This entry focuses on a few implementations of CR published in various top journals and conferences.
Cognitive Radio Software Testbed using Dual Optimization in Genetic Algorithm Jae Moung Kim; Sung Hwan Sohn; Ning Han; Guanbo Zheng; Young Min Kim; Joo Kwan Lee; Cognitive Radio Oriented Wireless Networks and Communications, 2008. CrownCom 2008. 3rd International Conference on
This implementation use a 2 step sensing process:
1. Coarse sensing, this is done very fast using a sliding window FFT (energy detection). It is used to identify channels
2. Sign sensing, this is used to confirm availability. This is done using Cyclostationary algorithm
The other interesting point is that the algorithm used for DFS is based on Genetic Algorithms
A Cognitive Radio Receiver Supporting Wide-Band Sensing Blaschke, V.; Renk, T.; Jondral, F.K.; Communications Workshops, 2008. ICC Workshops ‘08. IEEE International Conference on 19-23 May 2008 Page(s):499 - 503
Not as much a CR implementation, but it presents some practical ways for wide band sensing.
* Signals are sensed by finding the average received power
* Power allocation is found using FFT
A Cross-layer Cognitive Radio Testbed for the Evaluation of Spectrum Sensing Receiver and Interference Analysis Jongmin Park; Kwan-woo Kim; Taejoong Song; Sang Min Lee; Joonhoi Hur; Kyutae Lim; Laskar, J.; Cognitive Radio Oriented Wireless Networks and Communications, 2008. CrownCom 2008. 3rd International Conference on 15-17 May 2008 Page(s):1 - 6
This article presents a reconfigurable CR test bed. The system is completely software drive, and the advantage of that is it can instantly deploy different algorithms for testing. This can be interesting for my project, as if I set up a testbed, I can also test out all the algorithms there are, and maybe publish the results.
From this article, a few things were noted:
* We can simulate real condition by adding AWGN to the spectrum
* We should also consider channel fading
The system deployed here also uses Multi Resolution Spectrum Sensing (MRSS), which is a analog based, digitally assisted energy detection method. The system is entirely analog hardware, but does not need tunable filter to change its detection bandwidth. MRSS also do not require FFT nor ADC, and provide almost real time data.
I think MRSS is definitely worth looking into some more.
Integrated Resource Management in Cognitive Radio Marojevic, V.; Vucevic, N.; Reves, X.; Gelonch, A.; Mobile and Wireless Communications Summit, 2007. 16th IST 1-5 July 2007 Page(s):1 - 5 Digital Object Identifier 10.1109/ISTMWC.2007.4299309
This article is stored for future reading. The contents does not concern me at the moment, I just thought it would be a good article to read
Cognitive Radio Prototyping Jung, P.; Viessmann, A.; Spiegel, C.; Burnic, A.; Bai, Z.; Bruck, G.H.; Statnikov, K.; Waadt, A.; Wang, S.; Popon, X.; Velilla, R.R.; Saarnisaari, H.; Alles, M.; Brack, T.; Kienle, F.; Berens, F.; Rotolo, S.; Scalise, F.M.; Wehn, N.; Cognitive Radio Oriented Wireless Networks and Communications, 2008. CrownCom 2008. 3rd International Conference on 15-17 May 2008 Page(s):1 - 6
This article is not very useful, the prototype suggested is too advanced for my level to even attempt to understand
A real time cognitive radio testbed for physical and link layer experiments Mishra, S.M.; Cabric, D.; Chang, C.; Willkomm, D.; van Schewick, B.; Wolisz, S.; Brodersen, B.W.; New Frontiers in Dynamic Spectrum Access Networks, 2005. DySPAN 2005. 2005 First IEEE International Symposium on 8-11 Nov. 2005 Page(s):562 - 567
This is the more useful articles I have read of the whole lot. The article is also published by a team from UC Berkeley, which adds some weight to the content presented.
* There are not alot of practical data yet available for CR/SDR
* Some typical performance metric regarding protecting incumbents are:
o Time to detect PU
o Probability of not detecting a PU
o Probability of a wrong detection
o Time required to clear a channel
* Reliability is more important than fast channel detection when first connecting to a medium
* When testing/evaluation a CR system, we need to consider:
o PU and SU placements
o PU signal strength
The article also suggests some detection methods.
When not using a channel:
* Cyclostationary
* Pilot tone detection algorithms
When using a channel:
* Interleaved sensing (No need for dedicated hardware)
* Concurrent sensing using active cancellation of own signal (Require dedicated hardware)
Also interesting, is that their test setup has similar capabilities to the current we have at the lab.
A Cognitive Radio (CR) Testbed System Employing a Wideband Multi-Resolution Spectrum Sensing (MRSS) Technique Hur, Y.; Park, J.; Kim, K.; Lee, J.; Lim, K.; Lee, C.-H.; Kim, H.S.; Laskar, J.; Vehicular Technology Conference, 2006. VTC-2006 Fall. 2006 IEEE 64th 25-28 Sept. 2006 Page(s):1 - 5
This article is exactly the same as the Cross Layer article, more information have been added since this (earlier) article
Low-Complexity Adaptive Transmission for Cognitive Radios in Dynamic Spectrum Access Networks Pursley, M.B.; Royster, T.C.; Selected Areas in Communications, IEEE Journal on Volume 26, Issue 1, Jan. 2008 Page(s):83 - 94
This has been put away for future reading
Spectrum sensing in cognitive radio networks: requirements, challenges and design trade-offs Ghasemi, A.; Sousa, E.S.; Communications Magazine, IEEE Volume 46, Issue 4, April 2008 Page(s):32 - 39
This has been put away for future reading
OS-MAC: An Efficient MAC Protocol for Spectrum-Agile Wireless Networks Hamdaoui, B.; Shin, K.G.; Mobile Computing, IEEE Transactions on Volume 7, Issue 8, Aug. 2008 Page(s):915 - 930
This is not a CR hardware implementation article. But this article suggests a MAC protocol that is supposed to be close to ideal.
Requirements:
* N none overlapping channels
* 1 Control channel
The system is broken into a number of SUGs (Secondary User Groups), each using a channel. a SUG has the following properties:
* Two or more users can form a SUG
* Only 1 SU in a SUG can transmit at a time
* All SUs in a SUG receives the data
* There can be multiple SUGs in a network
* SUs communicate with others of the same SUG on the same channel
The notible advantage of this is that the protocol supports both 1 to 1 communication, as well as teleconferencing style communication.
To control the channels:
* All interchannel control frames are sent on the Control Channel
* Each SUG has a designated Delegated SU (DSU)
* The DSU periodically tune into the Control Channel to report it’s SUG’s traffic conditon
* Within a SUG, transmission follow the standard 802.11 random access scheme with the backoffs
A novel Cognitive Radio Concept deploying Petri Net based Scheduling
Viessmann, A.; Burnic, A.; Spiegel, C.; Bruck, G.H.; Jung, P.;
Wireless Algorithms, Systems and Applications, 2007. WASA 2007. International Conference on
1-3 Aug. 2007 Page(s):233 - 238
This again is beyond my ability at the moment. But there are a few points that can be used.
Detection can be one of the two forms:
1. Persistent, this uses a dedicated hardware which can inform channel changes in real time
2. Non persistent, this does not require dedicated hardwarem, but the tradeoff is system response time
This article also suggests a way in which new PHY and MAC code chain can be loaded on the fly, which means we can adapt to a number of different schemes and systems depending on the current situation.