Grid-Based Analysis Of GSM Field Intensity Coverage And Capacity

Field intensity coverage and capacity are important standards of GSM network quality. It can improve the performance of GSM network efficiently by optimizing field intensity coverage and capacity. Field intensity and coverage can reflect the wireless signal distribution, so they are important references for network optimizer. Theoretical Erl-B formula is not suitable for GSM network, so an modified Erl-B table generated from minute-level traffic data is needed.In this paper, parallel computing methods for GSM network field intensity coverage and Erl-B modification are studied. The area covered by a cell or sector is divided into regions, each of which is of 25×25 square meters. The field intensity of each region is represented by the intensity of the center point of this region. The field intensity analysis of a cell is thus reduced to the computation of the intensity of the center points, so that the field intensity coverage analysis for the whole network is computation-intensive. Parallel field coverage analysis averages the area covered by a cell or sector into n fan-shaped sub-sections (each sub-section is called a leaf-section). The field intensity coverage analysis of one cell is thus reduced to the parallel computation of field intensity coverage of n leaves.The modified Erl-B table is generated from the minute-level traffic data of the whole GSM network, so it is also computation-intensive. The traffic data is classified by carrier frequency. Each class of traffic data is used to generate channel-congestion rate-traffic shaped modified Erl-B formula with congestion-based or confidence-based method.Based on the above principle, the Grid-based Field Intensity Coverage and Capacity Analysis System (GFICAS) is presented to tackle the computing complexity of field intensity coverage and capacity analysis. Sub-tasks to analyze field intensity coverage in leaf-sections or to generate modified Erl-B formula of carrier frequencies are encapsulated into grid tasks, all of which are then allocated to free computing resources in the grid. Parallel executing of these tasks in the grid not only speeds up the tasks but also makes the analysis and modification more accurate.