CE Project: CE Bot

[ash]

Hey AJ! Nice to have you here, haha.
For a description on what the CE Bot is, check here: https://sites.google.com/a/crazyengi...rojects/CE-Bot

So far we are just discussing ideas. Please do take your time in reading the posts in this thread. I know you have a lot to catch up.. theres about 70 posts!

If you post something here, you are automatically part of the project Please do share your ideas!

Locomotion
Just to update on my previous post, the steering thing where two wheels turn with different angles is called Ackerman Steering:
Ackermann steering geometry - Wikipedia, the free encyclopedia

Okay, its not as complicated as I thought! Looks like we can reduce the number of servos after all. But I do wonder if we can still implement this system in the CE Bot..

Got some books on Robot sensors.. will spend some time reading it tonight.

[EDIT]
Here's a pretty good link on RC model car handling:
RcTek - Radio Controlled Model Car Handling Main Section Page

I don't understand half of it, but it has nice diagrams Theres stuff on Camber angling.. could probably achieve that on the CE Bot with extra servos (if we really need it).

[xheavenlyx]

The ideas you gave here are really good. But you know, we are making these assumptions by common sense, not excperience.

I would like us to actually have a look at previous rescue/fighter bots which have to be rugged, and powerful to at least maneuver on a rocky terrain.


For a start and to be fun I want you to see this video on Robot Wars. NOT SAYING WE FOLLOW THIS, but just to get an idea of the "size" when they look about 30 cm and 40kg, but in reality are BIg and about 100kg!!

YouTube - Constrictor (Robot Wars Battles) (Will blow you away!)
YouTube - Extreme 2 All Star Championship Grand Final (Will ... i have no words!)
YouTube - Robot Wars Extreme 1 Annihilator 2 - Round 1 (If you want to know the actual size @_@... unbelievable, it can eat a man....also see the house bot Sir Killalot! 216kg; petrol engine; lol )

This is just the start we will have more info here to study present designs. Ash, the steering mechanism is a good idea, but we will need atleast a 1/2 HP DC motor for each wheel (not sooo big), so for this to work we need good calculations which we will surely do. I will post more info here AND on celabs.tk under a new subpage "Case Studies" Dude, watch that video above (as an example:: one robot consumes 700+ watts, 100kg)

[xheavenlyx]

The Lord of Distruction bot for me: RAZER!

The World Championship semi Final of Robot Wars: YouTube - The Second World Championship Semi Final 2

Razor vs. "someone I forgot after watching the video"

We have a LOT to learn from this guy! Control, Power of the weapon and speed.


Presenting the best one yet when Razer KILLS (tearing apart and burnt - RIP) a house bot, house bots are supposed to be the arena gaurds and distroy the loser bots at the end or when they get cornered, Razer KILLS poor Matilda (Invincible and unforgiving!) :

YouTube - Robotwars Uk - R I P Matilda

(Forgive me if I am going off topic for a day but this is HELLISHLY exciting stuff!)

[ash]

Great links! I missed watching that show, haha

I'll try getting links to safety/rescue/recon robots. I came across some before.. even those that have very similar objectives to ours, eg university student research, etc..

Locomotion
Regarding the motor.. I think a 1/2 HP motor is too powerful for a single wheel! Probably a standard 12VDC (with gearing) is good enough. We don't want to carry too many batteries for the robot. The lead acid ones are quite heavy. Your getting too influenced by those combat robots

For some calcuations, one can use this:
How to Build a Robot Tutorial - Society of Robots

Communications
I found these links for a potential antenna designs for a 802.11 (2.4GHz) wireless modules:

1) 2.4Ghz Vertical Collinear Antenna for 802.11 Applications - WardrivingOnline.com
2) Can with conical horn - WardrivingOnline.com
3) Omni-directional "spider omni" antenna 2.4 GHz for 802.11b (WiFi) WLAN
4) iTrike: The World's First Solar-Powered Internet Rickshaw! (practical implementation of spider-omni directional).

I'll try getting polar charts for them. These designs are quite easy to fabricate. Probably the hardest thing is to integrate the N-type connectors to the WLAN modules (eg USB type). Thats what "hacking" is for! If we choose a directional type, we'd have to figure out a way to rotate it to point towards the other wireless module automatically. As I mentioned in previous post, we can implement xhx's optical sensor idea as a compass.

The reason I'm staying away from Low powered RF Modems for data (note: not remote control) transfer is their steep costs... perhaps $200-$300 a piece. Maybe when we get sponsors, we can purchase them (their baud rates are quite good) Besides, I want to do the testing soon. I can probably build an 2.4Ghz antenna next month or so after I do some initial testing between two computers. Thats probably the only part of the robot subsystem I can test at the moment.. the locomotion system is probably out of my grasp (unless I borrow some wheels, DC motors and a drive chain from the Robotics lab at uni, and fabricate the ackerman steering column. Fat chance! ).

CEans, please do try out any ideas here if you have access to tools and materials. We need to get as many experimental data as possible

Flowcharts
I'll draw a simple block diagram for the whole CE Bot later on Thursday night. That should give us an idea about why modular designs are useful

[ash]

All words and no pics makes Ash a dull boy Hence..



Just playing around with that Dia software. You can choose from many flow sheets. The Cisco set is quite interesting, as well as the isometric one.

Will add more pics soon..

[ash]

Here's a bit of vehicle dynamics, drawn by the Inkscape software (nothing wrong with promoting free software!). The steering geometry and other dimensions for CE Bot can be roughly guessed here using the Ackerman Equation, where x is the centerpoint of rotation:



Using that equation, we can also set the dimensions of the CE Bot itself in relation to a targetted steering angle. I think this equation applies for low speed maneuvers.. it should be ok for CE Bot Any mechanical CEans want to give input on this?

You'll notice that steering the front and rear wheels gives a very tight turn. As I've mentioned in my earlier posts, the middle set of wheels can be moved freely laterally. I'll draw a more detailed diagram next time.

Hopefully this info helpful for those who plan to draft a CE Bot design with 6 wheels on CAD

[xheavenlyx]

Ash, you gave the design on LEGO blocks a few posts earlier. That was really good, as the steering mechanism is implementable. On the design we will also look at how to form the outer chassis and structure on which we can place all the components. One thing I will do later (later meaning after college, only one week left! ) is to write down "which perticular sensors we will be using like Model number and collect their datasheets for dimensions, then we can start planning fitting this stuff in the CE bot and accordingly by wight choose the motor and wheel size AND the batteries!!

And Ash, thanks a lot for introducing the Ackerman Steering technique. That will be awesome for calculations! For more on Wireless (BT and/or WiFi) How to Build a Robot Tutorial - Society of Robots

I am noting down the main checkpoints in Random below for a rough idea. Please do revise them when you can:

================================================== ===
Ash - Test communication protocol. Donno if N- is finalized by FCC, but Latest Acer has N (One of my friend also told about it) So you check this. Please try to learn TCPIP communication basics if possible (1 day/week?)

XHX - We might have to program our Bot on C or Python (Python can be an awesome choice too, I learnt basics in 1 hr! and EXTREMELY Powerful: LINK: Pyro, Python Robotics: PyroModulePythonIntro for Python used in Robots (OSS-Open Source Software)) VERY USEFUL! Full Robot simulation! See if you CAN.

Ash - Further improve Steering design and description for final use.

XHX/Ash - Write on celabs.tk about Sensors that will/can be used finally (common and good and cost effective). This will take long so we can do this side by side. Dont forget Datasheets for each.

XHX/Ash - The motor calculation can come after sensors using that link you sent from Society of Robots. and then Power calculations. We will try to keep it minimum at first.

Dont forget, if you can see the Pyro Robot Sim, if not Ill anyway tell you how it goes. I am leaving for Saudi after 1 week, so there I have time (no fast net connection, only forum and celab.tk)
================================================== ======

[ash]

Nice work bro!
I'm checking out Python, looks neat! Love the efficiency and portability. But it might take me a while to learn it.. I'm still studying advanced C functions in my free time

Maybe we can get some Pyro experts from that site to assist with our CE Bot?

Me trying to get hold of an extra 802.11g USB module, and still searching for the schematics for my Aztech WL230 USB so I can "hack" it safely, haha. I've got two laptops (one with an in built webcam), and plan to take the latter in a drive around once the spider-omni antennas are finally fabricated. My final exams will end by 23rd June. Gosh.. arent engineering examinations perpetual or what?

Im hoping for some input from the mechanical engineers here about the steering thing!

[xheavenlyx]

Yea, I have installed Python and Pyro, and started a bit, when I am in Saudi (Thursday), I will start it there. Today I am going to submit my Thesis draft, Just one more week to go, HECTIICC!!!

And you will more people who have heard of MATLAB than Pyro... like the only person who came to know about RoboRealm was you I wish all CE were like you here.

I would really want to do some real life performance testing of a 802.11a/b/g Ad Hoc connection. I could probably test it between two laptops as a start. Can you suggest a methodology? Is a simple file transfer enough? Is there a way to have video streaming or conferencing between devices?

And yea, What "Hacking" are you talking about? Antenna? That is fine. I was thinking about converting the 802.11g to 802.11N lol. Anyway, Keep trying. If possible go to local comp shop and ask if they have junk hubs/switchers wireless, to salvage this connector from somewhere. College, computer shops, friends...anything

I will start Pyro and Python and let you. The most exciting part about these two is that they are completely modular!! And we can use C/C++ functions for tight timing applications, or even ASM (Assembly lang - learnt some in college and know it to control parallel port)

[ash]

Haha, that thesis work seemed like forever eh? Good luck with it!

Yep, I'm just gonna try soldering in the antenna cable connection. Also need to remove the PCB track that connects the existing embedded antenna. I've got some progress on researching how to do it, so should be no problem. My friend also has an extra wireless USB. So far so good!

By the way, ASM isnt portable much right? Even if they are a little similar, different architectures use different type of ASM yea? I've only covered the Intel syntax. At most, we'd probably use assembly language in uC's, since their memory is very limited. The assembly language in PIC's seem easy to learn (I suppose because of the limited number of registers )

[xheavenlyx]

Exactly, PIC's has a reduced instruction set than lets say 8051 uP or Pentium 4. P4 has grapix and mouse instructions too in ASM. As you said because of less registers and processing power.

ATMELS ATMEGA8 has a good set and they are kinda precise than PIC as they solve 1 instruction per clock cycle. PIC does one instruction in 4 clocks. Moreover, I have read a firmware for a new ATMEL uC which also has a USB function, which is pretty ok (not THAT tough). (NOTE: some new PIC also has USB).

We might (most probly?) be using a PIC or ATMEL uC with the ACER. Right? and yeaaa, ASM isnt very very different, we can know all the functions by reading their datasheets

[bayazidahmed]

Sorry for the delay guys. I got real busy with my new job.

Excellent development and good work from both of you.

I'll see what I can do now. Though I am an electronics and communication engineer but I am good in electronics and computers so I will be able to help with the coding part easily. And the bad news I am not so good in mehcanical and I really dont care which steering we use.
I havent read about the type of steering links given in the forum and in my opinion, I would prefer four independent dc motors which can be programmed individually, whatever type of steering you call it.
For turning we simply dont give power to one side of motors and rotate the other side. This is highly flexible, trackable (especially for antenna to track the source, I dont know how can we track the height) and easily programmable.

in btw, xheavenlyx, which city are u going to come in saudi?

[bayazidahmed]

There is something called dynamic C, Jackrabbit uC uses it.

We write the program in C, this compiler converts it into ASM for the uC.

The good thing about jackrabbit is, it has multi-tasking features, good memory and the main thing is, it has more number of ports and all ports are programmable as I/O (This is pretty good. We can control all the 4 motors using one uC and use this same uC for some different functions also like antenna tracking or in broader terms we can say anything related to movement can be controlled by this one uC, be it speed, direction, distance, antenna tracking, ...)