Abstract
A new echo canceller is presented based upon the memory adaptation principle. The echo canceller uses what has been dubbed the 'complement update' algorithm to double the speed of convergence and halve the memory requirements of conventional RAM based echo cancellers.

The effects of finite precision and nonlinearities in the echo canceller have also been examined. The conclusion is made that scaling after the A/D converter obviates A/D digital stopping problems if certain conditions hold. Also the echo canceller can compensate for nonlinearities that are odd symmetric and mild forms of other nonlinearities.

Abstract
Data transmission over long haul submarine optical fibre cables using linear modulation is investigated. The advantages of this method for branching submarine cables are discussed. An experimental multichannel system was constructed and typical results are presented.

Abstract
This paper examines the requirements of a Personal Communications Service (PCS) and compares them with the characteristics of the second generation digital cordless telephone known as CT2. It is concluded that CT2 possesses many of the requirements for a mass market personal communicator and CT2 may be able to be enhanced to fill that role.

Abstract
An analysis of CT2 system performance when CT2 base stations are closely located and operated with and without synchronisation is presented. The issue of the minimum base station separation allowable before synchronisation is required is also examined and some experimental results are presented.

Paper (566 k)

Abstract This paper presents an interference model for cellular and microcellular networks. This model enables a unified treatment of all reception environments from purely noise limited to purely interference limited through a parameter denoted the 'interference to noise ratio' or INR. Using the model, a simplified spatial description of mobile link outage contours is derived. Computer simulations are used to test the analytical theory. It is shown that as a link becomes more interference limited, larger variations in cell sizes result.

Paper (34 k)

Abstract This paper summarises the cell design issues for microcellular networks and explains how conventional large cell network design techniques may not be suitable for designing a ubiquitous microcellular network. A general interference model developed for arbitrary microcellular networks, using a parameter denoted the 'interference to noise ratio' or INR, is used to determine the microcell spacing required to maintain certain cell coverage targets as a function of mutual spill power. It is shown that a contiguous cell coverage requirement imposes significant limits upon the accumulated interference allowable in a microcell system.

Paper (1591 k)

Abstract This paper presents a general interference model for arbitrary cellular and microcellular networks which enables a unified treatment of all mobile link reception environments, from purely noise limited to purely interference limited, as a function of the ratio of interference power to receiver noise power at a receiver. This ratio is denoted 'INR' and given the symbol eta. The model enables a simplified spatial description of radio link outage contours to be derived, and it is shown that the maximum possible range for any mobile terminal is a simple function of the INR. Computer simulations are used to test the analytical theory and show that in comparison to large cell systems, microcellular networks are interference limited and exhibit larger variation in cell sizes. Examples show that this may present difficulties in engineering a microcellular system to meet a target coverage quality requirement.

Paper (361 k)

Abstract This paper describes an interference model that provides a simplified description of cellular and microcellular outage contours. The model and computer simulations indicate that cell size variation increases as microcellular systems become more interference limited. Some current cellular and microcellular systems are modelled.

Paper (104 k)

Abstract An interference model developed for arbitrary microcellular networks, based upon a parameter called the 'Interference to Noise Ratio' or INR, is used to derive interference and cell radius statistics for mobile stations in a microcellular network. The theoretical INR and cell radius statistics for simple interference environments show good agreement with numerical Monte Carlo simulations. For more complex environments, it is hypothesised that the Central Limit Theorem could be applied to approximate the behaviour of an interferer ensemble, enabling the cell radius statistics for a network to be expressed in terms of the system design parameters. Simulations of microcell ensembles that support this hypothesis are presented.

Paper (86 k)

Abstract This paper examines whether microcellular networks can be designed to provide contiguous cell coverage for a given proportion of mobile terminals. Simulation results using a general microcellular interference model show that contiguous cell coverage may not be possible for an acceptable proportion of mobile terminals even if the call loss rate is low.

Paper (126 k)

Abstract The distribution functions of cell radii in a microcell system are derived in closed form for three portable terminal distributions and tested using a Monte Carlo simulation. These distributions provide information about the quality of the microcell coverage. The results suggest that cell size reduction from near-far or same-cell interference from terminals using well spaced channels can be as severe as that resulting from adjacent channel interference from terminals in nearby microcells.

Paper (97 k)

Abstract This paper presents a novel analysis of cochannel reuse ratio distributions in DCA microcell systems, and compares the results with conventional FCA macrocell systems. Computer simulations show that DCA systems exhibit significantly closer channel reuse than FCA systems for a significant proportion of terminals. Mathematical analysis is used to show that this is a fundamental consequence of the microcell architecture and that the resultant interference distributions cannot be obtained using conventional cellular engineering techniques. A closed form expression for the cochannel reuse ratio distribution in DCA microcell systems is derived.

Paper (176 k)

Abstract This paper presents a novel analysis of cochannel and adjacent channel reuse ratio distributions in Dynamic Channel Allocation (DCA) microcell systems, and compares the results with conventional Fixed Channel Allocation (FCA) macrocell systems. Computer simulations show that DCA systems exhibit significantly closer cochannel and adjacent channel frequency reuse than FCA systems for a significant proportion of terminals. Mathematical analysis is used to show that this is a fundamental consequence of the microcell architecture and that the resultant interference distributions cannot be obtained using conventional cellular engineering techniques. A closed form expression for the channel reuse ratio distribution in DCA microcell systems is derived, which in turn establishes a theoretical lower limit to the channel reuse ratio in DCA systems.

Paper (64 k)

Abstract HIPERLAN (Type 1) is an ETSI high speed wireless LAN standard. The simulation results in this letter suggest that the best RSSI-based HIPERLAN channel assignment strategy is to always choose the channel with the lowest amount of measured interference, rather than the first channel measured below the Clear Channel Assessment (CCA) threshold.

Paper (286 k)

Abstract HIPERLAN is a high speed wireless LAN standard recently adopted by ETSI. It operates in 150 MHz of spectrum in the 5 GHz band, with up to five TDD channels provided for duplex data transmission at up to 23.5 Mb/s. This paper shows that in the presence of cochannel cells and full link utilisation, HIPERLAN performs poorly when using its DCA-type Clear Channel Assessment (CCA) scheme, and that the target performance levels are only achievable under optimal FCA-type channel assignments. Two main factors contribute to this poor performance: HIPERLAN has insufficient channels to meet C/I constraints in the presence of cochannel cells, and the CCA scheme results in poor channel distributions.

Paper (417 k)

Abstract This paper presents a new interference model for microcellular networks which integrates radio propagation parameters and user terminal mobility. This model uses a parameter denoted the 'interference to noise ratio' (INR) to obtain a simplified description of mobile link outage contours as a function of the location of the fixed and mobile radio ports. The INR is used to demonstrate that microcell networks are more interference limited than macrocell networks, and thus are more affected by user terminal mobility. Expressions are derived for the INR and user terminal cell radius distributions. It is shown that in microcell systems a significant proportion of terminals may not be able to meet a contiguous coverage criterion, and that closer microcell spacing can reduce rather than improve the coverage quality. Examination of cochannel and adjacent channel reuse ratios in DCA microcell systems suggest that the closer frequency reuse is primarily responsible for these coverage effects. Monte Carlo simulations are used to test the analytical theory. These results may form the basis of a design methodology for microcell systems.

Paper (163 k)

Abstract This paper presents a novel analysis of cochannel reuse in Dynamic Channel Assignment (DCA) microcellular systems with an in-cell channel reuse constraint. Mathematical analysis is used to show that DCA systems exhibit significantly closer cochannel reuse than Fixed Channel Assignment (FCA) systems despite the reuse constraint. Closed form expressions for the cochannel reuse ratio (CRR) distributions in DCA microcell systems are derived. It is shown that the DCA reuse distributions cannot be obtained using conventional cellular engineering techniques, and are a fundamental consequence of the microcell architecture.

Paper (244 k)

Abstract This paper presents the first closed-form analysis of cochannel and adjacent channel reuse ratio distributions in Dynamic Channel Assignment (DCA) cellular systems, and compares the results with conventional Fixed Channel Assignment (FCA) cellular systems. Computer simulations show that DCA systems exhibit significantly closer cochannel and adjacent channel frequency reuse than FCA systems for a significant proportion of terminals. Mathematical analysis is used to show that this is a fundamental consequence of DCA channel pooling and that the resultant interference distributions cannot be obtained using conventional cellular engineering techniques. The closed form expressions derived establish a theoretical lower limit to channel reuse in DCA cellular systems.