The Experts below are selected from a list of 123 Experts worldwide ranked by ideXlab platform
K M Wong - One of the best experts on this subject based on the ideXlab platform.
-
a riemannian distance approach to mimo radar signal design
International Conference on Acoustics Speech and Signal Processing, 2019Co-Authors: Y Y Shi, K M WongAbstract:We consider the signal design problem for a Multi-Input Multi-Output (MIMO) radar. The goal is to design a signal vector having a desired Covariance (Cov) matrix while ensuring that the side-lobes of the ambiguity functions are small. Since Cov matrices are structurally constrained, they form a manifold in the signal space. Hence, we argue that the difference between these matrices should not be measured in terms of the conventional Euclidean distance (ED), rather, the distance should be measured along the surface of the manifold, i.e., in terms of a Riemannian distance (RD). In either case, the signal optimization problem is quartic in the design variables. An efficient algorithm based on successive convex quadratic optimization is developed and is effective in producing good approximate solutions. Comparing the designs using ED and RD, we find that the convergence of the algorithm can be significantly faster by optimizing over the manifold (RD) than by optimizing in ED. More importantly, for tight constraints, the use of RD yields solutions which satisfy the constraints far better than the use of ED.
-
ICASSP - A Riemannian Distance Approach to MIMO Radar Signal Design
ICASSP 2019 - 2019 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP), 2019Co-Authors: Y Y Shi, K M WongAbstract:We consider the signal design problem for a Multi-Input Multi-Output (MIMO) radar. The goal is to design a signal vector having a desired Covariance (Cov) matrix while ensuring that the side-lobes of the ambiguity functions are small. Since Cov matrices are structurally constrained, they form a manifold in the signal space. Hence, we argue that the difference between these matrices should not be measured in terms of the conventional Euclidean distance (ED), rather, the distance should be measured along the surface of the manifold, i.e., in terms of a Riemannian distance (RD). In either case, the signal optimization problem is quartic in the design variables. An efficient algorithm based on successive convex quadratic optimization is developed and is effective in producing good approximate solutions. Comparing the designs using ED and RD, we find that the convergence of the algorithm can be significantly faster by optimizing over the manifold (RD) than by optimizing in ED. More importantly, for tight constraints, the use of RD yields solutions which satisfy the constraints far better than the use of ED.
M. R. Gastonguay - One of the best experts on this subject based on the ideXlab platform.
-
effect of nonmem minimization status and number of replicates on bootstrap parameter distributions for population pharmacokinetic models a case study
Clinical Pharmacology & Therapeutics, 2005Co-Authors: M. R. Gastonguay, A EltahtawyAbstract:Aims Bootstrap (BS) parameter distributions are often used to characterize estimation uncertainty and determine confidence intervals (CI) for population pharmacokinetic (PPK) model parameters. These results are used to guide inferences about clinical relevance of Covariate effects and other model components. The goal of this work was to compare BS parameter distributions using a published PPK model for oxaprozin (OX) under different minimization and re-sampling conditions. Methods Nonparametric BS analyses with NONMEM were conducted on a PPK model for OX and resulting parameter distributions were summarized by: 1) number of BS replicates (REPS), and 2) minimization (MIN) and $Covariance (Cov) status. Results For those runs reporting parameter estimates, BS CI for all parameters 1) did not change by more than 9% after 1000 BS REPS; 2) were unaffected by MIN status (<5% change), and most CI were unaffected by Cov status (<5% change in all but 1 parameter). Conclusions The number BS REPS should be investigated for each problem, but a general estimate of 1000 REPS may be a useful starting point. MIN status did not affect BS CI for this case. Clinical Pharmacology & Therapeutics (2005) 77, P2–P2; doi: 10.1016/j.clpt.2004.11.010
-
Effect of NONMEM minimization status and number of replicates on bootstrap parameter distributions for population pharmacokinetic models: A case study
Clinical Pharmacology & Therapeutics, 2005Co-Authors: M. R. Gastonguay, A. El‐tahtawyAbstract:Aims Bootstrap (BS) parameter distributions are often used to characterize estimation uncertainty and determine confidence intervals (CI) for population pharmacokinetic (PPK) model parameters. These results are used to guide inferences about clinical relevance of Covariate effects and other model components. The goal of this work was to compare BS parameter distributions using a published PPK model for oxaprozin (OX) under different minimization and re-sampling conditions. Methods Nonparametric BS analyses with NONMEM were conducted on a PPK model for OX and resulting parameter distributions were summarized by: 1) number of BS replicates (REPS), and 2) minimization (MIN) and $Covariance (Cov) status. Results For those runs reporting parameter estimates, BS CI for all parameters 1) did not change by more than 9% after 1000 BS REPS; 2) were unaffected by MIN status (
Y Y Shi - One of the best experts on this subject based on the ideXlab platform.
-
a riemannian distance approach to mimo radar signal design
International Conference on Acoustics Speech and Signal Processing, 2019Co-Authors: Y Y Shi, K M WongAbstract:We consider the signal design problem for a Multi-Input Multi-Output (MIMO) radar. The goal is to design a signal vector having a desired Covariance (Cov) matrix while ensuring that the side-lobes of the ambiguity functions are small. Since Cov matrices are structurally constrained, they form a manifold in the signal space. Hence, we argue that the difference between these matrices should not be measured in terms of the conventional Euclidean distance (ED), rather, the distance should be measured along the surface of the manifold, i.e., in terms of a Riemannian distance (RD). In either case, the signal optimization problem is quartic in the design variables. An efficient algorithm based on successive convex quadratic optimization is developed and is effective in producing good approximate solutions. Comparing the designs using ED and RD, we find that the convergence of the algorithm can be significantly faster by optimizing over the manifold (RD) than by optimizing in ED. More importantly, for tight constraints, the use of RD yields solutions which satisfy the constraints far better than the use of ED.
-
ICASSP - A Riemannian Distance Approach to MIMO Radar Signal Design
ICASSP 2019 - 2019 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP), 2019Co-Authors: Y Y Shi, K M WongAbstract:We consider the signal design problem for a Multi-Input Multi-Output (MIMO) radar. The goal is to design a signal vector having a desired Covariance (Cov) matrix while ensuring that the side-lobes of the ambiguity functions are small. Since Cov matrices are structurally constrained, they form a manifold in the signal space. Hence, we argue that the difference between these matrices should not be measured in terms of the conventional Euclidean distance (ED), rather, the distance should be measured along the surface of the manifold, i.e., in terms of a Riemannian distance (RD). In either case, the signal optimization problem is quartic in the design variables. An efficient algorithm based on successive convex quadratic optimization is developed and is effective in producing good approximate solutions. Comparing the designs using ED and RD, we find that the convergence of the algorithm can be significantly faster by optimizing over the manifold (RD) than by optimizing in ED. More importantly, for tight constraints, the use of RD yields solutions which satisfy the constraints far better than the use of ED.
A Eltahtawy - One of the best experts on this subject based on the ideXlab platform.
-
effect of nonmem minimization status and number of replicates on bootstrap parameter distributions for population pharmacokinetic models a case study
Clinical Pharmacology & Therapeutics, 2005Co-Authors: M. R. Gastonguay, A EltahtawyAbstract:Aims Bootstrap (BS) parameter distributions are often used to characterize estimation uncertainty and determine confidence intervals (CI) for population pharmacokinetic (PPK) model parameters. These results are used to guide inferences about clinical relevance of Covariate effects and other model components. The goal of this work was to compare BS parameter distributions using a published PPK model for oxaprozin (OX) under different minimization and re-sampling conditions. Methods Nonparametric BS analyses with NONMEM were conducted on a PPK model for OX and resulting parameter distributions were summarized by: 1) number of BS replicates (REPS), and 2) minimization (MIN) and $Covariance (Cov) status. Results For those runs reporting parameter estimates, BS CI for all parameters 1) did not change by more than 9% after 1000 BS REPS; 2) were unaffected by MIN status (<5% change), and most CI were unaffected by Cov status (<5% change in all but 1 parameter). Conclusions The number BS REPS should be investigated for each problem, but a general estimate of 1000 REPS may be a useful starting point. MIN status did not affect BS CI for this case. Clinical Pharmacology & Therapeutics (2005) 77, P2–P2; doi: 10.1016/j.clpt.2004.11.010
A. El‐tahtawy - One of the best experts on this subject based on the ideXlab platform.
-
Effect of NONMEM minimization status and number of replicates on bootstrap parameter distributions for population pharmacokinetic models: A case study
Clinical Pharmacology & Therapeutics, 2005Co-Authors: M. R. Gastonguay, A. El‐tahtawyAbstract:Aims Bootstrap (BS) parameter distributions are often used to characterize estimation uncertainty and determine confidence intervals (CI) for population pharmacokinetic (PPK) model parameters. These results are used to guide inferences about clinical relevance of Covariate effects and other model components. The goal of this work was to compare BS parameter distributions using a published PPK model for oxaprozin (OX) under different minimization and re-sampling conditions. Methods Nonparametric BS analyses with NONMEM were conducted on a PPK model for OX and resulting parameter distributions were summarized by: 1) number of BS replicates (REPS), and 2) minimization (MIN) and $Covariance (Cov) status. Results For those runs reporting parameter estimates, BS CI for all parameters 1) did not change by more than 9% after 1000 BS REPS; 2) were unaffected by MIN status (
