@article{
 title = {A Multioutput Convolved Gaussian Process for Capacity Forecasting of Li-Ion Battery Cells},
 type = {article},
 year = {2022},
 pages = {896-909},
 volume = {37},
 websites = {https://ieeexplore.ieee.org/document/9479693/},
 month = {1},
 id = {75b5a48c-1602-3b4b-8cd7-a98ccc4e9cde},
 created = {2021-11-08T20:03:10.663Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T20:03:10.663Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Chehade, Abdallah A. and Hussein, Ala A.},
 doi = {10.1109/TPEL.2021.3096164},
 journal = {IEEE Transactions on Power Electronics},
 number = {1}
}

@article{
 title = {Conditional Gaussian mixture model for warranty claims forecasting},
 type = {article},
 year = {2022},
 pages = {108180},
 volume = {218},
 websites = {https://linkinghub.elsevier.com/retrieve/pii/S0951832021006645},
 month = {2},
 id = {de6eff28-2b12-33fd-b18d-0502dccddabd},
 created = {2021-11-08T20:04:53.448Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2022-12-21T02:12:35.924Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Chehade, Abdallah and Savargaonkar, Mayuresh and Krivtsov, Vasiliy},
 doi = {10.1016/j.ress.2021.108180},
 journal = {Reliability Engineering & System Safety}
}

@article{
 title = {Uncorrelated Sparse Autoencoder With Long Short-Term Memory for State-of-Charge Estimations in Lithium-Ion Battery Cells},
 type = {article},
 year = {2022},
 pages = {1-12},
 websites = {https://ieeexplore.ieee.org/document/9959882/},
 id = {2e9bfb2b-3f4b-3234-84f6-e77603de779e},
 created = {2022-12-21T02:05:52.149Z},
 accessed = {2022-12-20},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2022-12-21T02:09:36.127Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Savargaonkar, Mayuresh and Oyewole, Isaiah and Chehade, Abdallah and Hussein, Ala A.},
 doi = {10.1109/TASE.2022.3222759},
 journal = {IEEE Transactions on Automation Science and Engineering}
}

@article{
 title = {VTrackIt: A Synthetic Self-Driving Dataset with Infrastructure and Pooled Vehicle Information},
 type = {article},
 year = {2022},
 keywords = {Index Terms-Autonomous vehicles,machine learning,reliability,safety,validation,verification},
 websites = {https://arxiv.org/abs/2207.11146v1},
 month = {7},
 day = {15},
 id = {d8acc8f2-cbc8-36c9-8d42-9d0a39b9a22e},
 created = {2022-12-21T02:08:53.640Z},
 accessed = {2022-12-20},
 file_attached = {true},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2022-12-21T02:10:29.413Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {Artificial intelligence solutions for Autonomous Vehicles (AVs) have been
developed using publicly available datasets such as Argoverse, ApolloScape,
Level5, and NuScenes. One major limitation of these datasets is the absence of
infrastructure and/or pooled vehicle information like lane line type, vehicle
speed, traffic signs, and intersections. Such information is necessary and not
complementary to eliminating high-risk edge cases. The rapid advancements in
Vehicle-to-Infrastructure and Vehicle-to-Vehicle technologies show promise that
infrastructure and pooled vehicle information will soon be accessible in near
real-time. Taking a leap in the future, we introduce the first comprehensive
synthetic dataset with intelligent infrastructure and pooled vehicle
information for advancing the next generation of AVs, named VTrackIt. We also
introduce the first deep learning model (InfraGAN) for trajectory predictions
that considers such information. Our experiments with InfraGAN show that the
comprehensive information offered by VTrackIt reduces the number of high-risk
edge cases. The VTrackIt dataset is available upon request under the Creative
Commons CC BY-NC-SA 4.0 license at http://vtrackit.irda.club.},
 bibtype = {article},
 author = {Savargaonkar, Mayuresh and Chehade, Abdallah},
 doi = {10.48550/arxiv.2207.11146}
}

Artificial intelligence solutions for Autonomous Vehicles (AVs) have been developed using publicly available datasets such as Argoverse, ApolloScape, Level5, and NuScenes. One major limitation of these datasets is the absence of infrastructure and/or pooled vehicle information like lane line type, vehicle speed, traffic signs, and intersections. Such information is necessary and not complementary to eliminating high-risk edge cases. The rapid advancements in Vehicle-to-Infrastructure and Vehicle-to-Vehicle technologies show promise that infrastructure and pooled vehicle information will soon be accessible in near real-time. Taking a leap in the future, we introduce the first comprehensive synthetic dataset with intelligent infrastructure and pooled vehicle information for advancing the next generation of AVs, named VTrackIt. We also introduce the first deep learning model (InfraGAN) for trajectory predictions that considers such information. Our experiments with InfraGAN show that the comprehensive information offered by VTrackIt reduces the number of high-risk edge cases. The VTrackIt dataset is available upon request under the Creative Commons CC BY-NC-SA 4.0 license at http://vtrackit.irda.club.

@article{
 title = {A Polynomial Regression Model with Bayesian Inference for State-of-Health Prediction of Li-ion Batteries},
 type = {article},
 year = {2022},
 keywords = {Battery Management System,Bayesian Statistics,Capacity,Li-Ion Battery,SOH},
 pages = {970-974},
 publisher = {Institute of Electrical and Electronics Engineers Inc.},
 id = {b35a3a46-02eb-3c65-8662-4d3a4ba722db},
 created = {2022-12-21T02:09:02.756Z},
 accessed = {2022-12-20},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2022-12-21T02:10:29.099Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {State-of-health (SOH) prediction is one of the key tasks of the Battery Management System (BMS) to ensure improved efficiency and safe operations of Lithium-ion (Li-ion) Batteries (LiBs). However, most of the existing SOH methods are either constrained by high model complexity or insufficient information about the historical degradation patterns of the battery cell. This paper proposes a Polynomial Regression Model with Bayesian Inference (PRMBI) for a robust SOH prediction of Li-ion batteries. The proposed PRMBI architecture leverages the strength of the semi-empirical modeling and data-driven methods for robust SOH prediction. The experimental results show that the proposed PRMBI significantly outperforms deep learning benchmark models when evaluated on battery cells that are still at early stages of degradation.},
 bibtype = {article},
 author = {Oyewole, Isaiah and Chelbi, Meriam and Chehade, Abdallah and Hussein, Ala A.},
 doi = {10.1109/ITEC53557.2022.9814038},
 journal = {2022 IEEE Transportation Electrification Conference and Expo, ITEC 2022}
}

State-of-health (SOH) prediction is one of the key tasks of the Battery Management System (BMS) to ensure improved efficiency and safe operations of Lithium-ion (Li-ion) Batteries (LiBs). However, most of the existing SOH methods are either constrained by high model complexity or insufficient information about the historical degradation patterns of the battery cell. This paper proposes a Polynomial Regression Model with Bayesian Inference (PRMBI) for a robust SOH prediction of Li-ion batteries. The proposed PRMBI architecture leverages the strength of the semi-empirical modeling and data-driven methods for robust SOH prediction. The experimental results show that the proposed PRMBI significantly outperforms deep learning benchmark models when evaluated on battery cells that are still at early stages of degradation.

@article{
 title = {A Novel Neural Network With Gaussian Process Feedback for Modeling the State-of-Charge of Battery Cells},
 type = {article},
 year = {2022},
 keywords = {Estimation,Gaussian processes (GPs),lithium-ion (Li-ion) battery,machine learning,state-of-charge (SOC)},
 pages = {5340-5352},
 volume = {58},
 publisher = {Institute of Electrical and Electronics Engineers Inc.},
 id = {0043bb70-5d6a-3945-a9ea-119b9cccc2fb},
 created = {2022-12-21T02:09:11.906Z},
 accessed = {2022-12-20},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2022-12-21T02:10:29.075Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {Although several state-of-charge (SOC) estimation methods have been proposed at the battery cell level, limited work has been done to identify the effect of cell aging on SOC estimations. To address this challenge, this article proposes a novel method for estimating the SOC of Lithium-ion (Li-ion) battery cells by accurately modeling the cell aging and degradation information. The proposed method, termed as 'NNGP,' is a deep neural network with Gaussian process feedback. The novel Gaussian process feedback helps the NNGP correlate SOC trends over consecutive charge-discharge cycles. Instead of time, the NNGP leverages available energy to correlate these SOC trends. The deep neural network within the NNGP then utilizes this information and other measured inputs to capture long-term cell aging and degradation trends. The NNGP leverages the most recent cell information to adapt its weights and estimate the SOC by employing an adaptive weighted training strategy. In our experiments on four Li-ion battery cells from three publicly available accelerated aging datasets, the NNGP clearly outperforms other benchmarked methods. The NNGP is also shown to be a useful prognostic tool capable of accurately estimating the SOC for up to 25 cycles in the future with an MAE below 3.5%. When tested under dynamic driving conditions and unknown initial SOC, the NNGP is shown to offer considerable improvements over other state-of-art methods. The derivation of the model followed by experimental evaluation is presented.},
 bibtype = {article},
 author = {Savargaonkar, Mayuresh and Chehade, Abdallah and Hussein, Ala A.},
 doi = {10.1109/TIA.2022.3170842},
 journal = {IEEE Transactions on Industry Applications},
 number = {4}
}

Although several state-of-charge (SOC) estimation methods have been proposed at the battery cell level, limited work has been done to identify the effect of cell aging on SOC estimations. To address this challenge, this article proposes a novel method for estimating the SOC of Lithium-ion (Li-ion) battery cells by accurately modeling the cell aging and degradation information. The proposed method, termed as 'NNGP,' is a deep neural network with Gaussian process feedback. The novel Gaussian process feedback helps the NNGP correlate SOC trends over consecutive charge-discharge cycles. Instead of time, the NNGP leverages available energy to correlate these SOC trends. The deep neural network within the NNGP then utilizes this information and other measured inputs to capture long-term cell aging and degradation trends. The NNGP leverages the most recent cell information to adapt its weights and estimate the SOC by employing an adaptive weighted training strategy. In our experiments on four Li-ion battery cells from three publicly available accelerated aging datasets, the NNGP clearly outperforms other benchmarked methods. The NNGP is also shown to be a useful prognostic tool capable of accurately estimating the SOC for up to 25 cycles in the future with an MAE below 3.5%. When tested under dynamic driving conditions and unknown initial SOC, the NNGP is shown to offer considerable improvements over other state-of-art methods. The derivation of the model followed by experimental evaluation is presented.

@article{
 title = {A controllable deep transfer learning network with multiple domain adaptation for battery state-of-charge estimation},
 type = {article},
 year = {2022},
 keywords = {Deep learning model,Lithium-ion batteries,Multiple domain adaptation,State-of-charge,Transfer learning},
 pages = {118726},
 volume = {312},
 month = {4},
 publisher = {Elsevier},
 day = {15},
 id = {c67a3d8e-0350-3b80-bda8-7b537f7ea6a8},
 created = {2022-12-21T02:09:17.111Z},
 accessed = {2022-12-20},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2022-12-21T02:10:29.307Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {Deep learning models have been drawing significant attention in the literature of state-of-charge (SOC) estimation because of their capabilities to capture non-trivial temporal patterns. However, most of such models ignore cell-to-cell variations or focus on short-term estimations that are not practical for battery cells with limited charging-discharging history. We propose a Controllable Deep Transfer Learning (CDTL) network for short and long-term SOC estimations at early stages of degradation. The CDTL utilizes shared knowledge between the target cells of interest and historical source cells with rich SOC data using controllable Multiple Domain Adaptation (MDA). Specifically, the CDTL consists of two long-short term memory (LSTM) networks, the source LSTM, and the target LSTM. The source LSTM is trained on SOC data from historical battery cells. The target LSTM is then trained using limited available SOC data from the target cell and the transferred knowledge from the source LSTM using controllable MDA with adaptive regularization. The contributions of the CDTL are two-folded. First, it reduces the likelihood of negative transfer learning using controllable MDA with adaptive regularization, which enhances the target LSTM generalizability for long-term SOC estimation. Second, the CDTL offers theoretical guarantees on the controllability and convergence of transferred knowledge from the source cell to target cell. The experimental results demonstrate that the proposed CDTL outperforms existing deep and transfer learning benchmarks with 1) a maximum improvement of 60% in root-mean-squared error (RMSE) for battery cells with the same chemistry, 2) an average improvement of 50% in RMSE across different battery chemistries, and 3) about 39% reduction in computational time.},
 bibtype = {article},
 author = {Oyewole, Isaiah and Chehade, Abdallah and Kim, Youngki},
 doi = {10.1016/J.APENERGY.2022.118726},
 journal = {Applied Energy}
}

Deep learning models have been drawing significant attention in the literature of state-of-charge (SOC) estimation because of their capabilities to capture non-trivial temporal patterns. However, most of such models ignore cell-to-cell variations or focus on short-term estimations that are not practical for battery cells with limited charging-discharging history. We propose a Controllable Deep Transfer Learning (CDTL) network for short and long-term SOC estimations at early stages of degradation. The CDTL utilizes shared knowledge between the target cells of interest and historical source cells with rich SOC data using controllable Multiple Domain Adaptation (MDA). Specifically, the CDTL consists of two long-short term memory (LSTM) networks, the source LSTM, and the target LSTM. The source LSTM is trained on SOC data from historical battery cells. The target LSTM is then trained using limited available SOC data from the target cell and the transferred knowledge from the source LSTM using controllable MDA with adaptive regularization. The contributions of the CDTL are two-folded. First, it reduces the likelihood of negative transfer learning using controllable MDA with adaptive regularization, which enhances the target LSTM generalizability for long-term SOC estimation. Second, the CDTL offers theoretical guarantees on the controllability and convergence of transferred knowledge from the source cell to target cell. The experimental results demonstrate that the proposed CDTL outperforms existing deep and transfer learning benchmarks with 1) a maximum improvement of 60% in root-mean-squared error (RMSE) for battery cells with the same chemistry, 2) an average improvement of 50% in RMSE across different battery chemistries, and 3) about 39% reduction in computational time.

@article{
 title = {A dual-LSTM framework combining change point detection and remaining useful life prediction},
 type = {article},
 year = {2021},
 pages = {107257},
 volume = {205},
 websites = {https://linkinghub.elsevier.com/retrieve/pii/S0951832020307572},
 month = {1},
 id = {e47bf46f-ae78-37a2-8003-14387fc0f165},
 created = {2020-10-04T16:04:16.609Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T20:03:10.909Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Shi, Zunya and Chehade, Abdallah},
 doi = {10.1016/j.ress.2020.107257},
 journal = {Reliability Engineering & System Safety}
}

@article{
 title = {Dynamic Adherent Raindrop Simulator for Automotive Vision Systems},
 type = {article},
 year = {2021},
 pages = {114808-114820},
 volume = {9},
 websites = {https://ieeexplore.ieee.org/document/9509513/},
 id = {ac586202-d6d3-3497-8041-604e4dd08615},
 created = {2021-11-08T19:55:56.622Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T20:03:10.915Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Hamzeh, Yazan and El-Shair, Zaid A. and Chehade, Abdallah and Rawashdeh, Samir A.},
 doi = {10.1109/ACCESS.2021.3103895},
 journal = {IEEE Access}
}

@article{
 title = {RMOPP: Robust Multi-Objective Post-Processing for Effective Object Detection},
 type = {article},
 year = {2021},
 websites = {http://arxiv.org/abs/2102.04582},
 month = {2},
 day = {8},
 id = {8e40ce1f-a506-3fd3-9261-2f5142979293},
 created = {2021-11-08T20:03:10.202Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T20:03:10.202Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 abstract = {Over the last few decades, many architectures have been developed that harness the power of neural networks to detect objects in near real-time. Training such systems requires substantial time across multiple GPUs and massive labeled training datasets. Although the goal of these systems is generalizability, they are often impractical in real-life applications due to flexibility, robustness, or speed issues. This paper proposes RMOPP: A robust multi-objective post-processing algorithm to boost the performance of fast pre-trained object detectors with a negligible impact on their speed. Specifically, RMOPP is a statistically driven, post-processing algorithm that allows for simultaneous optimization of precision and recall. A unique feature of RMOPP is the Pareto frontier that identifies dominant possible post-processed detectors to optimize for both precision and recall. RMOPP explores the full potential of a pre-trained object detector and is deployable for near real-time predictions. We also provide a compelling test case on YOLOv2 using the MS-COCO dataset.},
 bibtype = {article},
 author = {Savargaonkar, Mayuresh and Chehade, Abdallah and Rawashdeh, Samir}
}

Over the last few decades, many architectures have been developed that harness the power of neural networks to detect objects in near real-time. Training such systems requires substantial time across multiple GPUs and massive labeled training datasets. Although the goal of these systems is generalizability, they are often impractical in real-life applications due to flexibility, robustness, or speed issues. This paper proposes RMOPP: A robust multi-objective post-processing algorithm to boost the performance of fast pre-trained object detectors with a negligible impact on their speed. Specifically, RMOPP is a statistically driven, post-processing algorithm that allows for simultaneous optimization of precision and recall. A unique feature of RMOPP is the Pareto frontier that identifies dominant possible post-processed detectors to optimize for both precision and recall. RMOPP explores the full potential of a pre-trained object detector and is deployable for near real-time predictions. We also provide a compelling test case on YOLOv2 using the MS-COCO dataset.

@inproceedings{
 title = {Sparse Autoencoded Long Short-Term Memory Network for State-of-Charge Estimations},
 type = {inproceedings},
 year = {2021},
 pages = {474-478},
 websites = {https://ieeexplore.ieee.org/document/9490070/},
 month = {6},
 publisher = {IEEE},
 day = {21},
 id = {1894b0c2-df03-3e29-a2fc-2faa967117cb},
 created = {2021-11-08T20:03:10.376Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T20:03:10.376Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {Savargaonkar, Mayuresh and Oyewole, Isaiah and Chehade, Abdallah},
 doi = {10.1109/ITEC51675.2021.9490070},
 booktitle = {2021 IEEE Transportation Electrification Conference & Expo (ITEC)}
}

@inproceedings{
 title = {A Hybrid Long Short-Term Memory Network for State-of-Charge Estimation of Li-ion Batteries},
 type = {inproceedings},
 year = {2021},
 pages = {469-473},
 websites = {https://ieeexplore.ieee.org/document/9490188/},
 month = {6},
 publisher = {IEEE},
 day = {21},
 id = {eb713b8f-d6ea-35c3-98fb-72f27b45c830},
 created = {2021-11-08T20:03:10.521Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T20:03:10.521Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {Oyewole, Isaiah and Savargaonkar, Mayuresh and Chehade, Abdallah and Kim, Youngki},
 doi = {10.1109/ITEC51675.2021.9490188},
 booktitle = {2021 IEEE Transportation Electrification Conference & Expo (ITEC)}
}

@article{
 title = {Dynamic Adherent Raindrop Simulator for Automotive Vision Systems},
 type = {article},
 year = {2021},
 keywords = {Adherent raindrops,automotive domain,deep-learning,image degradation,object detection,raindrop simulator,recall,similarity metrics},
 pages = {114808-114820},
 volume = {9},
 publisher = {Institute of Electrical and Electronics Engineers Inc.},
 id = {00feb766-ac4c-31a4-85ba-93f48c843119},
 created = {2022-12-21T02:09:29.089Z},
 accessed = {2022-12-20},
 file_attached = {true},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2022-12-21T02:10:29.387Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {The automotive domain is highly regulated, with many safety-critical aspects to consider. This means that a great deal of testing is required to validate the performance of automotive systems, under all possible environmental conditions. For vision-based systems, camera images are among the most important input sources of information, and using high-quality images is integral to the system's performance. Rain, as a type of adverse weather condition, degrades the image quality which reflects negatively on the vision-based algorithms. Collecting representative sets of data under different rain conditions is required for system testing and performance evaluation. This usually is both costly and time-consuming. Augmenting the sets of real rained images in system testing is an attractive, feasible alternative. In this paper, we present an adherent rain simulator system, that adds simulated rain to clear image frames captured in real drive cycles. We test the quality of simulated rained images against real rained ones, using common image similarity metrics. We also compare the performance of deep learning-based object detectors, using our simulated rained images vs. real rained images. The results show that object detectors show similar performance using simulated and real rained images. A comparative analysis shows that our model produces more realistic raindrops, compared to a ray-tracing-based raindrop simulator.},
 bibtype = {article},
 author = {Hamzeh, Yazan and El-Shair, Zaid A. and Chehade, Abdallah and Rawashdeh, Samir A.},
 doi = {10.1109/ACCESS.2021.3103895},
 journal = {IEEE Access}
}

The automotive domain is highly regulated, with many safety-critical aspects to consider. This means that a great deal of testing is required to validate the performance of automotive systems, under all possible environmental conditions. For vision-based systems, camera images are among the most important input sources of information, and using high-quality images is integral to the system's performance. Rain, as a type of adverse weather condition, degrades the image quality which reflects negatively on the vision-based algorithms. Collecting representative sets of data under different rain conditions is required for system testing and performance evaluation. This usually is both costly and time-consuming. Augmenting the sets of real rained images in system testing is an attractive, feasible alternative. In this paper, we present an adherent rain simulator system, that adds simulated rain to clear image frames captured in real drive cycles. We test the quality of simulated rained images against real rained ones, using common image similarity metrics. We also compare the performance of deep learning-based object detectors, using our simulated rained images vs. real rained images. The results show that object detectors show similar performance using simulated and real rained images. A comparative analysis shows that our model produces more realistic raindrops, compared to a ray-tracing-based raindrop simulator.

@article{
 title = {Data-driven Adaptive Thresholding Model for Real-time Valve Delay Estimation in Digital Pump/Motors},
 type = {article},
 year = {2020},
 pages = {271–294},
 volume = {20},
 websites = {https://journals.riverpublishers.com/index.php/IJFP/article/view/301},
 month = {3},
 day = {9},
 id = {07b7cd6e-780b-3df3-a931-6e8a90f87198},
 created = {2020-03-12T20:00:16.951Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T19:47:06.741Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {<p>Valve characteristics are an essential part of digital hydraulics. The on/off solenoid valves utilized on many of these systems can significantly affect the performance. Various factors can affect the speed of the valves causing them to experience various delays, which impact the overall performance of hydraulic systems. This work presents the development of an adaptive statistical based thresholding real-time valve delay model for digital Pump/Motors. The proposed method actively measures the valve delays in real-time and adapts the threshold of the system with the goal of improving the overall efficiency and performance of the system. This work builds on previous work by evaluating an alternative method used to detect valve delays in real-time. The method used here is a shift detection method for the pressure signals that utilizes domain knowledge and the system’s historical statistical behavior. This allows the model to be used over a large range of operating conditions, since the model can learn patterns and adapt to various operating conditions using domain knowledge and statistical behavior. A hydraulic circuit was built to measure the delay time experienced from the time the signal is sent to the valve to the time that the valve opens. Experiments were conducted on a three piston in-line digital pump/motor with 2 valves per cylinder, at low and high pressure ports, for a total of six valves. Two high frequency pressure transducers were used in this circuit to measure and analyze the differential pressure on the low and high pressure side of the on/off valves, as well as three in-cylinder pressure transducers. Data over 60 cycles was acquired to analyze the model against real time valve delays. The results show that the algorithm was successful in adapting the threshold for real time valve delays and accurately measuring the valve delays. </p>},
 bibtype = {article},
 author = {Chehade, Abdallah and Breidi, Farid and Pate, Keith Scott and Lumkes, John},
 doi = {10.13052/ijfp1439-9776.2031},
 journal = {International Journal of Fluid Power},
 number = {3}
}

Valve characteristics are an essential part of digital hydraulics. The on/off solenoid valves utilized on many of these systems can significantly affect the performance. Various factors can affect the speed of the valves causing them to experience various delays, which impact the overall performance of hydraulic systems. This work presents the development of an adaptive statistical based thresholding real-time valve delay model for digital Pump/Motors. The proposed method actively measures the valve delays in real-time and adapts the threshold of the system with the goal of improving the overall efficiency and performance of the system. This work builds on previous work by evaluating an alternative method used to detect valve delays in real-time. The method used here is a shift detection method for the pressure signals that utilizes domain knowledge and the system’s historical statistical behavior. This allows the model to be used over a large range of operating conditions, since the model can learn patterns and adapt to various operating conditions using domain knowledge and statistical behavior. A hydraulic circuit was built to measure the delay time experienced from the time the signal is sent to the valve to the time that the valve opens. Experiments were conducted on a three piston in-line digital pump/motor with 2 valves per cylinder, at low and high pressure ports, for a total of six valves. Two high frequency pressure transducers were used in this circuit to measure and analyze the differential pressure on the low and high pressure side of the on/off valves, as well as three in-cylinder pressure transducers. Data over 60 cycles was acquired to analyze the model against real time valve delays. The results show that the algorithm was successful in adapting the threshold for real time valve delays and accurately measuring the valve delays. 

@article{
 title = {BLNN: An R package for training neural networks using Bayesian inference},
 type = {article},
 year = {2020},
 pages = {100432},
 volume = {11},
 websites = {https://linkinghub.elsevier.com/retrieve/pii/S235271101930322X},
 month = {1},
 id = {46684754-dd2e-3f58-a0c0-9dff324f19c0},
 created = {2020-03-12T20:17:12.588Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T19:47:09.550Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Sharaf, Taysseer and Williams, Theren and Chehade, Abdallah and Pokhrel, Keshav},
 doi = {10.1016/j.softx.2020.100432},
 journal = {SoftwareX}
}

@article{
 title = {Accelerating the Discovery of New DP Steel Using Machine Learning-Based Multiscale Materials Simulations},
 type = {article},
 year = {2020},
 pages = {3268-3279},
 volume = {51},
 websites = {https://link.springer.com/10.1007/s11661-020-05764-7},
 month = {6},
 day = {19},
 id = {5f31b087-9463-319e-96f0-52a0e2fb518e},
 created = {2020-04-21T19:23:40.633Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T19:47:14.647Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Chehade, Abdallah A. and Belgasam, Tarek M. and Ayoub, Georges and Zbib, Hussein M.},
 doi = {10.1007/s11661-020-05764-7},
 journal = {Metallurgical and Materials Transactions A},
 number = {6}
}

@article{
 title = {Robust Artificial Neural Network-Based Models for Accurate Surface Temperature Estimation of Batteries},
 type = {article},
 year = {2020},
 pages = {5269-5278},
 volume = {56},
 websites = {https://ieeexplore.ieee.org/document/9113323/},
 month = {9},
 id = {254f0c74-41a0-3d27-9961-ade1941bba19},
 created = {2020-06-13T22:41:09.648Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T19:47:18.790Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Hussein, Ala A. and Chehade, Abdallah A.},
 doi = {10.1109/TIA.2020.3001256},
 journal = {IEEE Transactions on Industry Applications},
 number = {5}
}

@article{
 title = {Power–law nonhomogeneous Poisson process with a mixture of latent common shape parameters},
 type = {article},
 year = {2020},
 pages = {107097},
 volume = {203},
 websites = {https://linkinghub.elsevier.com/retrieve/pii/S0951832020305986},
 month = {11},
 id = {e148e1bc-fbd7-3713-8bb5-e5576175296c},
 created = {2020-06-25T03:41:08.930Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T19:47:21.535Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Chehade, Abdallah and Shi, Zunya and Krivtsov, Vasiliy},
 doi = {10.1016/j.ress.2020.107097},
 journal = {Reliability Engineering & System Safety}
}

@inproceedings{
 title = {A Cycle-based Recurrent Neural Network for State-of-Charge Estimation of Li-ion Battery Cells},
 type = {inproceedings},
 year = {2020},
 pages = {584-587},
 websites = {https://ieeexplore.ieee.org/document/9161587/},
 month = {6},
 publisher = {IEEE},
 id = {a77e89fd-68a7-3162-a8a1-42df95acd396},
 created = {2020-08-11T18:14:11.014Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T19:47:26.040Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {Savargaonkar, Mayuresh and Chehade, Abdallah and Shi, Zunya and Hussein, Ala A.},
 doi = {10.1109/ITEC48692.2020.9161587},
 booktitle = {2020 IEEE Transportation Electrification Conference & Expo (ITEC)}
}

@inproceedings{
 title = {An Adaptive Deep Neural Network with Transfer Learning for State-of-Charge Estimations of Battery Cells},
 type = {inproceedings},
 year = {2020},
 pages = {598-602},
 websites = {https://ieeexplore.ieee.org/document/9161464/},
 month = {6},
 publisher = {IEEE},
 id = {d9401bd3-8c51-331b-8a01-1c1900b6d44b},
 created = {2020-08-11T18:14:11.037Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T19:47:34.960Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {Savargaonkar, Mayuresh and Chehade, Abdallah},
 doi = {10.1109/ITEC48692.2020.9161464},
 booktitle = {2020 IEEE Transportation Electrification Conference & Expo (ITEC)}
}

@inproceedings{
 title = {A Long Short-Term Memory Network for Online State-of-Charge Estimation of Li-ion Battery Cells},
 type = {inproceedings},
 year = {2020},
 pages = {594-597},
 websites = {https://ieeexplore.ieee.org/document/9161487/},
 month = {6},
 publisher = {IEEE},
 id = {87a65cc0-9924-3d1e-b622-35d9aff5c21a},
 created = {2020-08-11T18:14:11.305Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T19:47:32.954Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {Shi, Zunya and Savargaonkar, Mayuresh and Chehade, Abdallah A. and Hussein, Ala A.},
 doi = {10.1109/ITEC48692.2020.9161487},
 booktitle = {2020 IEEE Transportation Electrification Conference & Expo (ITEC)}
}

@article{
 title = {A Collaborative Gaussian Process Regression Model for Transfer Learning of Capacity Trends Between Li-Ion Battery Cells},
 type = {article},
 year = {2020},
 pages = {9542-9552},
 volume = {69},
 websites = {https://ieeexplore.ieee.org/document/9112271/},
 month = {9},
 id = {61ed88af-b01e-3ebb-b86c-a5d7604f12f1},
 created = {2023-02-18T15:50:29.778Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2023-02-18T15:50:29.778Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Chehade, Abdallah A. and Hussein, Ala A.},
 doi = {10.1109/TVT.2020.3000970},
 journal = {IEEE Transactions on Vehicular Technology},
 number = {9}
}

@article{
 title = {Structural Degradation Modeling Framework for Sparse Data Sets With an Application on Alzheimer’s Disease},
 type = {article},
 year = {2019},
 pages = {192-205},
 volume = {16},
 websites = {https://ieeexplore.ieee.org/document/8361039/},
 month = {1},
 id = {0a524121-4890-340f-a87f-fd6c5bc22bfe},
 created = {2018-10-22T01:11:19.375Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2019-05-07T16:45:05.962Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Chehade, Abdallah and Liu, Kaibo},
 doi = {10.1109/TASE.2018.2829770},
 journal = {IEEE Transactions on Automation Science and Engineering},
 number = {1}
}

@article{
 title = {Sensor Fusion via Statistical Hypothesis Testing for Prognosis and Degradation Analysis},
 type = {article},
 year = {2019},
 pages = {1774-1787},
 volume = {16},
 websites = {https://ieeexplore.ieee.org/document/8651299/},
 month = {10},
 id = {93867458-4403-3add-8eb7-f06ac7ad83b8},
 created = {2019-03-09T21:59:07.836Z},
 accessed = {2019-03-09},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-05-09T17:17:35.461Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 folder_uuids = {5c271c5e-8151-4427-8394-48c28155bd51},
 private_publication = {false},
 bibtype = {article},
 author = {Chehade, Abdallah and Shi, Zunya},
 doi = {10.1109/TASE.2019.2897784},
 journal = {IEEE Transactions on Automation Science and Engineering},
 number = {4}
}

@article{
 title = {Latent Function Decomposition for Forecasting Li-ion Battery Cells Capacity: A Multi-Output Convolved Gaussian Process Approach},
 type = {article},
 year = {2019},
 websites = {http://arxiv.org/abs/1907.09455},
 month = {7},
 day = {19},
 id = {a1194a01-4616-3c56-9c9b-f83e174b8248},
 created = {2019-07-17T21:06:14.854Z},
 accessed = {2019-07-17},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2019-07-23T15:01:23.508Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 folder_uuids = {5c271c5e-8151-4427-8394-48c28155bd51},
 private_publication = {false},
 abstract = {A latent function decomposition method is proposed for forecasting the capacity of lithium-ion battery cells. The method uses the Multi-Output Gaussian Process, a generative machine learning framework for multi-task and transfer learning. The MCGP decomposes the available capacity trends from multiple battery cells into latent functions. The latent functions are then convolved over kernel smoothers to reconstruct and/or forecast capacity trends of the battery cells. Besides the high prediction accuracy the proposed method possesses, it provides uncertainty information for the predictions and captures nontrivial cross-correlations between capacity trends of different battery cells. These two merits make the proposed MCGP a very reliable and practical solution for applications that use battery cell packs. The MCGP is derived and compared to benchmark methods on an experimental lithium-ion battery cells data. The results show the effectiveness of the proposed method.},
 bibtype = {article},
 author = {Chehade, Abdallah A. and Hussein, Ala A.}
}

A latent function decomposition method is proposed for forecasting the capacity of lithium-ion battery cells. The method uses the Multi-Output Gaussian Process, a generative machine learning framework for multi-task and transfer learning. The MCGP decomposes the available capacity trends from multiple battery cells into latent functions. The latent functions are then convolved over kernel smoothers to reconstruct and/or forecast capacity trends of the battery cells. Besides the high prediction accuracy the proposed method possesses, it provides uncertainty information for the predictions and captures nontrivial cross-correlations between capacity trends of different battery cells. These two merits make the proposed MCGP a very reliable and practical solution for applications that use battery cell packs. The MCGP is derived and compared to benchmark methods on an experimental lithium-ion battery cells data. The results show the effectiveness of the proposed method.

@inproceedings{
 title = {A Multi-Output Convolved Gaussian Process Model for Capacity Estimation of Electric Vehicle Li-ion Battery Cells},
 type = {inproceedings},
 year = {2019},
 pages = {1-4},
 websites = {https://ieeexplore.ieee.org/document/8790463/},
 month = {6},
 publisher = {IEEE},
 id = {3d87aa7e-140e-355e-a6dd-fe1d03a3f058},
 created = {2019-09-16T21:44:35.426Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2021-11-08T19:46:20.786Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {Chehade, Abdallah A. and Hussein, Ala A.},
 doi = {10.1109/ITEC.2019.8790463},
 booktitle = {2019 IEEE Transportation Electrification Conference and Expo (ITEC)}
}

@article{
 title = {The Sparse Reverse of Principal Component Analysis for Fast Low-Rank Matrix Completion},
 type = {article},
 year = {2019},
 websites = {http://arxiv.org/abs/1910.02155},
 month = {10},
 day = {4},
 id = {f5596fd6-ca64-3c49-81b6-de85b1c7cf6f},
 created = {2019-10-08T14:06:42.129Z},
 accessed = {2019-10-08},
 file_attached = {true},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2019-10-23T15:22:36.520Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 folder_uuids = {f674219e-688a-47a1-9a57-f2fbed097a63},
 private_publication = {false},
 abstract = {Matrix completion constantly receives tremendous attention from many research fields. It is commonly applied for recommender systems such as movie ratings, computer vision such as image reconstruction or completion, multi-task learning such as collaboratively modeling time-series trends of multiple sensors, and many other applications. Matrix completion techniques are usually computationally exhaustive and/or fail to capture the heterogeneity in the data. For example, images usually contain a heterogeneous set of objects, and thus it is a challenging task to reconstruct images with high levels of missing data. In this paper, we propose the sparse reverse of principal component analysis for matrix completion. The proposed approach maintains smoothness across the matrix, produces accurate estimates of the missing data, converges iteratively, and it is computationally tractable with a controllable upper bound on the number of iterations until convergence. The accuracy of the proposed technique is validated on natural images, movie ratings, and multisensor data. It is also compared with common benchmark methods used for matrix completion.},
 bibtype = {article},
 author = {Chehade, Abdallah and Shi, Zunya}
}

Matrix completion constantly receives tremendous attention from many research fields. It is commonly applied for recommender systems such as movie ratings, computer vision such as image reconstruction or completion, multi-task learning such as collaboratively modeling time-series trends of multiple sensors, and many other applications. Matrix completion techniques are usually computationally exhaustive and/or fail to capture the heterogeneity in the data. For example, images usually contain a heterogeneous set of objects, and thus it is a challenging task to reconstruct images with high levels of missing data. In this paper, we propose the sparse reverse of principal component analysis for matrix completion. The proposed approach maintains smoothness across the matrix, produces accurate estimates of the missing data, converges iteratively, and it is computationally tractable with a controllable upper bound on the number of iterations until convergence. The accuracy of the proposed technique is validated on natural images, movie ratings, and multisensor data. It is also compared with common benchmark methods used for matrix completion.

@inproceedings{
 title = {Monitoring Digital Technologies in Hydraulic Systems Using CUSUM Control Charts},
 type = {inproceedings},
 year = {2019},
 keywords = {Condition monitoring,Delays,Dynamics (Mechanics),Electronics,Errors,Fluid power systems,Fluids,Hydraulic drive systems,Hydraulics,Maintenance,Motors,Pistons,Pressure,Pumps,Quality control charts,Statistical process control,Time series,Valves},
 websites = {https://asmedigitalcollection.asme.org/FPST/proceedings/FPMC2019/59339/Longboat Key, Florida, USA/1071793},
 month = {10},
 publisher = {American Society of Mechanical Engineers},
 day = {7},
 id = {2a4c9787-b213-3fd0-821d-daa0c53c6f45},
 created = {2019-12-20T23:02:39.935Z},
 accessed = {2019-12-20},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2019-12-20T23:03:23.910Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 folder_uuids = {a2b1e262-6fa0-4430-a851-a73a28b3b5e4},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {Breidi, Farid and Chehade, Abdallah and Lumkes, John},
 doi = {10.1115/FPMC2019-1603},
 booktitle = {ASME/BATH 2019 Symposium on Fluid Power and Motion Control}
}

@misc{
 title = {The sparse reverse of principal component analysis for fast low-rank matrix completion},
 type = {misc},
 year = {2019},
 source = {arXiv},
 keywords = {Image reconstruction,Matrix completion,PCA,Recommender system,Sensor fusion,Subspace learning,collaborative filtering},
 id = {4cfcc6ee-a964-3f6a-93d8-2e0c4fdb7fb9},
 created = {2020-10-27T23:59:00.000Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2020-10-28T17:52:31.349Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {Copyright © 2019, arXiv, All rights reserved. —Matrix completion constantly receives tremendous attention from many research fields. It is commonly applied for recommender systems such as movie ratings, computer vision such as image reconstruction or completion, multi-task learning such as collaboratively modeling time-series trends of multiple sensors, and many other applications. Matrix completion techniques are usually computationally exhaustive and/or fail to capture the heterogeneity in the data. For example, images usually contain a heterogeneous set of objects, and thus it is a challenging task to reconstruct images with high levels of missing data. In this paper, we propose the sparse reverse of principal component analysis for matrix completion. The proposed approach maintains smoothness across the matrix, produces accurate estimates of the missing data, converges iteratively, and it is computationally tractable with a controllable upper bound on the number of iterations until convergence. The accuracy of the proposed technique is validated on natural images, movie ratings, and multisensor data. It is also compared with common benchmark methods used for matrix completion.},
 bibtype = {misc},
 author = {Chehade, A. and Shi, Z.}
}

Copyright © 2019, arXiv, All rights reserved. —Matrix completion constantly receives tremendous attention from many research fields. It is commonly applied for recommender systems such as movie ratings, computer vision such as image reconstruction or completion, multi-task learning such as collaboratively modeling time-series trends of multiple sensors, and many other applications. Matrix completion techniques are usually computationally exhaustive and/or fail to capture the heterogeneity in the data. For example, images usually contain a heterogeneous set of objects, and thus it is a challenging task to reconstruct images with high levels of missing data. In this paper, we propose the sparse reverse of principal component analysis for matrix completion. The proposed approach maintains smoothness across the matrix, produces accurate estimates of the missing data, converges iteratively, and it is computationally tractable with a controllable upper bound on the number of iterations until convergence. The accuracy of the proposed technique is validated on natural images, movie ratings, and multisensor data. It is also compared with common benchmark methods used for matrix completion.

@misc{
 title = {Latent function decomposition for forecasting li-ion battery cells capacity: A multi-output convolved gaussian process approach},
 type = {misc},
 year = {2019},
 source = {arXiv},
 keywords = {Convolution process,Lithium-ion battery cell,Multi-output Gaussian process,Multi-task learning,Remaining useful life,State-of-charge,Transfer learning,—Capacity},
 id = {44c50cdc-43e4-3a86-9c9d-13dc5535cf65},
 created = {2020-10-27T23:59:00.000Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2020-10-28T21:05:06.368Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {Copyright © 2019, arXiv, All rights reserved. —A latent function decomposition method is proposed for forecasting the capacity of lithium-ion battery cells. The method uses the Multi-Output Gaussian Process, a generative machine learning framework for multi-task and transfer learning. The MCGP decomposes the available capacity trends from multiple battery cells into latent functions. The latent functions are then convolved over kernel smoothers to reconstruct and/or forecast capacity trends of the battery cells. Besides the high prediction accuracy the proposed method possesses, it provides uncertainty information for the predictions and captures nontrivial cross-correlations between capacity trends of different battery cells. These two merits make the proposed MCGP a very reliable and practical solution for applications that use battery cell packs. The MCGP is derived and compared to benchmark methods on an experimental lithium-ion battery cells data. The results show the effectiveness of the proposed method.},
 bibtype = {misc},
 author = {Chehade, A.A. and Hussein, A.A.}
}

Copyright © 2019, arXiv, All rights reserved. —A latent function decomposition method is proposed for forecasting the capacity of lithium-ion battery cells. The method uses the Multi-Output Gaussian Process, a generative machine learning framework for multi-task and transfer learning. The MCGP decomposes the available capacity trends from multiple battery cells into latent functions. The latent functions are then convolved over kernel smoothers to reconstruct and/or forecast capacity trends of the battery cells. Besides the high prediction accuracy the proposed method possesses, it provides uncertainty information for the predictions and captures nontrivial cross-correlations between capacity trends of different battery cells. These two merits make the proposed MCGP a very reliable and practical solution for applications that use battery cell packs. The MCGP is derived and compared to benchmark methods on an experimental lithium-ion battery cells data. The results show the effectiveness of the proposed method.

@article{
 title = {A data-level fusion approach for degradation modeling and prognostic analysis under multiple failure modes},
 type = {article},
 year = {2018},
 pages = {150-165},
 volume = {50},
 websites = {https://www.tandfonline.com/doi/full/10.1080/00224065.2018.1436829},
 month = {4},
 day = {3},
 id = {e0e961a7-4df9-3a54-b9b8-4f1c11d8e18a},
 created = {2018-04-23T17:48:26.260Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2018-04-23T17:48:26.260Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Chehade, Abdallah and Song, Changyue and Liu, Kaibo and Saxena, Abhinav and Zhang, Xi},
 doi = {10.1080/00224065.2018.1436829},
 journal = {Journal of Quality Technology},
 number = {2}
}

@inproceedings{
 title = {Design of a Transparent Hydraulic/Pneumatic Excavator Arm for Teaching and Outreach Activities},
 type = {inproceedings},
 year = {2018},
 websites = {https://peer.asee.org/30266},
 month = {6},
 day = {23},
 id = {97abdf6f-2b80-3c18-982b-7e187d1d9bb1},
 created = {2019-02-10T17:39:30.781Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2019-02-10T17:56:05.794Z},
 read = {true},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 folder_uuids = {a2b1e262-6fa0-4430-a851-a73a28b3b5e4},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {Pate, Keith and Marx, Joseph and Chehade, Abdallah and Breidi, Farid},
 booktitle = {2018 ASEE Annual Conference & Exposition}
}

@article{
 title = {Optimize the Signal Quality of the Composite Health Index via Data Fusion for Degradation Modeling and Prognostic Analysis},
 type = {article},
 year = {2017},
 pages = {1504-1514},
 volume = {14},
 websites = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=7165684,http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7165684,http://ieeexplore.ieee.org/document/7165684/},
 month = {7},
 id = {06077f3c-4987-3e09-acbc-b8a9eea43a6d},
 created = {2016-09-18T17:51:57.000Z},
 accessed = {2016-09-18},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2017-12-10T18:15:01.426Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 citation_key = {Liu2015},
 folder_uuids = {5780280a-8847-41f0-9ec1-674ac415515f},
 private_publication = {false},
 bibtype = {article},
 author = {Liu, Kaibo and Chehade, Abdallah and Song, Changyue},
 doi = {10.1109/TASE.2015.2446752},
 journal = {IEEE Transactions on Automation Science and Engineering},
 number = {3}
}

@article{
 title = {Sensory-Based Failure Threshold Estimation for Remaining Useful Life Prediction},
 type = {article},
 year = {2017},
 pages = {939-949},
 volume = {66},
 websites = {http://ieeexplore.ieee.org/document/7924404/},
 month = {9},
 id = {a6bdec02-81e2-3816-a19f-08c0456e9a58},
 created = {2017-07-23T15:18:08.352Z},
 accessed = {2017-07-23},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2017-12-10T18:19:19.324Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Chehade, Abdallah and Bonk, Scott and Liu, Kaibo},
 doi = {10.1109/TR.2017.2695119},
 journal = {IEEE Transactions on Reliability},
 number = {3}
}

@phdthesis{
 title = {Data-driven Approaches for Condition Monitoring and Predictive Analytics.},
 type = {phdthesis},
 year = {2017},
 source = {ProQuest},
 id = {37d35f63-18f0-39ef-b28f-d2d6fa4583aa},
 created = {2017-12-10T18:18:15.151Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2017-12-10T18:20:41.493Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 bibtype = {phdthesis},
 author = {Chehade, Abdallah}
}

@inproceedings{
 title = {Optimal dynamic behavior of adaptive WIP regulation with multiple modes of capacity adjustment},
 type = {inproceedings},
 year = {2014},
 keywords = {Capacity adjustment,Control theory,WIP regulation},
 pages = {168-173},
 volume = {19},
 issue = {C},
 publisher = {Elsevier},
 id = {929fc184-7f26-372e-b186-320a191239a9},
 created = {2016-09-15T22:43:05.000Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2017-03-28T16:04:23.416Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 citation_key = {Chehade2014},
 private_publication = {false},
 abstract = {It is desirable to maintain consistent dynamic behavior of WIP regulation in work systems with multiple modes of capacity adjustment (floaters, overtime, etc.) and different adjustment periods, delays and limits in the various modes. Coordination of these modes is necessary in order to keep optimal dynamic behavior. In this paper, a control-theoretic model of WIP regulation is presented first that accommodates multiple capacity adjustment modes with different adjustment periods (per shift, per day, per week, etc.) and different delays in implementing adjustments. Then an algorithm is presented for adapting WIP adjustment parameters in the presence of capacity adjustment limits and mode priorities so that a specified dynamic performance goal continues to be met. Results of simulations driven by industrial data are used to illustrate the effect of limits and performance goals on dynamic behavior, and conclusions are drawn regarding the effectiveness of adaptive regulation of WIP by coordinating multiple modes of capacity adjustment.},
 bibtype = {inproceedings},
 author = {Chehade, A. and Duffie, N.},
 booktitle = {Procedia CIRP}
}

It is desirable to maintain consistent dynamic behavior of WIP regulation in work systems with multiple modes of capacity adjustment (floaters, overtime, etc.) and different adjustment periods, delays and limits in the various modes. Coordination of these modes is necessary in order to keep optimal dynamic behavior. In this paper, a control-theoretic model of WIP regulation is presented first that accommodates multiple capacity adjustment modes with different adjustment periods (per shift, per day, per week, etc.) and different delays in implementing adjustments. Then an algorithm is presented for adapting WIP adjustment parameters in the presence of capacity adjustment limits and mode priorities so that a specified dynamic performance goal continues to be met. Results of simulations driven by industrial data are used to illustrate the effect of limits and performance goals on dynamic behavior, and conclusions are drawn regarding the effectiveness of adaptive regulation of WIP by coordinating multiple modes of capacity adjustment.

@inproceedings{
 title = {Control theoretical modeling of transient behavior of production planning and control: A review},
 type = {inproceedings},
 year = {2014},
 keywords = {Control theory,Dynamics,Production planning and control},
 pages = {20-25},
 volume = {17},
 issue = {C},
 publisher = {Elsevier},
 id = {03b21765-90aa-393f-9a04-564d852ca420},
 created = {2016-09-15T22:43:05.000Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2017-03-28T16:04:23.416Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 citation_key = {Duffie2014},
 private_publication = {false},
 abstract = {To remain competitive manufacturers need to adapt to increasingly dynamic and turbulent markets-therefore production engineers and business managers need tools for mathematically modeling analyzing and designing agile and changeable production systems that incorporate policies that are robust in the presence of disturbances and mitigate the negative impacts of turbulence in the production environment The spectrum of potential contributions of control theory to understanding the dynamic behavior of production systems in the presence of turbulence is broad In this paper the focus is on classical control theoretical modeling of the transient behavior and fundamental dynamics of production planning and control which generally is considered to include scheduling sequencing loading and controlling Publications in this area in recent years are reviewed, with contributions reported in publications of the CIRP (International Academy for Production Research) receiving particular emphasis.},
 bibtype = {inproceedings},
 author = {Duffie, N. and Chehade, A. and Athavale, A.},
 booktitle = {Procedia CIRP}
}

To remain competitive manufacturers need to adapt to increasingly dynamic and turbulent markets-therefore production engineers and business managers need tools for mathematically modeling analyzing and designing agile and changeable production systems that incorporate policies that are robust in the presence of disturbances and mitigate the negative impacts of turbulence in the production environment The spectrum of potential contributions of control theory to understanding the dynamic behavior of production systems in the presence of turbulence is broad In this paper the focus is on classical control theoretical modeling of the transient behavior and fundamental dynamics of production planning and control which generally is considered to include scheduling sequencing loading and controlling Publications in this area in recent years are reviewed, with contributions reported in publications of the CIRP (International Academy for Production Research) receiving particular emphasis.

@article{
 title = {Dynamics of autonomously acting products and work systems in production and assembly},
 type = {article},
 year = {2012},
 keywords = {Autonomous,Dynamics,Production},
 pages = {267-275},
 volume = {5},
 id = {4c8ffee1-1fc7-3123-89ac-d2390a78edb5},
 created = {2016-09-15T22:43:05.000Z},
 file_attached = {false},
 profile_id = {ceb22f87-8e17-3a73-8ac5-df584316b424},
 last_modified = {2017-03-28T16:04:23.416Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 citation_key = {Jeken2012},
 private_publication = {false},
 abstract = {Autonomous production is characterized by local and autonomous decision making of intelligent logistic objects such as work systems that adjust production rates and parts that decide which products they " want" to become and which orders they will fill. It is important to understand and have confidence in dynamic interactions of these objects and their resulting performance. In this paper the dynamic interaction of autonomous products and work systems is investigated using a hybrid simulation model and a control-theoretic model. Results obtained using both models show that these dynamic interactions can be well behaved and predictable. Through linearized models of continuous input flows at nominal rates, tools of control theory are shown to build confidence in complex system dynamic behavior of interacting autonomous logistics objects when decision-making logic is modeled in a way that makes control-theoretic analyses tractable. ?? 2012 CIRP.},
 bibtype = {article},
 author = {Jeken, O. and Duffie, N. and Windt, K. and Blunck, H. and Chehade, A. and Rekersbrink, H.},
 journal = {CIRP Journal of Manufacturing Science and Technology},
 number = {4}
}

Autonomous production is characterized by local and autonomous decision making of intelligent logistic objects such as work systems that adjust production rates and parts that decide which products they " want" to become and which orders they will fill. It is important to understand and have confidence in dynamic interactions of these objects and their resulting performance. In this paper the dynamic interaction of autonomous products and work systems is investigated using a hybrid simulation model and a control-theoretic model. Results obtained using both models show that these dynamic interactions can be well behaved and predictable. Through linearized models of continuous input flows at nominal rates, tools of control theory are shown to build confidence in complex system dynamic behavior of interacting autonomous logistics objects when decision-making logic is modeled in a way that makes control-theoretic analyses tractable. ?? 2012 CIRP.