Difference between revisions of "WSN development overview"
From emboxit
(→Synapse) |
m (1 revision) |
(No difference)
|
Latest revision as of 16:51, 11 September 2015
Contents
Overview of Technology used for Wireless Sensor Networks (wsn)
- Self Organizing, Self Healing, FCC approval
- CC2500 has not 802.15.4 MAC. Contiki support: Only a small project for 2 nodes. Simplicity to be tested, TinyOS to be checked
- Squibdee openhardware project from Libelium is 'support of Arduino + XBee'. To be used as software example. ArduinoFIO is better design.
- PowWow based on MSP430 and CC2420 2.4GHz 802.15.4
- This combination is well supported by TinyOS, Contiki
- Very 'light' (in terms of Software)implementation of mesh WSN :)
- CC2420 board not easy to get in good price :(
- ATmega128RFA1
- Is supported by Contiki (TinyOS to be checked)
- Used by Synapse in series RF200
- Excellent performance but work only with their proprietary firmware.
- Used by Dresden in their modules.
- Dresden modules start at 20EU with good performance and FCC approval.
- JTAG and SPI connectors for Firmware download. They are open for Firmware Development
- Disadvantages of ATmega128RFA1: Chip is new, Contiki, TinyOS support might not be mature
- It is good idea to create Altium libraries for ATmega128RFA1. Not too hard to design a module, but no FCC and other approvals.
- Trends on wsn modules:
- MSP430 or AVR + CC2420 or other RF is the 'classic' solution. Single-chip is preferred for new modules
- Single chip ATmega128RFA1 for the low cost modules
- Single chip Freescale ARM7 MC13224 for the performance modules
- New star is single chip STM32W108 ARM-Cortex + 802.15.4. Included in Contiki Platforms but not yet in known commercial module
Second scan
- Mesh network is the choice due to self-healing and range.
Open source
- Open source environments are appropriate for University and European-founded projects
- Very interesting but big resources needed to manage the project
- ITEM [check below] project at Czech Technical university, is the most complete project found up to now...
- Libelium [check below] commercial products based on open-source
Dust networks
- Dust Networks [check below] has the best documentation on the wireless-HART protocol which is standardized by ISO.
- The only provider of wireless-HART, their devices have best characteristics.
- But Dust Networks does not have any distributor. No price and availability info. Their competition says they are very expensive!
DIGI
- Digi [check below] XBee modules are sold everywhere at logical prices
- Many public projects available
- Digimesh protocol is competitive to wireless-HART
- Used by intelligent-devices, Libellium
- Still the best starting point
Synapse
- Synapse [check below
- Can not support sleeping mesh
- Good documentation
- They do not describe the details of how the SNAP protocol works, in the detail Dust-networks does.
- No CPU needed, Python Engine is included in the module.
- Over the air Python application download
- Nodes are engaged in the mesh network instantly at power-on, out-of-the-box
- Different modules available with nice price/performance
- Very good pricing. The lowest cost module at $20 no CPU needed, very good low-power and range characteristics, SNAP-network-protocol and Python-Interpreter included.
SimpliciTI
- TI very simple OS for wireless networks
- API and source code for MSP430
- Peer-Peer capability
- The route Layer has to be developed, to implement the mesh, based on the peer-peer.
ANT
- ANT [check below] WSN protocol used for sports devices
- Professional approach, good network characteristics
- As it is designed for peripheral devices of an athlete, the range might be LOW.
- Chips by Nordic for some years and TI starting in 2011
- Big names including Nike are using it.
- Available boards only from Sparkfun which is Hobby-electronics oriented.
First approach (check newer text titled: second scan)
- Most used CPU is MSP430. Mostly BIG devices with 60-90KB Flash
- Second used CPU: AVR. Here also large Versions with 128KB Flash are used.
- Most used OS is TinyOS, second is Contiki.
- Most used RF module is XBEE.
- ZigBee network implementation.
- Used by amateurs and professionals.
- Does not need a wireless Operating system.
- Does not need a large processor.
- Book Building Wireless Sensor Networks with ZigBee, XBee, Arduino, and Processing
- To check if already tested TinyOS to control XBee via AT-commands.
- To check Arduino-FIO+XBee and over-air sketch download.
- Most Used RF chips:
- TI CC24xx 2.4GHz ZigBee devices with good support from TinyOS and Contiki
- TI CC11xx Sub GHz devices with TinyOs support but not as good as above
- A large serial flash (~8Mbit) is used to store the golden-image and the new over-air downloaded images of the firmware
- TinyOS supports over-air programing.
- TinyOs evaluation and implementation has a rather long learning curve.
- Companies that make Wireless sensor boards, have also ported and give to customers their version of TinyOs and Contiki.
Overview Links
- University Bern Presentation Seminar WS 06/07 Sensor Nodes Markus Anwander
- EET article: Wireless route to energy efficiency
- ONE-NET open-source, simple, good documentation, Star,Peer-Peer from Wikipedia
- ONE-NET Home
- TI SimpliciTI simple star, peer-peer, available source page
- FreakLabs opensource wireless home
- Jennic now a NXP company from Wikipedia