Ali Shakouri’s research while affiliated with Purdue University West Lafayette and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (462)


Fig. 1 (a) Different operations performed in a CNC machine to convert a raw material into final product (b) Part fingerprint examples by differences in duration and sequence of operations 2.2. Description of the data collected
Fig. 5 Current fingerprint of all different parts obtained by (a) averaging and (b) up-sampling method; the upsampling approach creates higher resolution fingerprints, but at an additional computational cost.
Comparison of cycle time (i.e., average time to make a part and average delays between two units)
Part Fingerprinting Based Productivity Monitoring of CNC Machines with Low-Cost Current Sensors
  • Preprint
  • File available

November 2024

·

16 Reads

Ajanta Saha

·

·

Jabir Jahangir

·

[...]

·

Ali Shakouri

Digital transformation of manufacturing industry, Smart Manufacturing, leverages continuous measurement of machines on the shop floor to make effective decisions and improve productivity metrics such as machine uptime (MMU) and overall equipment efficiency (OEE). However, despite the declining sensor cost, the initial financial and technological skill requirements of digital transformation pose significant barriers for the overwhelming majority (90%) of the manufacturers who are classed as small and medium enterprises (SMEs). To lower this barrier, here we demonstrate an inexpensive (~$40 per machine), data-efficient solution that extracts part-level productivity metrics of a CNC machine from its total current consumption alone. We introduce the concept of a part’s “fingerprint” and develop a set of methods that allows one to extract the fingerprints and utilize them to monitor each individual manufactured part and their cycle times. Testing on actual production data of over three months in a part-counting task, the algorithms show a good match (96.2% overall accuracy) with manually logged production data is achieved. The presented fingerprint framework is general: it can be extended to multi-sensors, and multi-modal analytics. We expect that such a simple, yet cost-effective, solution will be accessible for a wide range of discrete manufacturers, facilitating the beginning of their digital transformation journey.

Download

Adversarial Domain Adaptation for Metal Cutting Sound Detection: Leveraging Abundant Lab Data for Scarce Industry Data

October 2024

·

9 Reads

Cutting state monitoring in the milling process is crucial for improving manufacturing efficiency and tool life. Cutting sound detection using machine learning (ML) models, inspired by experienced machinists, can be employed as a cost-effective and non-intrusive monitoring method in a complex manufacturing environment. However, labeling industry data for training is costly and time-consuming. Moreover, industry data is often scarce. In this study, we propose a novel adversarial domain adaptation (DA) approach to leverage abundant lab data to learn from scarce industry data, both labeled, for training a cutting-sound detection model. Rather than adapting the features from separate domains directly, we project them first into two separate latent spaces that jointly work as the feature space for learning domain-independent representations. We also analyze two different mechanisms for adversarial learning where the discriminator works as an adversary and a critic in separate settings, enabling our model to learn expressive domain-invariant and domain-ingrained features, respectively. We collected cutting sound data from multiple sensors in different locations, prepared datasets from lab and industry domain, and evaluated our learning models on them. Experiments showed that our models outperformed the multi-layer perceptron based vanilla domain adaptation models in labeling tasks on the curated datasets, achieving near 92%, 82% and 85% accuracy respectively for three different sensors installed in industry settings.



Efficient Microscopic Image Instance Segmentation for Food Crystal Quality Control

September 2024

·

1 Read

This paper is directed towards the food crystal quality control area for manufacturing, focusing on efficiently predicting food crystal counts and size distributions. Previously, manufacturers used the manual counting method on microscopic images of food liquid products, which requires substantial human effort and suffers from inconsistency issues. Food crystal segmentation is a challenging problem due to the diverse shapes of crystals and their surrounding hard mimics. To address this challenge, we propose an efficient instance segmentation method based on object detection. Experimental results show that the predicted crystal counting accuracy of our method is comparable with existing segmentation methods, while being five times faster. Based on our experiments, we also define objective criteria for separating hard mimics and food crystals, which could benefit manual annotation tasks on similar dataset.




Enhanced imaging of electronic hot spots using quantum squeezed light

June 2024

·

12 Reads

·

2 Citations

Detecting electronic hot spots is important for understanding the heat dissipation and thermal management of electronic and semiconductor devices. Optical thermoreflective imaging is being used to perform precise temporal and spatial imaging of heat on wires and semiconductor materials. We apply quantum squeezed light to perform thermoreflective imaging on micro-wires, surpassing the shot-noise limit of classical approaches. We obtain a far-field temperature sensing accuracy of 42 mK after 50 ms of averaging and show that a 256 × 256 pixel image can be constructed with such sensitivity in 10 min. We can further obtain single-shot temperature sensing of 1.6 K after only 10 μ s of averaging, enabling a dynamical study of heat dissipation. Not only do the quantum images provide accurate spatiotemporal information about heat distribution but also the measure of quantum correlation provides additional information, inaccessible by classical techniques, which can lead to a better understanding of the dynamics. We apply the technique to both aluminum and niobium microwires and discuss the applications of the technique in studying electron dynamics at low temperatures.





Citations (53)


... Therefore, highly displaced squeezed light will accomplish both enhanced sensitivity and improved precision. A recent result of super-resolution imaging using highly displaced squeezed light attests to the critical usage of such a quantum state of light [22], and enhancements through the use of highly displaced squeezed light have been demonstrated in several spectroscopy and microscopy applications [23][24][25][26][27][28][29][30][31]. Additionally, one can use a high-displacement amplitude-squeezed light to control an optically driven energy transition process [32][33][34]. ...

Reference:

Fundamental limits to the generation of highly displaced bright squeezed light using linear optics and parametric amplifiers
Enhanced imaging of electronic hot spots using quantum squeezed light

... Recently, it has been shown that quadrature squeezed states of light can be used to obtain a true quantum advantage when compared to the best classical approaches, in very different contexts, improving the detection of gravitational waves 18,19 and single-cell imaging. 5,8,20 Bright intensity squeezed states of light 16,[21][22][23][24][25] are also valuable resources when considering achieving quantum advantage. This is because these optical states can be generated at relatively high powers (milliwatt-level powers) and, thus, can compete with classical sensing methods applied to certain applications, 24,26 provided that optical losses are relatively small. ...

Quantum Sensing of Thermoreflectivity in Electronics
  • Citing Article
  • April 2023

Physical Review Applied

... The part counts obtained through the two different metrics may exhibit discrepancies depending on the choice of methods or metrics. In such instances, ensemble averaging can be used to give a more stable part count [28]. Given that we use two sets of metrics in this work, the ultimate part count is determined as the weighted average of the two individual counts, expressed in Eq. (8), ...

A new paradigm of reliable sensing with field-deployed electrochemical sensors integrating data redundancy and source credibility

... Among the various oxide semiconductor candidates, In 2 O 3 has been extensively studied owing to its exceptional electron conductivity, even in ultrathin bodies of just a few nanometers ( 3 nm). [12][13][14][15][16][17][18][19][20][21][22][23] Most studies on the In 2 O 3 thin-film transistors (TFTs) employ high-k-gate dielectrics, such as HfO 2 , since high-k materials are essential for reducing the natural length of devices. [17][18][19][20][21][22][23] However, the factors that could potentially degrade benefits of high-k dielectrics in In 2 O 3 TFTs have not been thoroughly investigated. ...

Transient Thermal and Electrical Co-Optimization of BEOL Top-Gated ALD In2_{\text{2}}O3_{\text{3}} FETs Toward Monolithic 3-D Integration
  • Citing Article
  • April 2023

IEEE Transactions on Electron Devices

... In the study [6], anomaly detection in smart factories was explored using machine learning techniques, particularly addressing challenges related to diverse sensors and their transferability across different production lines. A combination of machine learning algorithms, including neural networks and transfer learning, was employed to facilitate knowledge transfer between production lines. ...

Anomaly Detection and Inter-Sensor Transfer Learning on Smart Manufacturing Datasets

Sensors

... Due to the high production speed and throughput requirements, quality inspection is performed directly on the moving sample. Hence, measurement systems combining fast measurements and high lateral resolution are required [8]. Optical measurement systems enable contactless displacement measurements with high measurement rates and are of high interest in these applications [9]. ...

Real‐Time Metrology for Roll‐To‐Roll and Advanced Inline Manufacturing: A Review

... Metal oxides have enabled many emerging applications in advanced electronics such as CMOS back-end-of-line (BEOL)compatible logic and memory components (Datta et al., 2019;Charnas et al., 2023;Kim et al., 2023). For instance, In-based oxides have been actively explored as channel material for BEOLcompatible transistors due to its high electron mobility, large area uniformity, excellent conformity on complex structure, and lowtemperature processability (Samanta et al., 2020;Han et al., 2021;Si et al., 2022;Zhang et al., 2022;Zheng et al., 2022;Liao et al., 2023;Zhang et al., 2023). Hf-based oxides are also currently used as highk dielectric in Si-based logic transistors and storage capacitors in dynamic random-access memory (DRAM) arising from its relatively high permittivity value (10-25), suitable band offsets to Si, sufficiently large bandgap (E g ), and high thermal stability (Wilk et al., 2001;Kim et al., 2013;Wang B. et al., 2018). ...

Alleviation of Self-Heating Effect in Top-Gated Ultrathin In2_{\text{2}}O3_{\text{3}} FETs Using a Thermal Adhesion Layer
  • Citing Article
  • January 2022

IEEE Transactions on Electron Devices

... One of the major restrictions of machine learning applications using customized databases is the cost of human labor. In the previous papers [3,4,5], it is demonstrated through experiments that the correlation between thin-film nitrate sensor performance and surface texture exists. In the previous papers, several methods for extracting texture features from sensor images are explored, repeated cross-validation and a hyperparameter auto-tuning method are performed, and several machine learning models are built to improve prediction accuracy. ...

Improvements to color image and machine learning based thin-film nitrate sensor performance prediction: New texture features, repeated cross-validation, and auto-tuning of hyperparameters
  • Citing Article
  • January 2022

Electronic Imaging

... The tighter distribution of EMF values has important ramifications for using cost-effective manufacturing processes to produce high-quality sensors at low cost. The production of disposable ISEs can be scaled up using roll-toroll manufacturing methods; 55 in this regard, we have determined the feasibility of depositing GrNP dispersions onto roll substrates by in-line spray coating, prior to membrane casting ( Figure S6). Graphite is also well known for its excellent thermal conductivity, and graphene-based dispersions and composites have been widely studied as media for heat dissipation and exchange. ...

Roll to Roll Manufacturing and In-Line Imaging and Characterization of Functional Films
  • Citing Conference Paper
  • September 2022

... The weights can be chosen to be constants; for instance, = = gives the simple average. The weights could also be tuned with time as a function of how reliably each metric accurately identifies the unknown parts [28][29][30]. ...

Embrace the Imperfection: How Intrinsic Variability of Roll-to-Roll Manufactured Environmental Sensors Enable Self-Calibrating, High-Precision Quorum Sensing
  • Citing Conference Paper
  • September 2022