The force-finding process of the cable-net in the deployable mesh reflector antenna,AstroMesh,is investigated to optimize the pretension distribution and satisfy surface accuracy.Since the geometry and topology of the...The force-finding process of the cable-net in the deployable mesh reflector antenna,AstroMesh,is investigated to optimize the pretension distribution and satisfy surface accuracy.Since the geometry and topology of the mesh reflector antennas are given as a constraint with the boundary condition assumed to be fixed,the force-finding process can be performed on a constant equilibrium matrix to obtain a feasible set of forces.Then,the equilibrium matrix can be rewritten in terms of force modes after the singular value decomposition.The object of force design is to minimize the deviation of member forces and,therefore,the surface accuracy can be guaranteed by transforming an optimization of the distribution of prestresses into an optimization with multiple prestress modes.Finally,numerical examples solved by the sequential quadratic programming(SQP)algorithm and the genetic algorithm are given to validate the efficiency of the proposed method.The comparison results show that the genetic method can converge to the optimized point after approximately 50 iterations while the converging process of the sequential quadratic programming method depends largely on the initial points.展开更多
Intensity modulated radiation therapy （IMRT） requires the determination of the appropriate multileaf collimator settings to deliver an intensity map. The purpose of this work was to attempt to regulate the shape bet...Intensity modulated radiation therapy （IMRT） requires the determination of the appropriate multileaf collimator settings to deliver an intensity map. The purpose of this work was to attempt to regulate the shape between adjacent multileaf collimator apertures by a leaf sequencing algorithm. To qualify and validate this algorithm, the integral test for the segment of the multileaf collimator of ARTS was performed with clinical intensity map experiments. By comparisons and analyses of the total number of monitor units and number of segments with benchmark results, the proposed algorithm performed well while the segment shape constraint produced segments with more compact shapes when delivering the planned intensity maps, which may help to reduce the multileaf collimator＇s specific effects.展开更多
To control missile＇s miss distance as well as terminal impact angle, by involving the timeto-go-nth power in the cost function, an extended optimal guidance law against a constant maneuvering target or a stationary t...To control missile＇s miss distance as well as terminal impact angle, by involving the timeto-go-nth power in the cost function, an extended optimal guidance law against a constant maneuvering target or a stationary target is proposed using the linear quadratic optimal control theory.An extended trajectory shaping guidance（ETSG） law is then proposed under the assumption that the missile-target relative velocity is constant and the line of sight angle is small. For a lag-free ETSG system, closed-form solutions for the missile＇s acceleration command are derived by the method of Schwartz inequality and linear simulations are performed to verify the closed-form results. Normalized adjoint systems for miss distance and terminal impact angle error are presented independently for stationary targets and constant maneuvering targets, respectively. Detailed discussions about the terminal misses and impact angle errors induced by terminal impact angle constraint, initial heading error, seeker zero position errors and target maneuvering, are performed.展开更多
Vertebral landmark localization is a crucial step in various spine-related clinical applications,which requires detecting the corner points of 17 vertebrae.However,the neighboring landmarks often disturb each other be...Vertebral landmark localization is a crucial step in various spine-related clinical applications,which requires detecting the corner points of 17 vertebrae.However,the neighboring landmarks often disturb each other because of the homogeneous appearance of vertebrae,making vertebral landmark localization extremely difficult.In this paper,we propose a multi-stage cascaded convolutional neural network(CNN)to split a single task into two sequential steps:center point localization to roughly locate 17 center points of vertebrae,and corner point localization to determine four corner points for each vertebra without any disturbance.The landmarks in each step were located gradually from a set of initialized points by regressing offsets using cascaded CNNs.To resist the mutual attraction of the vertebrae,principal component analysis was employed to preserve the shape constraint in offset regression.We evaluated our method on the AASCE dataset,comprising 609 tight spinal anteroposterior X-ray images,and each image contained 17 vertebrae composed of the thoracic and lumbar spine for spinal shape characterization.The experimental results demonstrated the superior performance of vertebral landmark localization over other state-of-the-art methods,with the relative error decreasing from 3.2e−3 to 7.2e−4.展开更多
基金The National Natural Science Foundation of China(No.51308106,51578133)the Natural Science Foundation of Jiangsu Province(No.BK20130614)+3 种基金the Specialized Research Fund for the Doctoral Program of Higher Education(No.20130092120018)the Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutionsthe Excellent Young Teachers Program of Southeast Universitythe Postgraduate Research&Practice Innovation Program of Jiangsu Province(No.KYCX18_0105)
文摘The force-finding process of the cable-net in the deployable mesh reflector antenna,AstroMesh,is investigated to optimize the pretension distribution and satisfy surface accuracy.Since the geometry and topology of the mesh reflector antennas are given as a constraint with the boundary condition assumed to be fixed,the force-finding process can be performed on a constant equilibrium matrix to obtain a feasible set of forces.Then,the equilibrium matrix can be rewritten in terms of force modes after the singular value decomposition.The object of force design is to minimize the deviation of member forces and,therefore,the surface accuracy can be guaranteed by transforming an optimization of the distribution of prestresses into an optimization with multiple prestress modes.Finally,numerical examples solved by the sequential quadratic programming(SQP)algorithm and the genetic algorithm are given to validate the efficiency of the proposed method.The comparison results show that the genetic method can converge to the optimized point after approximately 50 iterations while the converging process of the sequential quadratic programming method depends largely on the initial points.
基金supported by National Natural Science Foundation of China (No. 10805012)Fundamental Research Funds for the Central Universities (Nos. 2010HGXJ0216, 2012HGXJ0057, 2012HGXJ0062, 2012HGXJ0071, 2012HGBZ0190)Seed Foundation of Hefei University of Technology (2012HGZY0007)
文摘Intensity modulated radiation therapy （IMRT） requires the determination of the appropriate multileaf collimator settings to deliver an intensity map. The purpose of this work was to attempt to regulate the shape between adjacent multileaf collimator apertures by a leaf sequencing algorithm. To qualify and validate this algorithm, the integral test for the segment of the multileaf collimator of ARTS was performed with clinical intensity map experiments. By comparisons and analyses of the total number of monitor units and number of segments with benchmark results, the proposed algorithm performed well while the segment shape constraint produced segments with more compact shapes when delivering the planned intensity maps, which may help to reduce the multileaf collimator＇s specific effects.
基金co-supported by the National Natural Scienc Foundation of China (No. 61172182)
文摘To control missile＇s miss distance as well as terminal impact angle, by involving the timeto-go-nth power in the cost function, an extended optimal guidance law against a constant maneuvering target or a stationary target is proposed using the linear quadratic optimal control theory.An extended trajectory shaping guidance（ETSG） law is then proposed under the assumption that the missile-target relative velocity is constant and the line of sight angle is small. For a lag-free ETSG system, closed-form solutions for the missile＇s acceleration command are derived by the method of Schwartz inequality and linear simulations are performed to verify the closed-form results. Normalized adjoint systems for miss distance and terminal impact angle error are presented independently for stationary targets and constant maneuvering targets, respectively. Detailed discussions about the terminal misses and impact angle errors induced by terminal impact angle constraint, initial heading error, seeker zero position errors and target maneuvering, are performed.
基金supported by the National Natural Science Foundation of China(61872417,62061160490)the project of Wuhan Science and Technology Bureau(2020010601012167)the Open Project of Wuhan National Laboratory for Optoelectronics(2018WNLOKF025).
文摘Vertebral landmark localization is a crucial step in various spine-related clinical applications,which requires detecting the corner points of 17 vertebrae.However,the neighboring landmarks often disturb each other because of the homogeneous appearance of vertebrae,making vertebral landmark localization extremely difficult.In this paper,we propose a multi-stage cascaded convolutional neural network(CNN)to split a single task into two sequential steps:center point localization to roughly locate 17 center points of vertebrae,and corner point localization to determine four corner points for each vertebra without any disturbance.The landmarks in each step were located gradually from a set of initialized points by regressing offsets using cascaded CNNs.To resist the mutual attraction of the vertebrae,principal component analysis was employed to preserve the shape constraint in offset regression.We evaluated our method on the AASCE dataset,comprising 609 tight spinal anteroposterior X-ray images,and each image contained 17 vertebrae composed of the thoracic and lumbar spine for spinal shape characterization.The experimental results demonstrated the superior performance of vertebral landmark localization over other state-of-the-art methods,with the relative error decreasing from 3.2e−3 to 7.2e−4.
