Cooperative principles in Wimax

Cooperative diversity is an important type of cooperation that is useful in wireless networks. Although the purpose of employing relays is different i.e. to increase cell coverage and also help mobile stations (MS) to save power, there is an additional adavantage of being able to exploit spatial diversity from the multiple relay paths created.
Aside: Spatial diversity can be simplistically explained as the advantage that a receiver gets when he "sees" multiple copies of the incoming signal. Imagine the advantage that we have when we see with two eyes as compared to a scenario if humans had only a single eye (God forbid!).

The idea of wireless relay first originated as an information theoretical scenario in Cover and Gamal (1979). Some of the important research in this area is listed below: (Source: Wimax Evolution: Emerging technologies and applications: Wiley 2009)
1. The relay channel (Cover and Gamal (1979))
2. User cooperative diversity (Sendonaris et al.(1998))
3. Cooperative coding (Hunter and Nosratinia(2002))
4. parallel relay channel (Gastpar et al.(2002))
5. Multihop diversity (Boyer et al.(2004), Gupta and Kumar (2003))

A few of the practical scheduling issues in a wireless network are:
1. Which entities make the cooperation decision (Always a BS in the case of Wimax)
2. Whether to relay or not
3. How to relay the data (different cooperative diversity scheme)

Relay selection method is not specified in the current 802.16j draft. Ideally a decision should be made based on the CSI and the number of hops (delay constraint at the receiver). I would like to learn more about relay selection algorithms. It is quite obvious that as the network scales in size, this algorithm will become more and more complicated. The question is: Is there a sub optimal way which is also practically feasible for relay selection? A million dollar question....comments and suggestions are welcome.. :)

Literature survey

I have started reading about Wimax standard IEEE 802.16j. Since 802.16j is an ammendment to the original draft 802.16-2004 standard , I also need to know about this original draft.

Some points from 802.16-2004 draft:
For cellular and multiple access channels, the uplink and downlink transmit/receive specifications are different, which is in contrast to point to point standards such as 802.11.

The PHY has four different versions:
1. WirelessMAN-SC PHY: Single carrier , targetted for 10 to 66 Ghz range
2. WirelessMAN-SCa PHY: Single carrier, Non Line of Sight operation possible, below 11 Ghz
3. WirelessMAN-OFDM PHY: Based on OFDM, Non Line of Sight operation possible, below 11Ghz
4. WirelessMAN-OFDMA PHY

I am not sure as of now if a device can support multiple PHYs or if base station (BS) can serve subscriber stations (SS) with different types of PHYs. .


However Mobile multihop relays(MMR) are supported only in the OFDMA type of PHY.

Some points from 802.16j draft:
Relay routing and channel allocation is performed at the MAC layer. Section 6.3.28 in 802.16j draft deals with this issue. Relays can be of following 4 types:
1. Transparent
2. Non transparent centralized scheduling and security
3. Non transparent decentralized scheduling and centralized security
4. Non transperant decentralized scheduling and security.

A transparent relay is one where the SS is oblivious to the presence of intermediate relay
A non transparent relay is the one where the SS knows about the presence of intermediate relay

Scheduling refers to channel allocation where as decentralized security refers to the fact that a relay is non regenerative decode and forward type.
Routing decisions are taken centrally by the base station. Algorithm for optimal path selection is beyond the scope of the standard. (There is a room for applying our creativity here).

Here is a cool paper that I found. It was preetty useful:
http://www.fujitsu.com/downloads/MAG/vol44-3/paper10.pdf