tiny wheels on a giant machine, that would be LOL !
Besides that, if possible, we're going for max power and max velocity. We never know what would happen next.

My views may be a little bit of obscure and ridiculous, I think you get the idea. Well, it's just my point of view.

Tell us what do you think about a winning robot.

Cheers.
-CH-

Wednesday, 2 September 2009

3D modelling using Solidworks

I have been relearning Solidworks (a 3D CAD software widely used by engineers & designers) for a while and decided to do some simple modelling on my own after going through some of the tutorials. So i just picked the Technic studless straight beams as my subject due to its simplicity & my skill level :p

In engineering, just like in real life, there are many ways to solve a given problem and the rule of thumb is always to solve it using the simplest method. I completed the first 3M studless straight beam using 7 steps (or 7 features),

But i thought i used too many steps to complete the model, so i made another attempt to remodel the beam using different methods and this time around it only uses 3 steps to complete the model.

Satisfied, i continued to model the rest of the straight beams of varying lengths including 5m, 7m, 9m, 11m, 13m, and 15m & made a screen shot of all the different beams lined up together,



Overall, i had a fun experience using Solidworks to model the Technic beams. I understand there are a lot of LEGO CAD softwares out there that would satisfy most Mindstorms & Technic fans, but for the few who find them too limiting in features, eg designing your own custom parts, performing some stress analysis, or some motion analysis, then you should give Solidworks a try ! (provided money is not an issue for you :p)