Where are the large collaborative projects?
xanadu
Posts: 3,347
What's wrong with this picture? We connect every device in the house to the internet except our microcontrollers (give or take a few people).
We all have everything to build personal nodes that could connect via the internet to other nodes from like individuals. The nodes will perform some functions to make it worth while.
The two things screwing this up are:
1. Basic networking understanding
2. Nothing to connect to
If both of those issues were resolved would you participate in connecting with a network of microcontrollers via the internet?
What about a private IRC channel and Spinnerets?
Ideas on a protocol to use for communication are needed. I would pre-pay a 10 year contract on the rackspace to run whatever server is needed as a proxy. I think it that the internet isn't expanding enough into our market as fast as it should be. I've written multi user php software to control robots from a web server's serial port and never generated any interest. There is no internet based embedded network for everyone. We could lay the foundation for it if we wanted.
If you don't think something like this is worthwhile for anyone I would like to hear that as well. I hope that is the reason and not the two items above that could be easily resolved.
We all have everything to build personal nodes that could connect via the internet to other nodes from like individuals. The nodes will perform some functions to make it worth while.
The two things screwing this up are:
1. Basic networking understanding
2. Nothing to connect to
If both of those issues were resolved would you participate in connecting with a network of microcontrollers via the internet?
What about a private IRC channel and Spinnerets?
Ideas on a protocol to use for communication are needed. I would pre-pay a 10 year contract on the rackspace to run whatever server is needed as a proxy. I think it that the internet isn't expanding enough into our market as fast as it should be. I've written multi user php software to control robots from a web server's serial port and never generated any interest. There is no internet based embedded network for everyone. We could lay the foundation for it if we wanted.
If you don't think something like this is worthwhile for anyone I would like to hear that as well. I hope that is the reason and not the two items above that could be easily resolved.
Comments
So here is the challenge. Wifi enabled board, under 1 watt consumption and under $10. Four real world sensors - one digital input, one analog input, one analog output and one digital output (connected to a relay).
Does this exist yet?
We can get a WiFi box (WR703N) for ~$25 that can run *nix and probably do the hub end plus an interface to the cheap wireless.
The NRLF2401 are nice and cheap ~$1.50 ea. Now we just need a simple interface to this and a simple protocol, not the complex IP protocol.
Drac: How did your wireless networks work out?
My vision is basically a system where you can plug in a Propeller to the internet and be able to send and receive information from other people with similar setups. Robot control isn't really something I can see people doing a lot, however this would potentially allow the best way of doing so if you changed your mind.
"Internet based embedded network for everyone" would be the end result. Essentially you could either provide a resource such as storage, or offer up information such as sensor data.
Anyway the point is it would be large, and hopefully grow to thousands of nodes.
Lets you connect and display your sensor data if you want.
http://www.myrobots.com/
That will just give the Cylons a way to take over.
C.W.
Devices that can be reprogrammed wirelessly are especially scary.
C y l o n s
WiFi enabled under $10? Not going to happen soon in the hobby market. Perhaps embedded inside an appliance (or toy)
That's funny, last time I went on this rant it didn't exist. I'll have to check it out thanks!
What actually are you asking here?
Before I put that lamp in my bathroom on the internet I need it to have enough smarts to support TCP/IP.
But then of course I need it to be secure, say with TLS. I don't want anyone messing with my bath room lamp.
Perhaps I don't want to run cables to my bathroom so my my lamp had better also support WIFI.
Well, I can do all that already with a Raspberry Pi and a WIFI dongle.
But then, why would I do all that?
They have a serial protocol http://www.myrobots.com/wiki/Open_Serial_Protocol
@Cluso, no I haven't solved this. If we can crack it then it may well be a step towards the internet of things. Right now, the costs are too high, and the power consumption is too high. All the components exist though.
Power consumption: Even though 1W doesn't sound much, consider that a NiMh battery with 2000mAH and 1.2V can only store just over 2 Wh, so the battery is going to go flat in two hours. Fine for mains powered devices but not so good for robots. The answer is probably along the lines of having a mains powered node, and then mobile/battery powered nodes that wake briefly, talk, then go to sleep. For a mobile robot transmitting position data, it can wake any time. For receiving commands, it may need accurate timing eg - "you go to sleep, we'll talk again in 30 minutes".
Radio can be wifi on the robot, or 433Mhz serial on the robot and then serial to wifi/wired ethernet at the node. The 433Mhz solution is only a couple of dollars, so until wifi gets that cheap, this is the cheapest option.
myrobots looks like they are using 9600 baud, and that would fit nicely with various 433Mhz modules. The data packet is fairly long - that may preclude some modules. I don't think it will work with picaxe chips, and propeller chips are more expensive. Many raw RF modules work at 1200 or 2400 baud as they are working at audio frequencies.
Next challenge is data clashes. You can cheat and use multiple frequencies but then the powered node needs lots of different receivers. Or you can use a smarter protocol on just one frequency, and poll robots in turn. The robot code is fairly simple - arduino or picaxe sort of simplicity. The powered node gets a bit more complicated - not quite PC complexity, probably something ideally suited to the Propeller.
For absolute simple low cost and low power, it could be possible to get the robot node down to under $10. Use a picaxe and a raw 433Mhz module. Assume sparse transmissions of only a few bytes (under 16) and assume that there will be occasional data clashes when robots transmit at the same time.
The powered node needs to handle talking to the ethernet and handling all the data from the robots. The myrobots protocol could be the answer. They are using xbees which are a bit more expensive but might be able to do meshing more transparently. The powered node is going to be more expensive.
For long distance runs, you can up the wireless power but that flattens the batteries. Some other solutions have become cost effective, eg using ethernet over powerline modules eg Netcomm NP205 and similar part numbers. So if you have power, you can set up powered nodes and then have a cluster of robots within 50-100m of that node.
One real challenge is data clashes. Two robots transmitting every half an hour is fine. 20 robots transmitting every minute is not going to work. So then the powered node is going to need to arbitrate. Maybe it needs to tell each robot exactly how many seconds it needs to wait before transmitting again?
myrobots are on the right track, but their costs are too high. The wifi shield and arduino is $109. Doing the RF with 433Mhz modules can bring that down under $10. There could be scope for an interface that takes the myrobots protocol and chops up the packets into something a picaxe can handle and then reconstitutes them at the other end. Cheapest solution is something like this http://www.picaxe.orconhosting.net.nz/txrx433.jpg but RS232 modules have better receiver sensitivity http://www.picaxe.orconhosting.net.nz/txrxconf.jpg was one of the early experiments but there are better modules around now.
Any solution for robots would also apply for things like fridges, temp sensors, working out when the dog's water bowl is empty etc.
It's not about remote control and the associated risks. You could be read only. You could do whatever you wanted if the option existed. It's a framework for a mesh network using standard network protocols. Sharing sensor data is just the beginning.
One large collaborative project is propforth, but that is not what you are looking for... But one part of it might be.
Dlink DIR505 might be similar to what you want. It connects to the prop over serial, same as RPi or Beagle bone (or whatever its called) without the unused HDMI, and with WiFi. It is cheaper and more powerful and more flexible than spinnerette. It is a full linux system, internet, python etc, also router; you can connect to your home network via wifi, or you can accept connection from you laptop or other wifi device.
Even though we use propforth, any microcontroller that supports serial should be able to connect to the DIR505 reguardless of language or environment.
They cost $30, so its a bit over your stated price.
But this is the direction we are going, in addition to using android device as generic user interface (when PC is not practical). Came along if you want to, it will be fun!
There is no price requirement, $30 is reasonable. Can you confirm this is what you mentioned above? - http://www.amazon.com/D-Link-All-in-One-Mobile-Companion-DIR-505/dp/B0081TXIQA
There is also the Dlink DIR505L. The L model has a recharable battery. I think the battery would be nice, but I think I read the battery code is a bit touchy on the open source hack.
We just started on this part of the project, so Sal has not made the code braino-proof yet; so I don't have any instructions posted yet. I think Sal just got the cheapest one (without the battery) to start with, so that is what I will be ordering. If it work out like we expect, the support for the battery should not be a big hurdle.
We have to start with the load load of the open source version of the Dlink code for the 505, and the OpenWRT code. I plan to write the instructions as Sal walks me through the process, to ensure that the instruction match exactly on both units. it makes testing easier. We should be starting in the next week or so, if anyone is interested in following along.