The H∞ proportional-integral-differential(PID) feedback for arbitrary-order delayed multi-agent system is investigated to improve the system performance. The closed-loop multi-input multi-output(MIMO) control framewo...The H∞ proportional-integral-differential(PID) feedback for arbitrary-order delayed multi-agent system is investigated to improve the system performance. The closed-loop multi-input multi-output(MIMO) control framework with the distributed PID controller is firstly described for the multi-agent system in a unified way. Then, by using the matrix theory, the prescribed H∞performance criterion of the multi-agent system is shown to be equivalent to several independent H∞ performance constraints of the single-input single-output(SISO) subsystem with respect to the eigenvalues of the Laplacian matrix. Subsequently, for each subsystem,the set of the PID controllers satisfying the required H∞ performance constraint is analytically characterized based on the extended Hermite-Biehler theorem. Finally, the three-dimensional set of the decentralized H∞ PID control parameters is derived by finding the intersection of the H∞ PID regions for all the decomposed subsystems. The simulation results reveal the effectiveness of the proposed method.展开更多
Pedestrian attribute recognition in surveillance scenarios is still a challenging task due to the inaccurate localization of specific attributes. In this paper, we propose a novel view-attribute localization method ba...Pedestrian attribute recognition in surveillance scenarios is still a challenging task due to the inaccurate localization of specific attributes. In this paper, we propose a novel view-attribute localization method based on attention(VALA), which utilizes view information to guide the recognition process to focus on specific attributes and attention mechanism to localize specific attribute-corresponding areas. Concretely, view information is leveraged by the view prediction branch to generate four view weights that represent the confidences for attributes from different views. View weights are then delivered back to compose specific view-attributes, which will participate and supervise deep feature extraction. In order to explore the spatial location of a view-attribute, regional attention is introduced to aggregate spatial information and encode inter-channel dependencies of the view feature. Subsequently, a fine attentive attribute-specific region is localized, and regional weights for the view-attribute from different spatial locations are gained by the regional attention. The final view-attribute recognition outcome is obtained by combining the view weights with the regional weights. Experiments on three wide datasets(richly annotated pedestrian(RAP), annotated pedestrian v2(RAPv2), and PA-100 K) demonstrate the effectiveness of our approach compared with state-of-the-art methods.展开更多
A local control strategy is presented to switch the consensus value for the first-order multi-agent system.When the local proportional controller is employed in the multiagent systems,the Laplacian matrix of the syste...A local control strategy is presented to switch the consensus value for the first-order multi-agent system.When the local proportional controller is employed in the multiagent systems,the Laplacian matrix of the system is changed.It is proved that the changed Laplacian has the same properties as the Laplacian matrix of the original multi-agent system for the consensus.Based on this,the parameter of the local controller,which can guarantee that all the agents change the original consensus value into the desired one,is determined in terms of the matrix calculation and the stability criterion.In practice,the control system must be implemented in the discrete form.Thus,the influence of the sampling period on the stability of the discrete multi-agent system with the local controller is analysed.The simulation results show the validity of the proposed method.展开更多
Automated machine learning(AutoML)pruning methods aim at searching for a pruning strategy automatically to reduce the computational complexity of deep convolutional neural networks(deep CNNs).However,some previous wor...Automated machine learning(AutoML)pruning methods aim at searching for a pruning strategy automatically to reduce the computational complexity of deep convolutional neural networks(deep CNNs).However,some previous work found that the results of many Auto-ML pruning methods cannot even surpass the results of the uniformly pruning method.In this paper,the ineffectiveness of Auto-ML pruning,which is caused by unfull and unfair training of the supernet,is shown.A deep supernet suffers from unfull training because it contains too many candidates.To overcome the unfull training,a stage-wise pruning(SWP)method is proposed,which splits a deep supernet into several stage-wise supernets to reduce the candidate number and utilize inplace distillation to supervise the stage training.Besides,a wide supernet is hit by unfair training since the sampling probability of each channel is unequal.Therefore,the fullnet and the tinynet are sampled in each training iteration to ensure that each channel can be overtrained.Remarkably,the proxy performance of the subnets trained with SWP is closer to the actual performance than that of most of the previous AutoML pruning work.Furthermore,experiments show that SWP achieves the state-of-the-art in both CIFAR-10 and ImageNet under the mobile setting.展开更多
基金supported by National Natural Science Foundationof China(Nos.61273116 and 61074039)National Natural ScienceFund for Distinguished Young Scholar of China(No.61026016)Natural Science Foundation of Zhejiang Province(No.Y1111012)
文摘The H∞ proportional-integral-differential(PID) feedback for arbitrary-order delayed multi-agent system is investigated to improve the system performance. The closed-loop multi-input multi-output(MIMO) control framework with the distributed PID controller is firstly described for the multi-agent system in a unified way. Then, by using the matrix theory, the prescribed H∞performance criterion of the multi-agent system is shown to be equivalent to several independent H∞ performance constraints of the single-input single-output(SISO) subsystem with respect to the eigenvalues of the Laplacian matrix. Subsequently, for each subsystem,the set of the PID controllers satisfying the required H∞ performance constraint is analytically characterized based on the extended Hermite-Biehler theorem. Finally, the three-dimensional set of the decentralized H∞ PID control parameters is derived by finding the intersection of the H∞ PID regions for all the decomposed subsystems. The simulation results reveal the effectiveness of the proposed method.
基金supported by National Key R&D Program of China(No.2018YFB1308000)Natural Science Foundation of Zhejiang province(No.LY21F 030018)。
文摘Pedestrian attribute recognition in surveillance scenarios is still a challenging task due to the inaccurate localization of specific attributes. In this paper, we propose a novel view-attribute localization method based on attention(VALA), which utilizes view information to guide the recognition process to focus on specific attributes and attention mechanism to localize specific attribute-corresponding areas. Concretely, view information is leveraged by the view prediction branch to generate four view weights that represent the confidences for attributes from different views. View weights are then delivered back to compose specific view-attributes, which will participate and supervise deep feature extraction. In order to explore the spatial location of a view-attribute, regional attention is introduced to aggregate spatial information and encode inter-channel dependencies of the view feature. Subsequently, a fine attentive attribute-specific region is localized, and regional weights for the view-attribute from different spatial locations are gained by the regional attention. The final view-attribute recognition outcome is obtained by combining the view weights with the regional weights. Experiments on three wide datasets(richly annotated pedestrian(RAP), annotated pedestrian v2(RAPv2), and PA-100 K) demonstrate the effectiveness of our approach compared with state-of-the-art methods.
基金supported by the National Natural Science Foundation of China[grant numbers 61273116,61074039]the National Natural Science Fund for Distinguished Young Scholar of China[grant number 61026016].
文摘A local control strategy is presented to switch the consensus value for the first-order multi-agent system.When the local proportional controller is employed in the multiagent systems,the Laplacian matrix of the system is changed.It is proved that the changed Laplacian has the same properties as the Laplacian matrix of the original multi-agent system for the consensus.Based on this,the parameter of the local controller,which can guarantee that all the agents change the original consensus value into the desired one,is determined in terms of the matrix calculation and the stability criterion.In practice,the control system must be implemented in the discrete form.Thus,the influence of the sampling period on the stability of the discrete multi-agent system with the local controller is analysed.The simulation results show the validity of the proposed method.
基金This work was supported by Natural Science Foundation of Zhejiang Province,China(No.LY21F030018)National Key R&D Program of China(No.2018YFB 1308400).
文摘Automated machine learning(AutoML)pruning methods aim at searching for a pruning strategy automatically to reduce the computational complexity of deep convolutional neural networks(deep CNNs).However,some previous work found that the results of many Auto-ML pruning methods cannot even surpass the results of the uniformly pruning method.In this paper,the ineffectiveness of Auto-ML pruning,which is caused by unfull and unfair training of the supernet,is shown.A deep supernet suffers from unfull training because it contains too many candidates.To overcome the unfull training,a stage-wise pruning(SWP)method is proposed,which splits a deep supernet into several stage-wise supernets to reduce the candidate number and utilize inplace distillation to supervise the stage training.Besides,a wide supernet is hit by unfair training since the sampling probability of each channel is unequal.Therefore,the fullnet and the tinynet are sampled in each training iteration to ensure that each channel can be overtrained.Remarkably,the proxy performance of the subnets trained with SWP is closer to the actual performance than that of most of the previous AutoML pruning work.Furthermore,experiments show that SWP achieves the state-of-the-art in both CIFAR-10 and ImageNet under the mobile setting.
