Final Report Summary - COMIC (Cooperative Multipoint Transmission for Interference Coordination)
COMIC is divided in three main steps, one each year. The objective of the first step is the definition of the optimization problem without any limitation, in order to obtain the best scheduler that serves users with the best QoS.
A deep literature review was performed at the beginning of the first year, which pointed out that the capacity of multicell and multiuser systems is still an open issue. Some works have shown that the optimum coding scheme for the Broadcast Channel (BC) is Dirty Paper Coding (DPC), but the optimum parameters of this scheme cannot be computed yet. Therefore, the upper bound of these systems is not available for the research community. This upper bound is very important to know the point at which we are right now, and how much these systems could be improved. Due to all this, the work of this first year has mainly focused on finding the optimum parameters of DPC. The strategy followed was the finding of sufficient and necessary conditions for the DPC parameters to be optimum. These conditions provide a description of how the optimum solution is, and this can be used to design an algorithm. The main difficulty is that the problem is, in general, non-convex. Thus, conventional techniques cannot be used. At the end of the first year, the fellow reached a sufficient condition, which needed to be extended, in order to be necessary too. This result was sent to IEEE Communications letters, where it received a lot of interest but was rejected because just a sufficient condition was not considered as relevant enough. After the first year, the first step was still unfinished due to the complexity of the problem, that is, the sufficient condition is not enough to design algorithms which could reach the optimum. However, because of the interest that this work raised in the research community, it was decided to extend the first step to the second year too. All this was explained in the reports of the first and second years.
During the second year, and with the objective of finding a necessary and sufficient condition, the fellow started to look at the problem from a different perspective. In this sense, the fellow started the study of Successive Interference Cancellation (SIC) as the optimum transmission scheme for the Multiple Access Channel (MAC). The MAC can be seen as the uplink of a cellular system, whereas the BC is the downlink. SIC has a special structure which makes its parameters to be easier to calculate, as compared with DPC. Moreover, there are several works that highlight certain duality between the BC and the MAC, so that the optimum parameters of DPC could be computed in a dual MAC. For that reason, the main focus of the second year has been the resolution of the MAC problem. The findings have been quite significant. The fellow has acquired a deep understanding of the structure of the optimum parameters of the SIC. Moreover, the fellow has described this structure with several theorems, and, from them, the fellow has been able to write some necessary and sufficient conditions that the optimum parameters satisfy. This is the first time that something like this is done for the BC or the MAC. Moreover, these conditions allow the design of algorithms that converge to the optimum parameters. This is one milestone of the project for the first year, which, as explained before, was delayed due to the complexity of these channels. The results of this work have been published in the IEEE International Symposium on Information Theory (ISIT), and a more complete version is currently under review in the IEEE Transactions on Information Theory.
Although this has been the main line of research during the two years, the fellow also started the second step, which focuses on the design of practical algorithms. In this sense, the fellow has proposed an optimal algorithm to serve users with certain target quality of service in interference free HetNets. The assumption of HetNets without interference is still ideal, but this algorithm is more practical than the capacity-achieving algorithm pursued by the other line of research. This results of this work has been published in the following two papers:
• D. Calabuig, J. F. Monserrat and N. Cardona, "Fairness-driven fast resource allocation for interference-free heterogeneous networks," IEEE Communications Letters, vol. 16, no. 7, pp. 1092-1095, 2012.
• D. Calabuig, J. F. Monserrat and N. Cardona, "Proportionally fair scheduler for heterogeneous wireless systems," Transactions on Emerging Telecommunications Technologies, vol. 23, no. 1, pp 1-5, 2012.
The original plan for the third year described in the project proposal was the implementation of feasible CoMP schedulers and dissemination. These were, as planned, the main focuses of this year, although dissemination took most part of this year effort.
The expertise of the Mobile Communications Group (MCG) of UPVLC in all the layers of the mobile communications systems permitted a contract with the Spanish company Sistelbanda S.A. to aid in the development of a LTE femtocell. Within this contract, Dr. Daniel Calabuig is currently designing the uplink scheduler.
The collaboration between the MCG and Sistelbanda S.A. is expected to continue for several more years in which Sistelbanda S.A. wants to add some CoMP capabilities to their femtocells. The research of Dr. Daniel Calabuig during this project will be essential, and will differentiate the femtocells of this European company from those of the competency.
A detailed description of the dissemination activities is given in section 2 of this report. As a summary, the dissemination activities of this year include the publication of the results of the second year, technology transfer activities within the European project METIS, and some lectures about CoMP.
A deep literature review was performed at the beginning of the first year, which pointed out that the capacity of multicell and multiuser systems is still an open issue. Some works have shown that the optimum coding scheme for the Broadcast Channel (BC) is Dirty Paper Coding (DPC), but the optimum parameters of this scheme cannot be computed yet. Therefore, the upper bound of these systems is not available for the research community. This upper bound is very important to know the point at which we are right now, and how much these systems could be improved. Due to all this, the work of this first year has mainly focused on finding the optimum parameters of DPC. The strategy followed was the finding of sufficient and necessary conditions for the DPC parameters to be optimum. These conditions provide a description of how the optimum solution is, and this can be used to design an algorithm. The main difficulty is that the problem is, in general, non-convex. Thus, conventional techniques cannot be used. At the end of the first year, the fellow reached a sufficient condition, which needed to be extended, in order to be necessary too. This result was sent to IEEE Communications letters, where it received a lot of interest but was rejected because just a sufficient condition was not considered as relevant enough. After the first year, the first step was still unfinished due to the complexity of the problem, that is, the sufficient condition is not enough to design algorithms which could reach the optimum. However, because of the interest that this work raised in the research community, it was decided to extend the first step to the second year too. All this was explained in the reports of the first and second years.
During the second year, and with the objective of finding a necessary and sufficient condition, the fellow started to look at the problem from a different perspective. In this sense, the fellow started the study of Successive Interference Cancellation (SIC) as the optimum transmission scheme for the Multiple Access Channel (MAC). The MAC can be seen as the uplink of a cellular system, whereas the BC is the downlink. SIC has a special structure which makes its parameters to be easier to calculate, as compared with DPC. Moreover, there are several works that highlight certain duality between the BC and the MAC, so that the optimum parameters of DPC could be computed in a dual MAC. For that reason, the main focus of the second year has been the resolution of the MAC problem. The findings have been quite significant. The fellow has acquired a deep understanding of the structure of the optimum parameters of the SIC. Moreover, the fellow has described this structure with several theorems, and, from them, the fellow has been able to write some necessary and sufficient conditions that the optimum parameters satisfy. This is the first time that something like this is done for the BC or the MAC. Moreover, these conditions allow the design of algorithms that converge to the optimum parameters. This is one milestone of the project for the first year, which, as explained before, was delayed due to the complexity of these channels. The results of this work have been published in the IEEE International Symposium on Information Theory (ISIT), and a more complete version is currently under review in the IEEE Transactions on Information Theory.
Although this has been the main line of research during the two years, the fellow also started the second step, which focuses on the design of practical algorithms. In this sense, the fellow has proposed an optimal algorithm to serve users with certain target quality of service in interference free HetNets. The assumption of HetNets without interference is still ideal, but this algorithm is more practical than the capacity-achieving algorithm pursued by the other line of research. This results of this work has been published in the following two papers:
• D. Calabuig, J. F. Monserrat and N. Cardona, "Fairness-driven fast resource allocation for interference-free heterogeneous networks," IEEE Communications Letters, vol. 16, no. 7, pp. 1092-1095, 2012.
• D. Calabuig, J. F. Monserrat and N. Cardona, "Proportionally fair scheduler for heterogeneous wireless systems," Transactions on Emerging Telecommunications Technologies, vol. 23, no. 1, pp 1-5, 2012.
The original plan for the third year described in the project proposal was the implementation of feasible CoMP schedulers and dissemination. These were, as planned, the main focuses of this year, although dissemination took most part of this year effort.
The expertise of the Mobile Communications Group (MCG) of UPVLC in all the layers of the mobile communications systems permitted a contract with the Spanish company Sistelbanda S.A. to aid in the development of a LTE femtocell. Within this contract, Dr. Daniel Calabuig is currently designing the uplink scheduler.
The collaboration between the MCG and Sistelbanda S.A. is expected to continue for several more years in which Sistelbanda S.A. wants to add some CoMP capabilities to their femtocells. The research of Dr. Daniel Calabuig during this project will be essential, and will differentiate the femtocells of this European company from those of the competency.
A detailed description of the dissemination activities is given in section 2 of this report. As a summary, the dissemination activities of this year include the publication of the results of the second year, technology transfer activities within the European project METIS, and some lectures about CoMP.