Final Report Summary - CWNETPLAN (Combined indoor/oudoor Wireless Network Planning)
Project context and objectives
In order to increase the capacity of their network, it is becoming increasingly important for mobile operators to increase the indoor radio coverage. This is why new solutions (like picocells and femtocells) rely on the installation of small base stations directly inside buildings. However, such indoor emitters will often have a negative impact on the outdoor cells, due to the interference they produce. The aim of this project was to propose some solutions to help operators optimise both the indoor and outdoor networks at the same time. This project therefore focused on developing new propagation models that are necessary to evaluate the interference between indoor cells and outdoor cells. These models had to be able to compute the indoor-to-outdoor propagation (in order to evaluate the interference produced by indoor cells on the outdoor cells) and the outdoor-to-indoor propagation (in order to evaluate the interference produced by the outdoor cells on the indoor cells). Moreover this project also focused on optimising indoor cells in the presence of outdoor cells, which is also referred to as self-organisation. During the two years of this project, the following tasks were achieved.
Work performed
Development of new indoor-to-outdoor and outdoor-to-indoor propagation models
We proposed to combine a very accurate propagation model called MR-FDPF with an outdoor propagation model called IRLA. Indeed, MR-FDPF (multi-resolution frequency domain parflow) was shown to be very efficient for indoor propagation, but hard to use in outdoor scenarios due to its high memory requirements. On the other hand, IRLA (intelligent ray launching model) was more efficient for large outdoor scenarios and more open space. First we developed the outdoor-to-indoor model; the IRLA rays from the outdoor source were launched up to the borders of the simulation building. Then in the second stage, we propose a new technique to convert the rays on the borders of the building into MR-FDPF flows. The last step was to use these flows to perform the indoor MR-FDPF coverage. After the outdoor-to-indoor model was validated in different scenarios, we focused on the development of the new indoor-to-outdoor model. This model first computed the indoor MR-FDPF coverage and then a new technique was proposed to convert the flows at the borders of the building into rays to be computed with the IRLA model. This was done using an approach based on the SAGE algorithm. Both models (outdoor-to-indoor and indoor-to-outdoor) have attracted a great deal of attention from other researchers.
Development of combined indoor/outdoor optimisation methods
We focused on self-organising techniques to be implemented in the indoor cells, so that their negative impact (such as interference) on outdoor cells was minimised. First we did some studies on the access methods of femtocells, which were shown to have a big impact on interference. This is why we proposed the concept of hybrid femtocells, where part of the resource is open to all users, whereas the remaining part is restricted to usage by the femtocell owner. Then we fused of LTE optimisation of this purpose we chose to self-organise the power and the resource blocs of the femtocells. Our optimisation techniques were fully distributed, therefore minimising the impact on macrocells and reducing the signalling overhead in the network. Our work on femtocell optimisation led to many publications in journals and books.
Validate the techniques with real measurements
We focused on the calibration of the propagation models. This work was done in collaboration with two PhD students. Measurements were performed in Luton in three different scenarios; and in France (INSA, in conjunction with the IPLAN project) in one scenario. These scenarios included residential areas and university campuses and helped to calibrate the models.
In order to increase the capacity of their network, it is becoming increasingly important for mobile operators to increase the indoor radio coverage. This is why new solutions (like picocells and femtocells) rely on the installation of small base stations directly inside buildings. However, such indoor emitters will often have a negative impact on the outdoor cells, due to the interference they produce. The aim of this project was to propose some solutions to help operators optimise both the indoor and outdoor networks at the same time. This project therefore focused on developing new propagation models that are necessary to evaluate the interference between indoor cells and outdoor cells. These models had to be able to compute the indoor-to-outdoor propagation (in order to evaluate the interference produced by indoor cells on the outdoor cells) and the outdoor-to-indoor propagation (in order to evaluate the interference produced by the outdoor cells on the indoor cells). Moreover this project also focused on optimising indoor cells in the presence of outdoor cells, which is also referred to as self-organisation. During the two years of this project, the following tasks were achieved.
Work performed
Development of new indoor-to-outdoor and outdoor-to-indoor propagation models
We proposed to combine a very accurate propagation model called MR-FDPF with an outdoor propagation model called IRLA. Indeed, MR-FDPF (multi-resolution frequency domain parflow) was shown to be very efficient for indoor propagation, but hard to use in outdoor scenarios due to its high memory requirements. On the other hand, IRLA (intelligent ray launching model) was more efficient for large outdoor scenarios and more open space. First we developed the outdoor-to-indoor model; the IRLA rays from the outdoor source were launched up to the borders of the simulation building. Then in the second stage, we propose a new technique to convert the rays on the borders of the building into MR-FDPF flows. The last step was to use these flows to perform the indoor MR-FDPF coverage. After the outdoor-to-indoor model was validated in different scenarios, we focused on the development of the new indoor-to-outdoor model. This model first computed the indoor MR-FDPF coverage and then a new technique was proposed to convert the flows at the borders of the building into rays to be computed with the IRLA model. This was done using an approach based on the SAGE algorithm. Both models (outdoor-to-indoor and indoor-to-outdoor) have attracted a great deal of attention from other researchers.
Development of combined indoor/outdoor optimisation methods
We focused on self-organising techniques to be implemented in the indoor cells, so that their negative impact (such as interference) on outdoor cells was minimised. First we did some studies on the access methods of femtocells, which were shown to have a big impact on interference. This is why we proposed the concept of hybrid femtocells, where part of the resource is open to all users, whereas the remaining part is restricted to usage by the femtocell owner. Then we fused of LTE optimisation of this purpose we chose to self-organise the power and the resource blocs of the femtocells. Our optimisation techniques were fully distributed, therefore minimising the impact on macrocells and reducing the signalling overhead in the network. Our work on femtocell optimisation led to many publications in journals and books.
Validate the techniques with real measurements
We focused on the calibration of the propagation models. This work was done in collaboration with two PhD students. Measurements were performed in Luton in three different scenarios; and in France (INSA, in conjunction with the IPLAN project) in one scenario. These scenarios included residential areas and university campuses and helped to calibrate the models.