Robot Guy's Technology Journal

I received my BeagleBoard! This is a complete Linux embedded system in a 3" x 3" board. I was amazed at how small it really is when I took it out of the box. I connected it up to my monitor using an HDMI cable, and it did boot up and display a beagle on screen. I don't have the serial cable adapter I need for it, so will have to wait to tinker with it more until I get that. I got a Rev C3 board, which is the latest version, which is good, and am excited to start learning more about it and working with it. If all goes well, there will definitely be a place for BeagleBoard on at least one or two of my robots.

8-Dale

73 from N7PKT

I've been doing some tinkering with hybrid leg/wheel designs and built up a 5 joint leg with a steerable wheel attached at the end. I was looking at the completed assembly and was wondering how I could check it out and work with it. I did some rough mechanical alignment of the five servos according to how I want the assembly to move and the maximum ranges for each joint.

Since the only robot body I have that is large enough to support this leg is WALTER, I attached it to his body. I was amazed at how this hybrid leg looks on WALTER. His body is plently light enough with three decks that I think four of these legs will work fine to move him around, even with batteries and electronics installed. Since there are going to be four steerable wheels powered by gear head motors, I'll have to use two Sabertooth 2X5 motor controllers because it won't work to pair the motors up on a heavier duty controller. With each wheel being steerable, I need independent control of all the motors.

I'll definitely be using an SSC-32 servo controller to handle the hybrid legs. With five servos per leg and four motors, it will take up to 24 servo ports on the SSC-32 if I use PWM to control the motor controllers. With this new hybrid leg design, it may be possible for WALTER to adjust to terrain and maybe even do a bit of climbing.

8-Dale

73 from N7PKT

I received all the parts to build ARTIE as a regular ArduBot.  However, the PCB I got is not going to be workable for what I want to do with ARTIE.  The larger wheels and encoders aren't fitting the way I hoped they would. I've decided to go ahead and build ARTIE on one of those 5" Pololu chassis which has mounting for the wheels/encoders I have, reflective sensors, and proximity sensors. This chassis also has the possibility of adding decks for expansion, which is good.  I'll get the stuff I need for ARTIE from Pololu next month.

8-Dale

73 from N7PKT

This will be ARTIE's new chassis eventually. After I get him constructed and working in the standard ArduBot form, I'll get one of these chassis and an expansion deck and transfer ARTIE to it. This is one of the Pololu 5" round robot chassis which is designed to take all the components that will be ARTIE. This chassis is also available in several cool colors, so I can build several similar robots and have each one be a different color for easy visual identification.

Pololu also offers this cool reflective sensor array which will find a home on ARTIE. With this array installed on ARTIE, he will have the equivalent capability of the 3Pi robot for line following and line maze solving. ARTIE is basically my DIY 3pi and a whole lot more, with the ZigBee wireless communication link.

Here is a look at how the motors and encoders will go together, and a look at the parts kit.

Here are the motors for ARTIE, which are 100:1, 140 RPM. ARTIE will not be very fast, but he will be an investigator robot that requires slow deliberation. I can get motors that go all the way to 10:1 and 3000 RPM, which would make a very fast little robot. All these motors are interchangable for various purposes.  I will probably build a super speedy version of ARTIE next.

ARTIE will be powered by this cool High Capacity Lithium Backpack from LiquidWare, which fits snugly right under any Arduino or compatible controller board. This backpack is rechargable by plugging it into a standard USB cable! This still allows shields to be attached both above and below an Arduino board because it will always be the bottom of a stack. There couldn't be a more reasonable way to power a small Arduino compatible robot! I will be using these to power all my ARTIE type robots.

I am also getting one of the cool double wide extender shields, also from LiquidWare, for experimentation and placement on some Arduino compatible robots. This extender shield allows Arduino compatible boards to be attached to the top OR bottom, which maximizes flexibility in experimentation. I can't imagine expanding an Arduino compatible project any other way!

I also plan to get a ButtonShield for the user interface on ARTIE. I'll use the buttons for selection of various behaviors and/or tasks for ARTIE to perform. There are 26 buttons, and two mode keys, so there are many possible combinations for use in selecting behaviors and tasks.

8-Dale

73 from N7PKT

I have all the parts on order for ARTIE now.  This includes all the ZigBee (XBee) stuff to make him wireless and provide a wireless docking station for my PC. I'm very excited about this project because I will finally be able to experiment with robot swarming. The cost of participation in this aspect of robotics has finally come down to a point where I can build a group of small robots. Of course, both WALTER and ASTRID will also be a part of this when they are completed.

8-Dale

73 from N7PKT

I stumbled onto a very interesting DIY robot project I want to build. It's called the ArduBot. It is a robot very similar to the commercial Pololu 3pi robot. The big difference with the ArduBot is it is one you can build yourself, and I plan to do just that. You can customize it with pretty much anything you want to add to it, including sensors, wireless, etc.

In fact, I will be building my first Ardubot a bit differently from the Sparkfun Tutorial though. I'll be using slightly faster motors (24:1 instead of 30:1), larger wheels (42mm instead of 32mm), and adding wheel encoders. Using the encoders requires the larger wheels. Having the encoders will allow the ArduBot to know how fast it is going, how far it moves, and whether it is going forward or reverse. Encoders are a requirement to be able to do dead reckoning.

Unlike the 3pi, the ArduBot is designed to be 100% Arduino compatible.  In fact, the "brain" of the Ardubot is an Arduino or Arduino compatible controller board. I will also be using a slightly different Arduino type board for the controller of my ArduBot though.  I found the WiseDuino and RoboDuino. The WiseDuino has a battery backed real time clock, an EEPROM for data storage, and a place to plug in this XBee adapter for wireless communication. This will be ideal for my experimentation in robot swarming, and I am planning to build several ArduBots for this purpose. I wish the RoboDuino had these features, which would make the perfect robot controller for the ArduBot.

Maybe I can eventually create an RoboShield for the Arduino to add these features to the RoboDuino. I want to get a RoboDuino kit, so I can build it to allow for shields to be attached under it (instead of on top, like most shields would be added. Since the RoboDuino already has the nice 3 pin headers for attaching sensors and servos, I don't want to interfere with this, so the shield will have to go on the bottom.  I already have long pin headers that will all me to do this. I'll have to think about what other features I might want to go with a RoboDuino. This RoboShield would be able to attach either above or below an Arduino compatible board. I'd also want to add 3 pin headers for all the I/O pins in case it is attached above a board that doesn't already have them. I've been thinking about this RoboShield idea for quite awhile now, and I think it is finally time for me to do something with it. I'm already starting to learn how to use the Eagle EDA (schematic and PCB designer) software, which I'm now much more motivated to learn how to use.

All of my robots, including WALTER and ASTRID, will get XBee radios for wireless communication.

8-Dale

73 from N7PKT

I will add a wireless communication system to WALTER also, which will consist of WiFi-G and ZigBee. The AXON controller has three available UARTs (TTL logic level serial ports), so one will connect with a ZigBee wireless module, and another one will connect to a WiFi-G module. I'll be getting a ZigBee module (with this adapter) from Sparkfun and may get the WiFi-G module there also. This would leave one UART available for other things. I am not sure if I will need an SSC-32 now, since I've decided to remove the arm.

I'm still researching the WiFi-G to see what is available that might work best on WALTER. With this, there would be the possibility have allowing internet users to control WALTER in a limited way. With ZigBee, the area of swarm research will be opened, and WALTER will have the capability of communicating with other robots, like ASTRID.  I'll be putting the same communication system on ASTRID after she is up and running.

This will be the last complete rebuild I do with WALTER.

8-Dale

73 from N7PKT

I've been working on the new decks for W.A.L.T.E.R., and have them almost done now. I have created 3 new decks, and assembled them into a new 3D CAD model. So far, everything is looking pretty good and I am almost ready to make a test run and have somebody create them. I just need to find somebody who has a CNC machine that can accept data from a DXF file and create parts 11 1/2" in diameter. I don't want to spend too much money on a trial run of making the new decks.

I originally designed W.A.L.T.E.R. to have four decks.  However, I may stay with a three deck setup since I have decided to focus his development on just a few specific areas.

8-Dale

73 from N7PKT

After thinking about what I am trying to do with WALTER, and robotics in general, I've decided to remove the pan/tilt from the bottom deck and remove the arm from the top deck. I think I've been trying to do way too much with WALTER, and am going to orient him only to three specific areas now. Doing this will allow me to use both of the nice ball casters I recently purchased for him and solve the last mechanical problem I've been working on. I've been real unhappy with the placement of the new casters on the bottom deck due to the location of the pan/tilt.

From now on, I will orient WALTER completely for line following, maze solving, sensor experimentation, swarm research (with ASTRID), and behavior based robotics. With the AXON controller, I have the processing power and potential to do some really cool and interesting stuff in these areas. I think this will be the perfect fit for WALTER. I'm getting a Tracker sensor that I'll mount at the front of the bottom deck for line following.

This will be the last complete rebuild I do with WALTER.

8-Dale

73 from N7PKT

W.A.L.T.E.R. is now officially in pieces.  Many pieces!  Today I am connecting and testing the new electronics that will serve as the main control section.  I've resolved the power connection issues I had and will be routing all power from the SSC-32 servo controller instead of from the AXON controller. This is mainly because the AXON doesn't have power connections that are compatible with the battery packs I have. I'll have the SSC-32 setup with the VL=VS1 jumper installed and the VS1=VS2 jumpers removed. This allows me to power all the electronics from the VL/VS1 battery pack (7.2V@2800mAH NiMH) and use a 6V@2800maH battery pack for the VS2 supply which will power the servos.

I am looking at the possibility of using a RoboDuino (an Arduino compatible board designed for robotics) to handle reading several PING ultrasonic and a few Sharp IR range sensors, plus maybe the front pan/tilt where a pair of sensors will be mounted.

I'll adjust the spacers between the bottom and second decks to accomodate the new pan/tilt sensor base. The pan/tilt is already assembled and has a PING sensor on it.  I just have to add a Sharp IR sensor to that and mount everything to the bottom deck. I will move the batteries up to the second deck to put them right in between where the AXON/SSC-32 (third deck) and motor controller (bottom deck) are. This should make routing the power cables easier.

Today's main project is connecting power to the SSC-32 and routing power to the AXON so I can start tinkering with code. I seem to have misplaced my motor controller, so I have to find that also. The motor controller and motors will be powered from the 7.2V@2800maH battery that also powers the rest of the electronics and most sensors. With this setup and powering servos from a separate 6V@2800mAH battery, I should get a reasonable run time with W.A.L.T.E.R.

8-Dale

73 from N7PKT!

I've started working on the new control code for W.A.L.T.E.R., in anticipation of receiving the new Axon controller this coming week. I want to be able to start tinkering with the AXON as soon as it arrives. The first part of the code I am working on is what is required to read the PING Ultrasonic and Sharp GP2D12 IR sensors I have.  These will be the workhorses of W.A.L.T.E.R.'s sensory perception. I like to mount these as a pair because what one doesn't see, the other one usually will. This pairing increases the chances that W.A.L.T.E.R. will see an obstacle.

Since the existing routine for processing the PING doesn't do things quite the way I want to do it, I have cloned the existing routine and renamed it so I can modify it to do things my way. I will eventually mount several of the PING sensors in various locations around the body and want to be able to process them all inside a loop. If this works out as I hope, I'll create a routine for the GP2D12 that processes them in the same way.

One pair of sensors will be on a front mounted pan/tilt base, which will allow for coverage of the area in front of and slightly to either side of center that isn't covered by the sensors on the sides of the robot. It will also allow seeing down, which I hope will help keep W.A.L.T.E.R. from falling down stairs and off cliffs.

Once I have completed the sensor processing code for the PING and GP2D12, I will be able to move on to writing motor control code for the Sabertooth 2X5 that controls the two GHM-4 gear motors that move W.A.L.T.E.R. around. Once this is completed, I'll have all the basics for W.A.L.T.E.R. and can move on to working on actual behavior code that will control how he sees and interacts with his environment.

8-Dale

73 from N7PKT

I finally did it!

I've selected and ordered the first of what will be several controller boards for W.A.L.T.E.R. I've been looking at the Axon controller for awhile now, but jave not purchased anything due to there being so many good controllers available. I am extremely picky about controller features due to some not so good experiences in the past. However, the Axon does have the features I need for W.A.L.T.E.R., at least for now, and can be used as an auxilliary controller in the future when I decide what I want to use for the on board Linux based main controller. For the time being, after I get a bluetooth tranceiver for the Axon, my main Windows/Linux PC will be the main controller.

Essentially, the Axon will be the main controller for W.A.L.T.E.R. for now as I learn to program it and use all of its features to the fullest. The Axon is programmed in the 'C'/C++ language, which is a major requirement. I'm through using proprietary hardware that requires me to be at the mercy of the company who makes a controller, their proprietary development system, and proprietary programming language! From now on,  I require 100% Open Source and standard development tools for W.A.L.T.E.R. and any other robot I may build in the future. My main programming language on the main controller, whatever it ends up being, will be Python, which is also 100% Open Source as well as platform independent. Python is available for Windows as well as Linux, UNIX, and most other operating systems.

The only other type of controller I would consider is one that has its development system integrated into the board itself, such as the ServoPod.  This is definitely another controller I want to work with in the future.

My next step is to get a few pictures of W.A.L.T.E.R. in his current disassembled state and get them online here or on a Wiki, to start tracking my progress for version 2.0. I'm quite excited about being able to actually start working on everything for W.A.L.T.E.R. v2.0 now!

8-Dale

73 from N7PKT

OK, it's time to stop thinking about what I want to do with W.A.L.T.E.R. and start doing something about it! I have him down on my work table, got Python installed, and am ready to start getting back to making things happen. I have a lot of rewiring to do to get ready to make W.A.L.T.E.R. operational again, and he still needs a new brain, but I am getting closer to seeing him roll around again. Python is an interpreted programming language that is great for many things, including scripting and prototyping of applications. I like to use it as a control language because changes can be made and the results seen right away. I hope to use Python as the main programming language for W.A.L.T.E.R.'s software.

I still don't know what I am going to do about casters for him, because there aren't any stock casters that will work properly. Ideally, I would like to use two omni-wheel casters, one on the back and one on the front, but this would require some custom design and machining which I am not able to do. Well, I can do the designing that would be required, but my PC isn't currently up to running my 3D CAD software at this time. That would still leave the machining. I'd want an adjustable design that would allow the casters to be made the proper height and allow that to be changed if I use different sized wheels or otherwise affect the height the casters would need to be at.

I'm definitely convinced the way to go for processing power on W.A.L.T.E.R. is to have one master processor with a lot of power that has several smaller processors for given functions, such as sensor processing. One processor might even be assigned to a single sensor, depending on the sensor's complexity. I'll be tinkering with Arduinos and Sanguinos for at least some of these tasks, and I have one of each here now. There is also the onboard arm which will require dedicated sensor and movement processing. I'd like to mount a small camera on the end of the arm near the gripper.

For now, I have some untrasonic (sonar) and infrared sensors, as well as an accelerometer and digital compass. I want to eventually get a GPS unit for W.A.L.T.E.R. when he ventures outdoors. I have quite a bit of stuff to tinker with right now and it will take some serious time to integrate everything. I'm planning to pair an IR sensor and an ultrasonic sensor, which should give the best chance of "seeing" any object for distance detection.  If one sensor misses something, the other one will usually "see" it.

I'm going to be changing the deck setup for W.A.L.T.E.R. I've designed a new pan/tilt unit that I'll mount on the front which will have a +/- 60 degree range of motion horizontally and +60 degree and -90 degree vertical motion. This will allow W.A.L.T.E.R. to "see" with a very wide range. There's also going to be two pair of sensors (IR and ultrasonic together) mounted fixed that will point off to the sides about 30 degrees either side of center. Beyond that and the accelerometer for tilt and speed, compass for course, plus eventual GPS, I only want to add a thermopile (heat) sensor for detection of warm bodies (like my cat and Humans).

Well, that's my story and I am hoping I can stick to it!

8-Dale. 73 from N7PKT.

I was dinking around checking various repeaters I programmed into my radio today and happened to be listening to the 442.525 machine. It turns out there is an ATV net that meets there every Saturday at 11:00AM and Monday at 7:00 PM Pacific. This is the Oregon Amateur Television Association.They also have a streaming ATV feed that is active during their nets. I couldn't interact much because I can't seem to get a good signal into either of the repeaters that are supposed to be near where I live, but I was able to talk to one of the guys on 146.43 FM Simplex after the net.

One of the things I am very interested in doing with W.A.L.T.E.R is to have a video feed coming from him. Now, I realize there are several ways to accomplish this without using Ham radio, but would that be any real fun?? I started doing some research into hardware that might make this possible and did locate some small low power ATV transmitters and receivers that might just work for this application. Of course, I would only need a transmitter on W.A.L.T.E.R. with an appropriate camera and antenna. This will be a project for the long term though and nothing I plan to actually work on anytime soon.  I just want to get it down as an idea.

73 from N7PKT,

8-Dale