Grant: Discovery Grants Program - Individual
McGill University
$46,000.00
Numerical Methods for Parameter Estimation and Applications
EF 2023-24
Résumé
In many applications such as wireless communications, control, and global navigation satellite systems (GNSS) etc, one needs to recover an unknown integer-valued parameter vector in a linear model, which may be subject to some constraints. One main objective of this proposal is to develop numerically efficient and reliable algorithms and software for this estimation problem and related problems. The optimal approach to estimating the integer valued parameter vector is to solve an integer least squares (ILS) problem, which is typically NP-hard. The dimensions of the estimation problems in applications are becoming larger. For example, in massive multiple input multiple output (MIMO) wireless communications base stations with arrays of many antennas are being used to serve many terminals simultaneously and in GNSS based navigation and positioning various systems are combined so that more measurements are used to give more reliable and accurate positioning. The main challenge is to develop faster algorithms for finding optimal solutions or good suboptimal solutions for various ILS problems for those applications. This is a multidisciplinary area and it is being studied by people from different communities such as communications, cryptograph, information theory, control, GNSS and operations research etc. We have made significant contributions in this area and will continue to work in this area to meet the new challenges. We will study the problems using both conventional approaches and machine learning approaches. Specially we will develop: (1) More effective and efficient lattice reduction algorithms and column permutation algorithms, which will be used for integer parameter estimation. (2) More computationally efficient and accurate suboptimal estimators for various linear model based integer-valued parameters. (3) More efficient and accurate algorithms for sparse integer-valued signal recovery. (4) Algorithms for problems arising in applications such as GNSS. (5) Software packages for integer-valued parameter estimation. In addition to applying machine learning techniques for integer parameter estimation, we will also develop more effective and efficient graph representation learning methods for general purposes by using our expertise in matrix computations and optimization. The results of the proposed research will be useful not only to people in related research areas, but also to people in related industry.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Quebec, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
RGPIN-2023-04125
Name nom
Chang, Xiao-Wen
Department département
School of Computer Science
Organization id
61
Institution établissement
McGill University
Province en
Québec
Province fr
Québec
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2023
Award amount
46000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1507
Committee name en
Computer Science
Committee name fr
Sciences informatiques
Area of application code
1207
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Information, computer and communication technologies
Area of application fr
Technologie de l'information, des ordinateurs et des réseaux de communications
Research subject code
2715
Research subject group en
Information technology
Research subject group fr
Optimization
Research subject en
Technologie de l'information
Research subject fr
Optimisation
Application title
Numerical Methods for Parameter Estimation and Applications
Keyword
Compressed Sensing; Discrete Sparse Signal Recovery; GNSS positioning; Graph representation learning; Integer Parameter Estimation; Lattice Reduction; Least Squares; Massive MIMO detection; Numerical Linear Algebra; Numerical Optimization
Application summary
In many applications such as wireless communications, control, and global navigation satellite systems (GNSS) etc, one needs to recover an unknown integer-valued parameter vector in a linear model, which may be subject to some constraints. One main objective of this proposal is to develop numerically efficient and reliable algorithms and software for this estimation problem and related problems. The optimal approach to estimating the integer valued parameter vector is to solve an integer least squares (ILS) problem, which is typically NP-hard. The dimensions of the estimation problems in applications are becoming larger. For example, in massive multiple input multiple output (MIMO) wireless communications base stations with arrays of many antennas are being used to serve many terminals simultaneously and in GNSS based navigation and positioning various systems are combined so that more measurements are used to give more reliable and accurate positioning. The main challenge is to develop faster algorithms for finding optimal solutions or good suboptimal solutions for various ILS problems for those applications. This is a multidisciplinary area and it is being studied by people from different communities such as communications, cryptograph, information theory, control, GNSS and operations research etc. We have made significant contributions in this area and will continue to work in this area to meet the new challenges. We will study the problems using both conventional approaches and machine learning approaches. Specially we will develop: (1) More effective and efficient lattice reduction algorithms and column permutation algorithms, which will be used for integer parameter estimation. (2) More computationally efficient and accurate suboptimal estimators for various linear model based integer-valued parameters. (3) More efficient and accurate algorithms for sparse integer-valued signal recovery. (4) Algorithms for problems arising in applications such as GNSS. (5) Software packages for integer-valued parameter estimation. In addition to applying machine learning techniques for integer parameter estimation, we will also develop more effective and efficient graph representation learning methods for general purposes by using our expertise in matrix computations and optimization. The results of the proposed research will be useful not only to people in related research areas, but also to people in related industry.