Flexible Convergence of Wireless Standards and Services
Provision of a broadband communications service to users on board a train is challenging owing to a number of different factors. There is a fundamental trade-off in the capability of a communications standard between coverage and supported data rates, as it is uneconomic and invariably difficult to make high data rate services available anywhere within a national geographic area. The long-range mobility of trains inhibits provision of an adequate service from high capacity localised coverage networks, and wide area networks are unable to support high data rate access at high speeds. It has been shown that simultaneous use of wide area GSM coverage with more localised higher data rate access to the UMTS or WLAN standards makes the provision of a wide range of broadband communications services to trains feasible. Results have highlighted the limitations of a conventional single standard terminal for a generic information download process, common to the majority of applications. The GSM family of standards are unable to effectively support large downloads due to its low data rate, and UMTS or WLAN are inadequate for small downloads as a consequence of its limited coverage capability.
FLOWS has described the concept of multi-standard friendly MIMO techniques, defined as MIMO techniques which can be applied to multiple standards simultaneously while requiring minimal changes to the non-MIMO versions of the standards. The importance of this concept has been outlined in FLOWS report D14. Four MIMO techniques have been identified as relatively multi-standard friendly: MIMO-JT (Joint Transmission); space-time block codes (STBC), spatial multiplexing, and linear beamforming. FLOWS have further identified the related concept of convergence-enabling MIMO techniques: MIMO techniques, which can simplify the implementation of multi-standard systems while allowing backwards-compatibility with legacy terminals. Systems based on the linear transformation concept on which MIMO-JT is based fall into this category, in that this framework can be adapted to any standard by the use of an appropriate linear transformation, and indeed terminals and base station equipment can adapt to the available equipment at the further end of the link. Moreover the use of MIMO may make the system easier to implement, because more degrees of freedom are available. The multi-standard spectral domain signal-processing receiver proposed by FLOWS is also suitable for the implementation of such a framework. A more detailed description is provided in FLOWS report D21.
The previous result shows that intelligent selection of the most appropriate radio access network for successive downloads to a mobile user terminal significantly enhances performance compared with a conventional terminal. We envisage even greater benefits of multi-standard capability if we are able to concurrently download information over multiple standards. The average throughput rate available to a user is increased and concurrent download capability can reduce the delay associated with waiting for the completion of an on-going download. The potential benefits of concurrent download capability were determined in the provision of a generic information download service to mobile commuters; passing through localised regions of UMTS coverage with ubiquitous GPRS coverage. The user terminal keeps track of the remaining download time of queued requests over each standard, and when an information download request is generated, selects the standard it believes will result in the earliest completion of the requested download. Results show that when a high and sustained level of demand is placed on the system from a user, concurrent download capability over the two standards offers significant benefits. Much lower delay values are experienced due to the parallel transmission capability and because of the fundamental increase in overall throughput capability.
Evaluation of performance of TR-STBC transmission over FLOWS propagation scenarios 1 and 2 (train and street)
The performance of TR-STBC schemes have been presented in previous sources for simple two-tap multipath channels. However, an evaluation of the performance of TR-STBC in a realistic terrestrial scenario requires a more accurate modelling of the propagation environment. Channel models were developed in WP2 to correspond to propagation scenarios defined under FLOWS. Here we chose two scenarios, a macro-cellular train scenario and a micro-cellular street scenario and used the IST-TUL wideband double-directional channel model (WDDCM) developed in WP2 to determine the performance of the proposed TR-STBC receiver in the two propagation scenarios. It was found that the receiver can exploit multipath diversity whilst suppressing multi-user interference, in both scenarios (the train scenario offers greater multipath diversity due to greater delay spread). However, the receiver is only able to fully exploit spatial diversity, for realistic transmit (base-station) antenna separations, in the street scenario. This is due to the much larger angular spread provided in this micro-cellular scenario.
An important issue to address is the practicality of the proposed multi-standard selection algorithms. A subset of the algorithms use knowledge of the location of a user relative to the network coverage regions and/or general statistical information about upcoming network coverage to make a more informed decision in the standard selection process. This information needs to be made available to the user terminal without any modifications to the network entities or standards, and so the feasibility with which it can be obtained has been determined. The ability of the algorithms to select the most suitable standard depends on the accuracy of the information available in the selection process. It is therefore important to examine the impact of inaccurate information in this process; to determine the effectiveness of the algorithms with realistic estimates of the required information as would be experienced in a practical implementation. There are several standardised techniques for determining the location of a mobile phone including cell ID, network assisted GPS, and time difference of arrival, and this information can be made available to third party applications via the Open Service Access (OSA) Application Programming Interface (API). Obtaining accurate information about forthcoming network coverage is difficult, as it is hard to predict and dependent on a multitude of factors including the characteristics of the immediate surroundings and mobile objects. Coverage maps can be used to provide broad estimates of upcoming coverage. Alternatively, trial rail journeys can be carried out to determine coverage availability or this information could be gathered from the user terminals as people travel around. One of the algorithms only requires general statistical information regarding network coverage periods, which could be embedded in the user terminal software. Detailed performance evaluation has shown that the proposed algorithms remain effective in the presence of highly inaccurate information regarding the location of the user relative to the network coverage regions.
A potential benefit of multi-standard capability is the inherent diversity, owing to the independence of the communication channels associated with the different standards. Transmission of the same information across two radio channels enables correct reception in the presence of a single channel fade. The occurrence of two simultaneous channel fades is much less frequent than a single channel fade, and so this approach reduces the probability of packet loss at the expense of additional redundancy in the transmission process. This concept has been investigated in WP6, by considering data packet transmission from a single user, across a point-to-point wireless communications link subject to multipath fading. Variable sized data packets are segmented into smaller sub-packets of an appropriate size for transmission across the communications link, with duplicate copies of each sub-packet placed in two separate queues for transmission over two independent standards. The throughput and delay performance of the proposed approach has been evaluated with stop-and-wait and go-back-N automatic repeat request (ARQ) protocols over Rayleigh channels. Simulation results show that despite the 50% reduction in maximum useful capacity, the two-branch diversity system can offer a higher level of useful throughput for large data packet transmissions, when the probability of a transmission encountering a fade becomes significant. This is primarily due to the reduced regularity of packet retransmission owing to the much lower probability of two simultaneous channel fades compared with a single channel fade. The most significant performance enhancements are observed in the worst-case delay performance, with the two-channel system able to meet much tighter delay constraints. The greatest benefit is observed with go-back-N ARQ, as the redundancy and additional delay associated with each retransmission is much greater than with stop-and-wait ARQ.
Space-time block coding (STBC) is a MIMO technique offering increase in robustness to fading by exploitation of spatial diversity in terrestrial wireless channels. STBC is highly suited to small, low power user terminals as it requires only low-complexity processing at the user terminal. However, STBC traditionally assumes frequency-flat fading, and so is not ideally suited for frequency-selective channels, such as the WCDMA links employed in third generation cellular wireless standards (e.g. UMTS). Time-reversal STBC (TR-STBC) allows both spatial and multipath delay diversity gains to be achieved over frequency-selective channels. WCDMA standards such as UMTS use orthogonal spreading codes on the down-link to minimise multi-user interference. However multipath dispersion on frequency-selective channels causes this interference to reappear. Chip-level equalisation can be used to reverse the frequency-selective fading and remove the interference. FLOWS has demonstrated a combination of TR-STBC techniques with chip-level equalisation applied to the WCDMA downlink. This allows spatial and multipath diversity to be exploited, with little deterioration in performance for the case of multiple active users per cell. UMTS-standard common pilot sequences were used for MIMO channel estimation, and the performance of the technique has been realistically evaluated on time-variant channels.
In order to manage the simultaneous use of standards, a convergence manager (CM) is introduced as a functional entity to manage the inter-working between multiple standards and the mapping between different services. Several proposals on the possible locations of CM and the flexibility and optimisation of systems were studied. Every proposed location, such as either in the terminal or at the network side or both including location with the protocol stack has pros and cons, given the different aspects that can be taken into consideration, e.g. architecture, complexity of implementation, end user’s and network operator’s point of view. There are two main issues related to the CM assessed in FLOWS: - The location of the CM - The functionality of the CM Regarding the location of the CM, several options are viable. The CM functionality may be distributed over several elements. As potential hosting entities, the terminal, access network components, core network components, backbone network components and server entities have been investigated, each including significant pros and cons. To make the resulting architecture applicable to real world networks, it is furthermore of importance to consider also the migration path from existing standards towards a CM enabled architecture. An analysis of the convergence manager location was done taking into account the simultaneous use concepts and their mapping into services and standards. It was concluded that: - In order to deal with simultaneity, it is imperative that the MT has a CM. - The location of CMs in the terminal and in the core network is the case that provides the most flexible management of all the simultaneous use concepts. - For the case where, at both ends of the communication, two compatible terminals exist, each with a CM, simultaneous use concepts can easily be used on an end-to-end perspective. In the case where one of the extremes is a server with a CM, the same approach can be considered. For the case where different simultaneous use concepts are being used by each extreme of the communication, a CM is needed in the core-network to allow the concepts adaptation. FLOWS has defined scenarios and performed simulations in order to evaluate the CM options related to the issues above, i.e. its location and its functionality. The evaluation criteria may be split into benefits for users, operators and service providers. This is documented in FLOWS D12. The functionality of the CM is dependent upon the location in the network and the level of converged operation required. It is seen that the mapping of services to standards may be performed based on different policies, each having pros and cons for the parties involved (the customer, the network operator, the service provider, etc). A unique CM perspective/strategy was considered, based on a highest throughput / best system approach. The highest throughput concept means that for data and video applications, the highest available/possible transmission bitrate is selected based on the current networks loads. In respect to the best system approach, priority is given to the most appropriate application-oriented access technology (e.g., data applications first priority WLAN; Voice Calls first priority GSM). Two approaches to CM have been studied: one is a dynamic mapping of services in response to time variations in the channel and traffic. The other is a static mapping based on average channel and traffic. Two scenarios have been concentrated on: isolated cell for HL/2 and UMTS and multiple cells for UMTS and GSM. Several CM scheduling policies were discussed, and the quantitative results of simulation are produced to illustrate the comparison of different scheduling policies and the benefit of CM. The CM is also used to optimise link adaptation. Results from the evaluation based on simulations can be found in FLOWS D19. Additional investigations on the possible location and functionality of the convergence manager have been made with focus on easy integration into existing networks [Stad04a]. Two approaches have been studied: - An integration of the UMTS and WLAN core network components by combining the GGSN (Gateway GPRS Support Node) and PDG (Packet Data Gateway). - The use of session layer mechanisms, based on SIP signalling. Both concepts have been evaluated with respect to: - Signalling delay and overhead. - Flexibility in terms of service categories and mobility scenarios.
Joint Transmission (JT) is a novel receiver-oriented transmission scheme and belongs to the class of Preequalization techniques using channel state information at the transmitter. This MIMO technique allows simplified detection schemes at the mobile terminal (e.g. a simple matched filter). Simplified receiver structures help to overcome the implementation complexity problem, when a multi-standard terminal has to be build, and enables therefore a more unified baseband processing across the different air-interfaces at the terminal. Joint transmission can therefore be understood as convergence-enabling MIMO technique supporting the objectives of FLOWS. During the project lifetime, the JT idea was further developed in detail and derivates were created with certain advantages. The grade of maturity of this technology was increased and the performance was evaluated using the realistic scenarios developed by FLOWS.
A new MIMO receiver architecture for MIMO antenna systems was conceived and evaluated. This uses orthogonal coding to multiplex together the received signals from several antennas so that they can be converted from RF to baseband using a single radio receiver. This technique avoids the necessary replication in transceiver that is nominally required for multiple antenna systems and hence there is a resultant cost saving. In addition the technique does not suffer the loss of receiver performance that other methods such as time multiplexing are subject to. The orthogonal codes can be based on Walsh functions. This technique is described in FLOWS D9. It is also possible that the same technique could be used for combining signal from differing wireless systems so that they could be multiplex for down conversion though a generic RF front-end.
A detailed consideration of the way that two or more wireless standards may be operated simultaneously suggests that the following combinations are possible: Wholly parallel Advantages - Full transmission and reception at any instance from any standard - No need for coordination between standards and their respective radio hardware - No need for the convergence manager to manage the operation of the lower layers (signal from the lower layers to the convergence manager is still required for optimum operation.) Disadvantages - Interference issues between the wireless systems that probably cannot be resolved and may limit radio link performance - Both standards would need full support and functionality - Limited scope for sharable functions in the hardware (full set of system functions are needed for both standards) Parallel with limitations Advantages - Mostly the same as the wholly parallel case (though limitation need to be specifically defined) Disadvantages - True simultaneous transmission from both standards may not be possible - This may also apply to some combination of simultaneous transmission and reception - The convergence manager would need to manage these situations to avoid the collisions (collisions could be treated as lost of corrupt packets.) Interleaved transmit and receive slots Advantages - May solve some of the limitation of the example above - Otherwise, no benefits which are obvious! - This example is described in FLOWS deliverable D8. Disadvantages - Requires the more careful co-ordination and management of the transmit and receive frame slots - Best suited to TDMA/TDD based standards so not possible with standards that use continuous transmissions such as UMTS FDD - There may be an impact on other connections that are already in operation within the network - The overall through-put will be constrained by the need to provide the necessary slot timing flexibility. Operation of wholly uplink and wholly downlink on separate standards Advantages - Solves some issues concerning the managing of transmit and receive slot timing in previous cases. - Fits neatly to some applications Disadvantages - Breaks the correct operation of most if not all existing standards - Probably need a mechanism for providing a return channel via the other standard (for acknowledgements etc.) - No scope for link or route diversity - Interference between simultaneous transmission and reception may still be a problem Rapid connection set-up and closure Advantages - Perform the switching on a rapid basis so that only one complete duplex connection (for one standard) is operational at any instance Disadvantages - There will be limits to how rapidly this could be achieved - Significant throughput delays will be introduced - It may not appear as simultaneous operation to the user - No scope for link or route diversity Use systems that are packet based Advantages - Only transmits when needed - Continuous duplex connection are not needed - Transmissions from each standard can contend for priority between themselves Disadvantages - Limits the choice of standards (GPRS etc) - Packet and transmission collisions still need to be avoided and hence transmit and receive slot timings still need to be coordinated between the standards though this does not need to be performed on a regular TDMA basis. - Has an impact for time bounded services (e.g. voice and video) Conclusion: Generally, constraints may need to be exercised on when transmissions can take place from one or both wireless systems. The network must therefore not expect that a particular standard will be able to transmit data or a packet at any instant. Delays may be introduced into the data flow so as to accommodate appropriate slot timings. Is it expected that these delays may be no longer than one or two frame periods.
A proposal was developed for a common signalling of Layer 1 parameters. The convergence manager needs information on the link capacity hence there is a need signalling from Physical and Link layers. S/N, C/N, BER and delay are not comparable between the standards owing to differing bandwidth, QoS etc. A measurement abstract is needed to provide an indicator of how good a certain bearer and a global Radio Link Service Capacity Indicator is proposed. This new capacity indicator based on service definitions and these provide an indication, how often a service, which is defined by a service definition with its QoS constraints, can transported via this link. Advantage is here, that all QoS aspects are considered inherently and the accuracy is limited to full number of services, which can be mapped. However, difficulties might occur if the service to be mapped is not one of those service definitions used for that indicator. The global Radio Link Service Capacity Indicator (RLSC) for one direction of a radio link specifies the number of radio bearers needed for a typical service, which could be transported simultaneously over this radio link by maintaining the requested quality of services (QoS). This number could be given for typical services named Speech, ISDN, DSL and VIDEO (examples) together with reliability information for this indicator.
The most likely source of interference between closely colocated cellular and 5 GHz WLAN systems is the 3rd harmonics from the former. GSM1800 harmonics can fall into all of the lower 5 GHz band. Harmonics from either UMTS FDD band II or UMTS TDD transmitter can fall into part of the upper 5GHz band. Apart from the obvious solution of using appropriate filters at the output of the transmitter another approach is possible. Knowledge of the operating frequency of the cellular system can be passed to the WLAN so that is can avoid operating where harmonics interference is known to occur.
The effect of inter-element coupling and the directional response of elements on the channel correlation matrix and hence capacity of the MIMO channel. It makes use of the finite scatterers model of the double-directional wireless channel and the Steyskal model of element coupling. Most of the results are valid for an arbitrary array, but some are obtained specifically for the case of a ULA with "patch-like" directional response. It is shown that the effect of coupling can be modelled by modifying the directional response of the elements, using the response of the elements taken as part of the array. It is also shown that coupling affects the array gain, which in turn affects capacity, as well as correlation, which affects capacity in the reverse direction. The effect can be an increase or reduction depending on the phase of the coupling. If the directional responses are all the same the capacity will tend to be reduced compared to omni-directional elements; if they are different this may result in an increase.
A key aim of FLOWS is the investigation of the feasibility of multi-standard terminals; and a further advantage of spectral-domain processing employed in the FLOWS TR-STBC receiver is that it is also clearly relevant to standards based on OFDM. Thus it enables a common signal processing architecture to be applied to both WCDMA and such wireless standards, for example Hiperlan/2, IEEE 802.11, DVB-T. The main feature of this is the use of a fast Fourier transformation (FFT), as required in OFDM transmission, which can be implemented using high-speed dedicated FFT devices. FLOWS have proposed a multi-standard architecture for a WCDMA/Hiperlan-2/DVB-T receiver. The architecture features common ASIC devices for FFT, whose processing power is shared amongst the three standards. Since the three standards also employ convolutional codes, a dedicated Viterbi decoder ASIC can also be shared between the standards. Other standard-specific processing routines could be implemented on DSP or on FPGA using the software-defined radio concept.
Assessment of system performance in realistic terrestrial propagation scenarios requires the consideration of the effects of mobility. Doppler spreading can lead to severe degradation is system performance. A chip-level equalisation receiver based on linear interpolation between tap coefficients was proposed under FLOWS as a simple way of combatting Doppler spread on the WCDMA downlink. Joint multipath-Doppler (2D) matched filtering was proposed in another source as a simple method for combining signal energy dispersed in both multipath (delay) and Doppler domains. Under FLOWS, the technique was applied to the WCDMA downlink, using 2D channel estimation employing UMTS-standards common pilot sequences and 2D MMSE chip equalisation to suppress multi-user interference. Spectral-domain processing was applied to the functional steps to reduce computational complexity and enhance multi-standard friendliness. FLOWS has demonstrated that joint multipath-Doppler equalisation can provide significant performance benefits compared to multipath equalisation alone, for fast fading channels corresponding to, for example, a user terminal on a high-speed train. Further, some multi-user interference suppression can still be achieved over the equivalent 2D RAKE receiver.
Concerning the physical issues that inherently and transversally determine the performance of several standards, especially in their simultaneous application, work has evolved from the generation of a Wideband Double Directional Propagation Channel Model (WDDCM). After having defined its parameters, and going forth with an assessment stage, its simulator implementation has been provided to partners within other WPs. The flexibility that the IST-TUL WDDCM involves allows for its implementation within MIMO frameworks. Additionally, although being a physically-based single-bounce scatterer/cluster model, work has explicitly shown how such model closely relates to the FTW MIMO model, also presented in FLOWS WP2. For the use at the propagation level, several propagation scenarios have been defined. These were anyhow related to issues at other levels, e.g., traffic or simultaneous use of standards. Finally, it has been shown, together with the produced simulator, that including coupling effects within the antenna array elements does not significantly affect the array radiation pattern. Such study has been possible due to the flexibility in the use of the WDDCM simulator. Results were provided considering both coupling and no-coupling among antenna array elements. For this purposes, a real dual-band (UMTS and HIPERLAN/2 frequency bands) microstrip patch antenna linear array, developed within the scope of FLOWS (WP3), was considered. Generally, for the given antenna array, no significant differences in the radio channel characteristics were observed when considering coupling effects.
A set of 10 concepts of simultaneous use was identified, and structured into a space of simultaneous use, where services, systems and operators act as building components. A high level analysis is performed, where the pros and cons of the different concepts are identified. Illustrative examples were also provided for some concepts. Several parameters were listed for the evaluation of simultaneous use concepts (e.g., cost, throughput and complexity), depending on considering a users or an operator’s perspective. Related to this analysis, several proposals on the possible locations for the CM and the flexibility and optimisation of systems were discussed. Every proposed location has pros and cons, considering those different aspects and perspectives. Moreover, using the channel simulator developed in WP2, statistical distributions were obtained for the capacity increases coming from the use of MIMO; various scenarios of standards mixing were considered.
Space-Time Turbo-coded turbo parallel interference cancellation multi-user detection (ST-TuC-PIC-MUD)
Parallel interference cancellation (PIC) is a promising, powerful, and relatively low complexity method for mitigating multi-user interference especially on the up-link of WCDMA systems such as UMTS, where chip equalisation is not applicable. Moreover turbo-codes, which are part of the UMTS standard, allow the use of soft feedback from the decoder in an iterative joint PIC and decoding algorithm, which has been shown to be very effective, giving a BER performance even in a heavily-loaded multi-user CDMA system very close to the single user bound. Space-time turbo-codes, a form of turbo-coded spatial multiplexing, allows the bandwidth efficiency of an uplink connection to be increased by a factor given by the number of transmit antenna elements available. FLOWS has demonstrated a space-time turbo-coded turbo parallel interference cancellation receiver that allows this increase in link capacity while still maintaining a BER performance very close to the single user bound for a SISO system (using a single antenna at both ends of the link). Moreover it has shown that with turbo-codes it is not necessary to use an MMSE filter in the receiver, which must be re-calculated at every iteration.
FLOWS have presented the framework for development of scenarios that links user studies to the definition of services and the technical developments in the FLOWS project. A common set of reference scenarios was defined; User Scenarios and a Services Scenario to create FLOWS User-centred Scenarios aimed at testing the convergence of technologies, and a set of Service Provision scenarios describing conditions of multi-standard deployment in particular locations were considered. A Service Scenario describing a limited set of Data Applications and Network Services was presented, which is characterised in a quantitative way using measurements and models from other sources.
The issue of providing WLAN information via the UMTS radio interface in order to assist a WLAN-capable UE to find and access WLAN Access Points was raised at 3GPP RAN2 #41 answering an incoming liaison statement from SA WG1. Knowing that work was taking place in the IST project FLOWS on the need for the harmonisation of signalling to optimise the system performance when multiple different radio standards are used simultaneously, a submission was made to 3GPP. This offered some possible options for the speeding up of the WLAN selection procedure, by considering what information may be useful to be signalled over the UTRAN, contained in the UMTS broadcast information which could assist WLAN UE s to find I-WLAN access points as long as this information is limited in size. Feedback was requested from SA1 on the usefulness of the following UTRAN signalling enhancements to facilitate handover to WLAN access points on: - The Possible WLAN AP notification message also giving an indication of how often the UE should scan for available WLAN AP. (This can control the trade-off between UE battery life and reduction of the effective coverage area of the WLAN if the UE has already moved through part of the coverage area before performing a scan.) - The UE sending a measurement report to the UTRAN about the WLAN AP in response to the notification of WLAN APs possibly being available in the area. (This measurement could typically be a measure of WLAN signal quality, such as SNR or BLER.) - Additional information being included in the Possible WLAN AP notification message to assist scanning, such as: -- WLAN operating frequency; -- SSID; -- WLAN timing offsets; -- WLAN QoS provision. No feedback was obtained since further work on I-WLAN was deferred until release 7.
A general framework for deriving a channel model which is consistent with one's state of knowledge has been provided based on Information Theoretic tools. This has led to the proposal of a new time variant frequency selective double directional MIMO model. For this model, an asymptotic analysis (in the number of antennas) of the achievable transmission limit was conducted using tools of random matrix theory. A central limit was provided on the asymptotic behaviour of the mutual information and validated in the finite case by simulations and measurements performed at 2.1 GHz (using results of the FLOWS D10 deliverable). The results are both useful in terms of designing a system based on criteria such as quality of service and in optimising transmissions in multi-user networks. The problem of modelling channels is crucial for the efficient design of wireless systems. Unlike the Gaussian channel, the wireless channel suffers from constructive/destructive interference signalling. This yields a randomised channel with certain statistics to be discovered. Recently, the need to increase spectral efficiency has motivated the use of multiple antennas at both the transmitter and the receiver side. Hence, in the case of the i.i.d Gaussian model and perfect channel knowledge at the receiver, it has been proved that the ergodic capacity increase is min(r,t) bits per second per hertz for every 3dB increase ( is the number of receiving antennas and is the number of transmitting antennas) at high Signal to Noise Ratio (SNR). However, for realistic channel models, results are still unknown and may seriously put into doubt the MIMO hype. As a matter of fact, the actual design of efficient codes is tributary of the channel model available: the transmitter has to know in what environment it is transmitting in order to provide the codes with the adequate properties: as a typical example, in Rayleigh fading channels, when coding is performed, the hamming distance (also known as the number of distinct components of the multi-dimensional constellation) plays a central role whereas maximizing the Euclidean distance is the commonly approved design criteria for Gaussian channels. As a consequence, channel modelling is the key in better understanding the limits of transmissions in wireless and noisy environments. In particular, questions of the form: what is the highest transmission rate on a propagation environment where I only know the mean of each path, the variance of each path and the directions of arrival? are crucially important. It will justify the use (or not) of MIMO technologies for a given state of knowledge.
A novel transmission scheme was proposed, that can adapt the available amount of channel information at the transmitter to the current situation. If the channel is static, all information is re-transmitted to the transmitter, and the scheme is then equivalent to the (optimum with bit-loading) SVD scheme. On the other hand, if the channel varies fast, no information is re-transmitted, and then scheme is then (essentially) equivalent to V-BLAST. If the channel is somewhere in between, a variable number of parameters are re-transmitted. A complexity analysis shows that the additional complexity compared to the SVD scheme is cubic in the number of antennas and linear in the number of tones (OFDM). This complexity has to be paid for the increased flexibility in adapting to the channel. Furthermore, It is shown by capacity computations, that increasing the number of re-transmitted parameters also increases performance.
Design, fabrication and test of a range of antenna types and antenna arrays consisting of: - Multiband, wideband and ultra wideband printed antenna prototypes (L-bent patch, PIFA, fractal patch, meandered monopole, IFA, three monopole, branched monopole, half-bowtie, flat-ring monopole, spiral) to be used in small handsets, PDAs and laptop PCs. Multi-standard operation is achieved, supporting GSM 1800, UMTS, IEEE 802.11b, Bluetooth, HiperLAN/2 and IEEE 802.11a; - Multiband MIMO antenna system prototypes integrated into small handsets (PIFAs, IFA, half-bowtie). Multi-standard operation is achieved with small interaction with the user and adequate for a wide range of multipath environments; - Multiband MIMO antenna system prototypes integrated into PDAs (IFAs, half-bowties). Multi-standard operation is achieved, supporting GSM 1800, UMTS, IEEE 802.11b and Bluetooth, and having small interaction with the user; - A MIMO antenna system consisting of 19 L-bent patches integrated into the back of a laptop PC cover. Multi-standard operation is achieved, supporting GSM 1800, UMTS, IEEE 802.11b, Bluetooth, HiperLAN/2 and IEEE 802.11a.