ZigBee controlled RN-1

[bauermech]

While at RoboGames, Dan Albert and I were in charge of making the soccer teams comprised of eight RN-1's work. Dan brilliantly designed a programmable break-out board for the XBee. We placed one inside the slightly modified RN-1 head, and the other in his modified Play Station controllers. We wrote some control handling code, soccer routines, threw in a couple of gyros, and we were set. I must admit that with everything going on, the soccer routines (as well as my own RN-1's routines) could have been much better, but it turned out to be quite a success among the audience none the less. Dan Albert, David Culkins, and I made up ComBots Team USA and took the gold against the UK, Brazilian and Columbian teams. We all used the same two sets of three RN-1 players to compete, and simply changed their cute little jerseys to differentiate between the countries. The games were slow-paced, but ultimately, everyone had a great time. Only in Robo-soccer could the Americans do this well

I know Dan is in the process of making the break-out boards available for purchase soon, so keep a look out. They work great and are very easy to implement into the RN-1.

Sorry I don't have any pics or video yet... I left my camera at RoboGames So, if anyone picked up, or happened to see a camera in the vicinity of my table after I raced into my cab, PLEASE let me know!

[Chrisbie]

I love to see a video of that... Any idea how much the module cost? Hope its not expensive...

[bauermech]

The XBee is about $20 for each module... plus Dan's break-out board. He used the analogue red PS controller.

Good news is I have one tape in my possession crammed w/ video, but nothing to play it back on yet.

[DanAlbert]

Thanks for the kudos Matt.

RoboGames is the BEST!

If you haven't been there you haven't lived.

I hope to soon have the ZigBee hack available to purchase.

I'm not sure if people want the full PlayStation hack or just a wireless RS232 solution.

Any help on this forum would be ....well .... helpful.

Here are some pics of the ZigBee hack we used.
Left is the RoboNova ZigBee Receiver. (Goes inside the head)
Middle is the PlayStation conversion module. (Converts the controller signals to Ascii codes)
Right is the head expansion plate used to widen the head to accept the receiver module.

Two of the head modules could be used to convert TTL serial to wireless and back again. Good for boards like the SSC-32 and such.

I will be making a new board that will have a max232 level shifter on it and a DB9 connector to connect to the PC. Maybe a USB port? Guys... give me some feedback on what you want.

I think the cost is less than Bluetooth. Boards will sell for about $50 each.

Go Team USA...The Gold again in 2007 .... Yea

[robodude666]

Mmmm nice nice, but I am not 100% sure what this is, or what this "bee" stuff your talking about is. As I understand, a hack to use the playstation controller to control the robot? or as I read it again to make the robot play soccer... eH?

First feedback, can you make the image as a link? or a smaller image linking to the bigger one? Not all of us have a 1600 x 1280px screen

Sounds good, cheaper than bluetooth is always good $50 wireless board is pretty good. Though I seen some bluetooth ones for like 60 I think.

[limor]

Hi Dan,

Have you managed to communicate over ZigBee at full 115kbps and to what distance? also, what would you say are the main advantages over Sparkfun's BlueSmirf bluetooth module?

I think that in order to sell maybe a couple of hundred of these units, they should have the following features:

• come in a package that's "plug-n-play" without requiring the user to solder anything
• Price should be benchmarked against alternative bluetooth and zigbee solutions that may require soldering, voltage conversion etc. and also, Hitec will come out soon with a bluetooth module but the price will probably be around 100EU ($130)
• CE certified (I saw this mentioned on MaxStream's site so it should be ok)
• the package should probably come with two boards, one attached to a serial connector for the PC (or gamepad?), and a second to a headphone jack for Robonova and Robotis or optionally to a 3-pin kondo connector for KHR-1&2 controllers. ie: to all 3 popular models (haven't seen the RB1000 serial connector yet).

Assuming this is realizable, RoboSavvy would be interested in participating in a batch production of these zigbee boards maybe in conjunction with the modded gamepad. Lets discuss this over email

[tempusmaster]

Hi Dan,

Have you managed to communicate over ZigBee at full 115kbps and to what distance? also, what would you say are the main advantages over Sparkfun's BlueSmirf bluetooth module?

I think that in order to sell maybe a couple of hundred of these units, they should have the following features:

• come in a package that's "plug-n-play" without requiring the user to solder anything
• Price should be benchmarked against alternative bluetooth and zigbee solutions that may require soldering, voltage conversion etc. and also, Hitec will come out soon with a bluetooth module but the price will probably be around 100EU ($130)
• CE certified (I saw this mentioned on MaxStream's site so it should be ok)
• the package should probably come with two boards, one attached to a serial connector for the PC (or gamepad?), and a second to a headphone jack for Robonova and Robotis or optionally to a 3-pin kondo connector for KHR-1&2 controllers. ie: to all 3 popular models (haven't seen the RB1000 serial connector yet).

Assuming this is realizable, RoboSavvy would be interested in participating in a batch production of these zigbee boards maybe in conjunction with the modded gamepad. Lets discuss this over email

Limor,

Good points, I agree completely.

While I haven't hacked the ZigBee into my KHR-1 yet, I have had a lot of experience using it with the Surveyor robot platform and really like the performance. I've been able to operate the Surveyor reliably several rooms away - probably over 50 ft. (that's a rough guess, I haven't measured it) with intervening interior walls.

Also, ZigBee is already on the approved list for the ROBO-ONE competitions, so I'm assuming it has all the Japanese certifications already.

I will definitely buy one, perhaps a few more as well, if the two of you go ahead and make it a real product that is useable with the KHR series.

[robodude666]

Hmmm,

So a wireless RF module, eh? I would agree with limor. It would be great if it were "plug-n-play" as I do not like to solder on expensive stuff. As for its 115Kbps.. Does it really matter? Most files youl be sending across will be a few kb and if you want to send variables across or something they will be a few bytes or bits. Not like your going to be sending 1MB of RoboBasic! Plus the RN can hold only like 512kb I think...

Would this only work with the stock board for the RN, KHR, and RB? Or can it work with other boards such as ServoPod or OOPIC, or maybe even that new POB-PROTOBOARD?

Would be great if you can give us some of the basic features of this new product. Once I buy me my RoboNova next month (almost at $1000 ) I may need a wireless module. As I am sure many people would like a wireless module but don't really understand how to attach the Sparkfun Bluetooth SMiRF module.. Umm, think thats what it is called.

[limor]

As for its 115Kbps.. Does it really matter? Most files youl be sending across will be a few kb and if you want to send variables across or something they will be a few bytes or bits. Not like your going to be sending 1MB of RoboBasic! Plus the RN can hold only like 512kb I think...

The KHR-1 robot board (RCB-1) has only 2 speeds it can work at 115kbps and 9600bps. The slow speed only allows for motion sequence program execution and not for programming and other features. So it would be preferable if the ZigBee can sustain 115kbps (and the specs say it can). KHR-1 has sold some 500 units outside of Japan and some 3500 in Japan so there's a market out there of KHR-1 owners that would like to see it loose its umbilical cord.

[robodude666]

Ahh, I see. I didn't know that. Sowwy ^_^'' Well, looking foward on the progress you make with this Dan.

[DanAlbert]

The real key fact here is that not one vendor that I am aware of has actually 100% followed the ZibBee standard. They all sort of have their own flavor. MaxStream is no exception. However that being said. MaxSteam has a great bunch of techs. They have some beta firmware that will push the speed to 115200. Although I didn't see the need for this speed. It's always nice to have it. The module I used easily had 100ft range. It is reported to have several hundred feet range but I couldn't see my bot after 100 ft.
The pro model has been reported to work well over 1.5 miles. Considering the very low current (100mA) requirements this is amazing. It blows away BlueTooth and WiFi for most data communications. Yea, sure if you need to stream video go with 802.11G.

I have been asked by a few people to sell them a version of this hack.
Currently I am trying to make it more bullet proof. It's too easy to have someone hook up a 5 Volt (ior higher) line to a 3.3 input and fry the chip.
I estimate about 3 weeks until I have some product. Yet I still don't know whether or not to include the PlayStation 2 controller. I found it a real easy interface for controlling a biped robot. It's that stupid BozoBasic that the RoboNova comes with that makes it so difficult to program.
We need to rewrite it in C. Wish we had some schematics.

[subpilot]

I have my Wibotix WiPort 802.11 board working and I'm really happy with it. It gives me two extremely high speed serial ports, Ethernet, CAN, I2C and 11 GPIO pins. It also has exactly the same footprint as my Servopod controller. The drawback is that it takes 300ma @ 3.5V to run it.

I have been messing with the Servopod and it is really looking promising. I can control all my servos and I have working interfaces to SRF10 rangers and my accelerometers. I don't have a clean way to get servo position yet other than doing a bit bang polling scheme but I'm working on it. I think the best method may ultimately require adding a mux to the board in order to get signals to the timer input pins.

[DanAlbert]

The ServoPod looks like a good product. (it even has Forth)
Which servos are you using? Why do you need to bitbang them? Can't you set up an ISR via a timer?

[subpilot]

The only Forth I had any experience with was on my HP calculator so it's a bit of a learning curve. The ISOMax OS has a really simple way to measure the servo position pulse width but only applies to the 14 timer pins. Once I get a bit more up to speed I'm sure that I can probably use an internal timer but I also need to see if you can generate an interupt from all the servo I/O lines.

[DanAlbert]

The neat thing about forth is that you can create new commands from old ones in a threaded manner.

Unlike C++, Forth is really small and efficient and very easy to understand.
Except for the Reverse Polish Notation stack usage.

As far as getting an ISR from the multiple servos, i would think you need to put the all the pins on a logic analyzer and see what the timing and waveforms look like. I have a digiview that cost about $500.00. The best money I ever spent.

I may want to get a ServoPod for my homebrewed biped. I have an SSC-32 but I can't program it. Are you happy with your's.