 |
Project objectivesHow to enlarge the bandwidth and make the whole system benefit from it?
Operation on wider channels How to enhance system spectral efficiency and provide ubiquitous high-rate coverage?
Advanced cooperative transmission/reception techniques How to enlarge the bandwidth and make the whole system benefit from it?It is clear enough that reaching Gbps peak throughput requires not only higher spectral efficiency but also much larger bandwidth. The total spectrum needed by pre-IMT-2000, IMT-2000 and IMT-advanced systems by year 2020 is estimated to range between 1.2GHz and 1.7 GHz in [ITU-R M2078]. Based on this data, ROCKET will focus on techniques to exploit a large spectrum not only for hot-spot users but for the benefit of every user in the system. Operation on wider channelsThe maximum channel bandwidth is 20MHz in IEEE802.16e [802.16e], although WiMAX profiles are currently defined only up to 10MHz. For such bandwidths and for the usual link budgets, an increase in the available bandwidth translates into a proportional increase in offered bit rate. It is also well known that system capacity can be further enhanced with the use of MIMO techniques in highly dispersive environments, but these are not efficient in dealing with pathloss and shadowing. Preliminary results in FIREWORKS show how cooperative relaying can exploit wideband short-range links to create Virtual Antenna Arrays, thus also increasing the throughput in remote areas. ROCKET will investigate how the whole system may benefit from wideband links. Opportunistic spectrum usageAllocating one wide block of contiguous spectrum for primary use by one given system may not always be feasible. In this case, it may be necessary to explore and exploit the locally available spectrum, after ensuring that primary systems operation is not disturbed and possibly applying coexistence and radio resource sharing mechanisms. This topic has already been investigated by several 7th Framework Programme - ICT PROJECTs [ORACLE, E2RII] and standards [802.16h, 802.22]. ROCKET will elaborate on existing technologies for the opportunistic usage of the spectrum and apply them to our targeted system. In this regard, liaison activities with the aforementioned projects and with E3 will be favoured. Scattered multi-band operationSome devices may be able to operate simultaneously in multiple-bands. As an example, a fixed infrastructure relay may be equipped with an air interface at typical cellular frequencies (1 to 3GHz) for communication with the Subscriber/Mobile Stations, and a higher frequency air interface (e.g. 5GHz or higher) for communication with the base station or other relays. ROCKET will investigate how this multi-band capability impacts the system behaviour and performance. How to enhance system spectral efficiency and provide ubiquitous high-rate coverage?In addition to increased bandwidth, offering 1Gbps throughput will require a high peak spectral efficiency. ROCKET will thus consider the use of spatial multiplexing techniques, multi-user cooperative schemes, advanced multi-user coding and high-order modulations, which have an impact on RF front-end and/or baseband compensation scheme design. Minimizing the MAC overhead regardless of the bandwidth is seen as another source of the spectral efficiency the user will eventually experience. Particular attention will be drawn on increasing the spectral efficiency in remote and shadowed locations, which typically represent most of the covered area. As the relay stations and base stations will experience different propagating and interference situations, the constraints for opportunistic use of the spectrum may also be different, thus leading to scenarios of potentially enhanced reuse. Advanced cooperative transmission/reception techniquesThe deployment of conventional and cooperative relays for coverage extension and spectral efficiency increase is addressed in FIREWORKS and preliminary results of the project show the benefit of using simple relay terminals especially at cell edge and in shadowed areas. ROCKET will investigate advanced topics in cooperation at the PHY and MAC level which can all further improve system spectral efficiency. The expected result is a more homogeneous experienced quality in the cell for a given deployment of base stations. Improved single-user and multiuser cooperative techniquesAlthough there is already a vast literature on cooperative schemes, some aspects of single-user cooperative relaying are still not well understood and require further investigation, for instance closed-loop techniques exploiting different levels of Channel State Information at the source and relay terminals. Also new protocols that use multiple relays or joint scheduling of UL and DL transmissions promise superior spectral efficiency. Base Stations and Relay Stations cooperative transmissionIn ROCKET, we will investigate how cooperation can be extended to multiple base stations with relays potentially attached to different base stations. Preliminary results show that very high uplink spectral efficiency gains can be achieved at cell edge by cooperatively decoding the signal received on 3 (or more) base stations connected by high-speed fibre backhaul. Joint multiple cooperative linksCooperative links to different destinations typically interfere with each other when performed on the same time-frequency resource. However, interference cancellation techniques can be implemented in the transmitter and in the receiver. For instance a base station aware of the data transmitted by another base station (through a backbone connection) may cooperate to help it. But it may also reuse the same resource and transmit its own independent data to another user by Dirty Paper Coding pre-cancellation or using Nested Lattice Codes. Interference cancellation may also be performed at the destination by exploiting the degrees of freedom provided by multiple receive antennas. Base Stations and Relay Stations coordinationDynamic radio resource reuseSince for higher spectral efficiency the system will likely operate under low frequency reuse factor, it is important to consider dynamic reuse schemes. Some solutions aiming at avoiding catastrophic inter-cell interference already exist (e.g. fractional frequency reuse), but at the expense of spectrum efficiency loss. These solutions have to be re-visited in ROCKET to take into account the existence of relays and the cooperation capabilities. Inter-base station coordination will thus be required, for instance to avoid that two relays belonging to two different cells interfere each other. Advanced inter-cell radio resource allocation and scheduling are envisioned, with different levels of centralised/decentralised control. Cluster-based link adaptationTo improve the average cluster spectral efficiency in a cellular-based communication system, further research should focus on cooperation between base stations. In ROCKET, we will look into cluster-based link adaptation algorithms aiming at combining cluster-based power allocation with both cell-based modulation and coding algorithms and cooperation protocol. Optimal route selection in multi-hop and cooperative communicationIn general, the problem of routing in a meshed network with a limited number of hops to reach the wired backhaul is greatly reduced compared to pure ad-hoc routing. However, ROCKET will also consider multiple parallel relays, more than 2-hop paths, relay-to-relay transmission and cooperation. Therefore, routing cannot be considered separately from association/disassociation of the relays, the radio resource allocation and the improvements of the physical layer. Appropriate cross-layer routing metrics must be derived. Ultra-Efficient MAC layerOverhead reductionIn order to optimize the system spectral efficiency, the MAC must be scalable and must remain efficient even at very high PHY bit rates. A good trade-off must be found between exploiting the degrees of freedom offered by the PHY layer versus controlling the MAC overhead. For instance, allowing time-frequency selective scheduling with small chunk size improves PHY throughput but increases the MAC overhead. A redefinition and evaluation of MAC procedures for the schemes proposed in ROCKET is therefore required. Reconfigurable MACSince the system will be able to operate at different central frequencies, in different bandwidths, some MAC parameters will have to be adapted dynamically. For instance, operation at higher frequency means higher Doppler and the need to reduce channel sounding intervals. Adaptive schemes and resource allocationIn FIREWORKS, several adaptive resource allocation and scheduling algorithms are investigated, such as integrated bit, power and sub-carrier allocation, or user grouping for SDMA scheduling. As additional degrees of freedom will be introduced in ROCKET, especially with multi-user cooperative transmissions, it will be required to design low-complexity algorithms in order to allocate the radio resource and schedule mixed QoS users and allocate the resource according to criteria such that the benefits of the PHY layer are realized. |