Mengdi Xu

Mengdi Xu

Auto-generated profile, based on publicly available information

Mengdi Xu

Department of Mechanical Engineering Carnegie Mellon University

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
16 October 2017 in IEEE Wireless Communications Letters
Verified
Journal Article
Closed access
MXMengdi Xu
Mengdi Xu
MJMing Jin
Ming Jin
QGQinghua Guo
Qinghua Guo
YLYouming Li
Youming Li
A multichannel selection strategy is proposed for RF energy harvesting cognitive radio networks, where multiple channels are available, and each secondary user (SU) harvests RF energy from the channel being occupied by active primary users and transmits data using other idle channels. To maximize the average throughput of SUs which contend for the use of idle channels, the multichannel selection strategy is formulated to a convex optimization problem. Further, we derive an analytical solution to the optimization problem. Numerical results are provided to demonstrate the superior performance of the proposed strategy and verify the analytical solution.

17 July 2019 in IEEE Access
Verified
Journal Article
Open access
MXMengdi Xu
Mengdi Xu
SLShousheng Liu
Shousheng Liu
Zeshui XuZeshui Xu
Zeshui Xu
WZWei Zhou
Wei Zhou
Data envelopment analysis (DEA) is a model for evaluating the effectiveness of relative effectiveness of decision-making units with multiple input and output data based on non-parametric modeling using mathematical programming (including linear programming, multi-parameter programming, stochastic programming, etc.). Due to the complexity of real life, sometimes it is hard to get the exact value of the input and output data directly. The fuzzy DEA (FDEA) proposed to solve this problem well and is widely used in practice. However, the data of FDEA have certain subjectivity, and in addition, some indicators cannot be quantified intuitively. Owing to the complexity of society, there are often some causal relationships in the indicator system. As a method of combining uncertainty and graph theory for uncertainty reasoning, Bayesian network (BN) can effectively deal with the causal chain problem existing in the index and discover the potential relationship between data. The BN is often used to process accurate numerical information and does not handle uncertain information with ambiguity favorably. In order to solve the above issue, the interval-valued intuitionistic fuzzy number (IVIFN) is introduced into the BN to construct the interval-valued intuitionistic fuzzy BN (IVIFBN). Then, based on the index data obtained by the IVIFBN, the crossover efficiency is proposed, and the super efficiency interval FDEA (SEIFDEA) model is constructed. According to the different optimisms of the decision makers, the Hurwitz decision criterion is introduced for sorting. In addition, the model is applied to the performance evaluation system of logistics enterprises. Compared with the traditional DEA model, the validity and superiority of the model in fuzzy environment are verified.

1 August 2019
Verified
Conference Paper
Closed access
HWHongtao Wu
Hongtao Wu
JMJiteng Mu
Jiteng Mu
TDTing Da
Ting Da
MXMengdi Xu
Mengdi Xu
RTRussell H. Taylor
Russell H. Taylor
GCGregory S. Chirikjian
Gregory S. Chirikjian
Multi-mosquito object detection and 2D pose estimation are essential steps towards fully automated extracting PfSPZ-infected mosquito salivary glands for manufacture of PfSPZ Vaccine, which has been shown to protect against malaria in multiple clinical trials in the US, Europe, and Africa. This paper presents a deep learning approach to perform cluster condition classification and bounding box detection of multiple mosquitoes in an image. It also estimates the 2D pose of each non-clustered mosquito by body part detection. This approach is based on two popular convolutional neural network (CNN) architectures, Mask R-CNN and DeeperCut. In addition, we propose a cascaded image processing approach to achieve the multi-mosquito detection, cluster condition classification, and body parts detection in a multi-step manner. We compare the two approaches in terms of their functionality, robustness, accuracy, and speed. We hope our effective approaches would push forward the automation of PfSPZ Vaccine production to facilitate the prevention and elimination of this disease worldwide.

24 February 2021 in Minerals Engineering
Verified
Journal Article
Closed access
FGFangyu Guo
Fangyu Guo
QHQi He
Qi He
YXYaowen Xing
Yaowen Xing
YZYoufei Zhang
Youfei Zhang
SDShihao Ding
Shihao Ding
Xiahui GuiXiahui Gui
Xiahui Gui

1 August 2019
Verified
Conference Paper
Closed access
MXMengdi Xu
Mengdi Xu
RTRussell H. Taylor
Russell H. Taylor
GCGregory S. Chirikjian
Gregory S. Chirikjian
SLShengnan Lyu
Shengnan Lyu
YXYingtian Xu
Yingtian Xu
IULIAN IORDACHITAIULIAN IORDACHITA
IULIAN IORDACHITA
This paper describes the design of a fully automated apparatus to dispense mosquitoes into isolate units. This automation system consists of several process units including (1) facilitating the water vortex with a fan-shape rotor to gently transport the mosquitoes to the sorting slides with a conical geometry, (2) exploiting slides to guide mosquitoes to turntables driven by gears one by one, and (3) reorienting the mosquito until its proboscis points outward along the radial direction of the cone, aided by computer vision. This automation system serves as the first processing stage for collecting mosquito salivary glands. The sporozoites contained in the mosquito glands are the source material to produce Sanaria’s first generation PfSPZ vaccines. The Mosquito Staging System can dramatically enhance the mass production of Malaria Vaccine which is essential to prevent the propagation of Malaria.

14 June 2023 in arXiv
Verified
Preprint
Open access
PDF file
SLShiqi Liu
Shiqi Liu
MXMengdi Xu
Mengdi Xu
PHPiede Huang
Piede Huang
YLYongkang Liu
Yongkang Liu
KOKentaro Oguchi
Kentaro Oguchi
DZDing Zhao
Ding Zhao
Continual reinforcement learning aims to sequentially learn a variety of tasks, retaining the ability to perform previously encountered tasks while simultaneously developing new policies for novel tasks. However, current continual RL approaches overlook the fact that certain tasks are identical under basic group operations like rotations or translations, especially with visual inputs. They may unnecessarily learn and maintain a new policy for each similar task, leading to poor sample efficiency and weak generalization capability. To address this, we introduce a unique Continual Vision-based Reinforcement Learning method that recognizes Group Symmetries, called COVERS, cultivating a policy for each group of equivalent tasks rather than individual tasks. COVERS employs a proximal policy optimization-based RL algorithm with an equivariant feature extractor and a novel task grouping mechanism that relies on the extracted invariant features. We evaluate COVERS on sequences of table-top manipulation tasks that incorporate image observations and robot proprioceptive information in both simulations and on real robot platforms. Our results show that COVERS accurately assigns tasks to their respective groups and significantly outperforms existing methods in terms of generalization capability.