Weimin Xiao
Weimin Xiao has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
Anthony Soong
Huawei U.S. Wireless R&D
1 shared publication- Bing Liu
Department of Computer Science, University of Illinois at Chicago, Chicago, IL, USA
1 shared publication George Calcev
Huawei U.S. Wireless R&D
1 shared publicationGrothusen John
Department of Anesthesiology and Critical Care, Perelman School of Medicine at the University of Pennsylvania, PA, USA
1 shared publication- Hao Ge
School of Chemistry and Chemical Engineering, State Key Laboratory of High-Efficiency Utilization of Coal and Green Chemical Engineering, Ningxia University, Yinchuan 750021, China
1 shared publication - Jiajun Song
Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
1 shared publication
Latest Publications
- 14 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 20 (4) , 5692-5701
Multi-state Ingredient Recognition via Adaptive Multi-centric Network
Ingredient recognition has received significant attention due to its numerous industrial applications, such as intelligent retail terminals and intelligent cooking devices. However, ingredient recognition has the following challenges: 1) dynamic changes in the number of categories; 2) greater diversity and regionality of ingredients; and 3) large visual differences among different states of ingredients. In this article, we propose an adaptive multi-centric network (AdMNet) to solve the problem of ingredient recognition. AdMNet is based on the idea of retrieval, which consists of two main parts, the adaptive multi-centric nearest-neighbor central mean (AdM-NCM) classifier, and the context-aware attentional pooling (CAP) module. The AdM-NCM classifier adaptively establishes category-centric vector groups to recognize ingredients via optimizing the minimum clustering variance, where each state of the ingredient has its corresponding centric vector. The CAP module combines contextual information and multiple attention mechanisms. It captures more focused and discriminative features with higher weights assigned to fine-grained features, which results in better feature representation. In addition, we collect a large-scale ingredient dataset, ISIA Ingredient-201 with 201 classes and 100 442 images. To prove the greater robustness and generalization of our method, we compare the metrics in basic scenarios and realistic scenarios with those of other methods. Specifically, the base scenario is the regular setup, and the real scenario is similar to the class incremental learning setup. The experimental results show that our method reaches the state of the art on both basic scenarios and realistic scenarios with small samples.
- 14 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 20 (4) , 5692-5701
Multi-state Ingredient Recognition via Adaptive Multi-centric Network
Ingredient recognition has received significant attention due to its numerous industrial applications, such as intelligent retail terminals and intelligent cooking devices. However, ingredient recognition has the following challenges: 1) dynamic changes in the number of categories; 2) greater diversity and regionality of ingredients; and 3) large visual differences among different states of ingredients. In this article, we propose an adaptive multi-centric network (AdMNet) to solve the problem of ingredient recognition. AdMNet is based on the idea of retrieval, which consists of two main parts, the adaptive multi-centric nearest-neighbor central mean (AdM-NCM) classifier, and the context-aware attentional pooling (CAP) module. The AdM-NCM classifier adaptively establishes category-centric vector groups to recognize ingredients via optimizing the minimum clustering variance, where each state of the ingredient has its corresponding centric vector. The CAP module combines contextual information and multiple attention mechanisms. It captures more focused and discriminative features with higher weights assigned to fine-grained features, which results in better feature representation. In addition, we collect a large-scale ingredient dataset, ISIA Ingredient-201 with 201 classes and 100 442 images. To prove the greater robustness and generalization of our method, we compare the metrics in basic scenarios and realistic scenarios with those of other methods. Specifically, the base scenario is the regular setup, and the real scenario is similar to the class incremental learning setup. The experimental results show that our method reaches the state of the art on both basic scenarios and realistic scenarios with small samples.
- 26 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 343-346
A dynamic model for the electromechanical coupling of the ball screw feed system
The efficiency of computer numerical control (CNC) machine tools is directly associated with the machine’s ball screw feed system (BSFS). The electromechanical coupling effect of the mechanical structure and the servo system affects the system’s performance. The present study reported a mathematical model for explaining the electromechanical coupling effect by mathematically associating the servo system and the mechanical structure using MATLAB/Simulink. By adding a sinusoidal signal as an input signal in MATLAB/Simulink, the acceleration signals of BSFS were obtained by adjusting the frequency of the input signal. The electromechanical coupling model is verified by simulation and experiments. The simulation results showed that the frequency and amplitude are directly related. No significant difference was noticed between the generated “peak” frequency and the input signal frequency. It is shown that the errors between experimental results and theoretical ones are 16.8% and 2%, respectively, which is acceptable and verifies this established electromechanical coupling model.
- 26 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 343-346
A dynamic model for the electromechanical coupling of the ball screw feed system
The efficiency of computer numerical control (CNC) machine tools is directly associated with the machine’s ball screw feed system (BSFS). The electromechanical coupling effect of the mechanical structure and the servo system affects the system’s performance. The present study reported a mathematical model for explaining the electromechanical coupling effect by mathematically associating the servo system and the mechanical structure using MATLAB/Simulink. By adding a sinusoidal signal as an input signal in MATLAB/Simulink, the acceleration signals of BSFS were obtained by adjusting the frequency of the input signal. The electromechanical coupling model is verified by simulation and experiments. The simulation results showed that the frequency and amplitude are directly related. No significant difference was noticed between the generated “peak” frequency and the input signal frequency. It is shown that the errors between experimental results and theoretical ones are 16.8% and 2%, respectively, which is acceptable and verifies this established electromechanical coupling model.
- Open Access
- 30 September 2022
- Published by Transpopmed in Translational Perioperative and Pain Medicine
- Vol. 8 (3)
Impacts of COVID-19 Epidemic on Labor and Delivery: Data from Anesthesia Quality Control Registry of Hubei, China
- 30 September 2022
- Published by Transpopmed in Translational Perioperative and Pain Medicine
- Vol. 8 (3)
Open AccessImpacts of COVID-19 Epidemic on Labor and Delivery: Data from Anesthesia Quality Control Registry of Hubei, China
- 1 September 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Performance Analysis of Vehicle Platooning Using a Cellular Network
This paper investigates the performance of longitudinal control of a platoon of vehicles using communication of messages through a cellular network. We describe a platooning algorithm that utilizes, at each following vehicle, information communicated by the preceding vehicle and the lead vehicle of the platoon. The information received from the lead vehicle includes the position, speed and acceleration of the lead vehicle, whereas the information received from the preceding vehicle includes the target acceleration the preceding vehicle is attempting to achieve. The information is transmitted in messages that are received by a cellular base station, which then retransmits the messages to the intended vehicles. We perform an analysis of communication system capacity for a highway setting with dense configurations of platoons. We show that (a) the additional latency due to transmission through the infrastructure does not impact platooning performance, and (b) LTE can even support extreme vehicle densities with no impact to platooning.
- 1 September 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Performance Analysis of Vehicle Platooning Using a Cellular Network
This paper investigates the performance of longitudinal control of a platoon of vehicles using communication of messages through a cellular network. We describe a platooning algorithm that utilizes, at each following vehicle, information communicated by the preceding vehicle and the lead vehicle of the platoon. The information received from the lead vehicle includes the position, speed and acceleration of the lead vehicle, whereas the information received from the preceding vehicle includes the target acceleration the preceding vehicle is attempting to achieve. The information is transmitted in messages that are received by a cellular base station, which then retransmits the messages to the intended vehicles. We perform an analysis of communication system capacity for a highway setting with dense configurations of platoons. We show that (a) the additional latency due to transmission through the infrastructure does not impact platooning performance, and (b) LTE can even support extreme vehicle densities with no impact to platooning.
- 20 March 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Load-Based Resource Allocation and Interference Coordination for Multi-Carrier Dense Networks
The multi-carrier ultra-dense network (MC-UDN) has been considered as a promising candidate for future 4.5G and 5G wireless networks to support the explosive growth of data traffic. In a MC-UDN, it is crucial to efficiently allocate network resources including the network nodes and carriers according to traffic loads and channel and interference conditions. In this paper, carrier- level resource allocation jointly in the time domain and frequency domain is proposed. We formulate optimization problems based on queuing theory for two traffic models to optimize the average user experience in the network. Simulation and analytic results show that our proposed scheme can greatly enhance user experience and provide a much more efficient allocation strategy than existing ones, such as resource allocation in the frequency domain only.
- 20 March 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Load-Based Resource Allocation and Interference Coordination for Multi-Carrier Dense Networks
The multi-carrier ultra-dense network (MC-UDN) has been considered as a promising candidate for future 4.5G and 5G wireless networks to support the explosive growth of data traffic. In a MC-UDN, it is crucial to efficiently allocate network resources including the network nodes and carriers according to traffic loads and channel and interference conditions. In this paper, carrier- level resource allocation jointly in the time domain and frequency domain is proposed. We formulate optimization problems based on queuing theory for two traffic models to optimize the average user experience in the network. Simulation and analytic results show that our proposed scheme can greatly enhance user experience and provide a much more efficient allocation strategy than existing ones, such as resource allocation in the frequency domain only.
- Open Access
- 31 October 2005
- Published by Association for Computing Machinery (ACM)
- p. 60-67
Opportunity map
- 31 October 2005
- Published by Association for Computing Machinery (ACM)
- p. 60-67
Open AccessOpportunity map
- 30 March 2004
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 5, 3251-3255
Optimal detection of M-QAM signal with channel estimation error
In this paper, we study the optimal detection of QAM signaling with channel estimation error. Optimal detectors, Max- imum Likelihood (MaxLike) detector and a Modified Minimum Distance (ModDist) detector, are derived for the case of a known Gaussian channel distribution and unknown channel distribution, respectively. These detectors differ from the traditional minimum distance detector which ignores the effect of channel estimation error. Soft decision metrics for coded performance are also gener- ated using the soft output of these detectors based on Generalized Likelihood Ratio (GLR). Numerical results are given to compare the performance of these detectors. Although only marginal im- provement is observed when using the proposed detectors with- out coding, significant gains are obtained when used together with coding.
- 30 March 2004
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 5, 3251-3255
Optimal detection of M-QAM signal with channel estimation error
In this paper, we study the optimal detection of QAM signaling with channel estimation error. Optimal detectors, Max- imum Likelihood (MaxLike) detector and a Modified Minimum Distance (ModDist) detector, are derived for the case of a known Gaussian channel distribution and unknown channel distribution, respectively. These detectors differ from the traditional minimum distance detector which ignores the effect of channel estimation error. Soft decision metrics for coded performance are also gener- ated using the soft output of these detectors based on Generalized Likelihood Ratio (GLR). Numerical results are given to compare the performance of these detectors. Although only marginal im- provement is observed when using the proposed detectors with- out coding, significant gains are obtained when used together with coding.