Some time has passed since my last post. Two reasons for that, first that I've working in a preliminary report of my work here in Cambridge and secondly I also spent some free time travelling around ;). Now, in this post I make a review on the companies which already have
cognitive radio commercial products. I started to write this post some time ago, but I just found
a post in CRTWireless blog commenting that James Neel will be speaking about this at the
ICO-COST 902 meeting November 23-25 in Bologna. From his post:
"I take the position that CR is already a commercial product (e.g., xG’s cellular stuff, Cisco’s CleanAir, the CRC WiFi system, 802.11 h,y, self-organizing networks in 3GPP, Cambridge’s stuff)"
Nice to see that I'm not the only one thinking this way. I will start the review with xG Technology, since their product is in my opinion the closest to market. In fact
they have already signed a contract with a military contractor to license the technology xMax.
xG Technology: xMax
With xMax
xG Technology translates the concept of mobile phones to the (unlicensed) band of 900MHz. The core of xMax technology is it capability of sensing for other systems in the band and determine if interference has reached unacceptable levels, and in such case change band. This process can be carried out up to 33 times a second. The main practical problems a product of this kind has to deal with are synchronization issues and robustness of the detection scheme.
While xG Technology webpage mainly offers a
high level description of this cognitive radio technology, it also points out some characteristics of the physical layer:
- Handover decisions are made by fussing multiple samples and measurements. This allows the system to avoid unnecessary band switches due to a temporary interference or degraded network conditions.
- Compared to this the modulation is very simple: xMax uses a conventional single carrier modulation with BPSK modulation. However it allows the bonding of multiple of these single carriers channels working in different bands. Then, while the technology supports up to 1Mbps per channel with BPSK modulation, it allows the use of 18 RF channels simultaneously, and thus it can achieve up to 18Mbps. Higher rates could be achieved by employing adaptive modulation instead of BPSK.
Spectrum Bridge / Microsoft: Occupancy Database Solutions
Other companies, like Spectrum Bridge, have a different focus. They offer
database solutions which will be required in order to implement the unlicensed access using geolocation devices. However this is
yet to come. In the mean time we can already
play searching the spectrum for white holes.
CRC: CORAL
The
Communications Research Centre Canada has developed a
commercial WiFi-based cognitive radio development platform:
"The world's first commercial WiFi-based cognitive radio development platform is available from the Communications Research Centre Canada (CRC). The system will be of interest to both researchers and wireless Internet service providers building multipoint relay and other types WiFi networks. Called CORAL, the system can undertake radio interference sensing and autonomously adapt to the sensed interference."
Cisco: CleanAir Technology
Cisco's CleanAir Technology allows the wireless nodes to build an interference map and reconfigure the network to optimize the performance. This technology applies even for the widely used unlicensed 802.11n networks. However, while the resulting systems are self-healing and self-optimizing, this process is probably not agile enough to be used with licensed bands.
IMEC: Cobra
IMEC is a research company focused in nano-electronics.
Recently they launched a cognitive baseband radio (
COBRA) high performance architecture targeting 4G requirements at up to 1Gbit/s throughput and multiple asynchronous concurrent streams.
Note however that this project is just a reconfigurable physical layer chip and it is in wireless prototype stage. Hence, if successful it could be used in the future to build actual 4G cognitive radio systems.
Adapt4
Adat4 commercializes proprietary data transmission systems in the 217-220 MHz band which can be used for surveillance, monitoring...
Adapt4's products claim to be pioneer in implementing practical, cognitive radios.
"The XG1's and frequency management feature enables the radios to constantly scan the entire band to look for and avoid interference from other users (including other secondary or primary licensees). Should a frequency in use by the XG1 suddenly become interfered with, the radio will instantly select a free channel from its constantly updated Free Channel List. Each of the radio's 1 to 45 channels is individually managed in this way, so that segments of bandwidth throughout the entire band can be used. And to minimize further the possibility of extended interference, the radio will constantly change its frequency even without interference to minimize its profile in the band."
However, they offer low data rates (30 to 180 kbps of data throughput) and they do not specify how sensible and agile this detection process is. Note that most of current systems already implement a "dynamic channel selection" in the configuration process.
Here I covered just a few of the companies working on projects related to cognitive radio. A more exhaustive list can be found on the
Wireless Innovation forum Member list.
Cognitive Radio Standards for unlicensed access
We have seen different the approaches to Cognitive Radio by different companies. However, there exits an important standardization effort to avoid finishing with many proprietary non-interoperable systems. Here one slide of the talk by James Neel and Jeff Reed at the Atlantic Council (Oct 29, 2010)
"Second Wave of Wireless Communications" that summarizes some of these efforts.
I already wrote about the
802.16h (the cognitive Wimax) in a
previous post. I will also try to go over the main points of the remaining standards, with special focus on
802.11af and 802.22. Since the intended market for the last one noticeably overlaps with 802.16h, I'm specially curious about the differences between these two.
Labels: cognitive radio, hardware