An Adaptable and Unsupervised TinyML Anomaly Detection System for Extreme Industrial Environments

Antonini, Mattia; Pincheira, Miguel; Vecchio, Massimo; Antonelli, Fabio

doi:10.3390/s23042344

Cited by 28 publications

(13 citation statements)

References 41 publications

(71 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The key to their adoption as a standard means of monitoring is to lower their price while keeping them highly accurate. This is an active field of research with many other approaches beyond camera-based traps [25] and many other applications [47]. In this work, we aim to overcome some important technical limitations of vision-based systems namely: manual annotation and insect congestion on the bucket.…”

Section: Discussionmentioning

confidence: 99%

Image-Based Insect Counting Embedded in E-traps that Learn Without Manual Image Annotation and Self-Dispose Captured Insects

Saradopoulos¹,

Potamitis²,

Konstantaras³

et al. 2023

Preprint

View full text Add to dashboard Cite

This study describes the development of an image-based insect trap diverging from the plug-in camera insect trap paradigm. In short, a) it does not require manual annotation of images to learn how to count targeted pests and, b) it self-disposes the captured insects, and therefore is suitable for long-term deployment. The device consists of an imaging sensor integrated with Raspberry Pi microcontroller units with embedded deep learning algorithms that count agricultural pests inside a pheromone-based funnel trap. The device also receives commands from the server which configures its operation while an embedded servomotor can automatically rotate the detached bottom of the bucket to dispose of hydrated insects as they begin to pile up. Therefore, it completely overcomes a major limitation of camera-based insect traps: the inevitable overlap and occlusion caused by the decay and layering of insects during long-term operation, thus extending the autonomous operational capability. We study cases that are underrepresented in literature such as counting in situations of congestion and significant debris using crowd counting algorithms encountered in human surveillance. Finally, we perform comparative analysis of the results from different deep-learning approaches (YOLO7/8, crowd counting, deep learning regression) and we open-source the code and a large database of Lepidopteran plant pests.

show abstract

Section: Discussionmentioning

confidence: 99%

Image-Based Insect Counting Embedded in E-traps that Learn Without Manual Image Annotation and Self-Dispose Captured Insects

Saradopoulos¹,

Potamitis²,

Konstantaras³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…The need for efficient utilization of real-time integrated systems was emphasized for monitoring manufacturing processes and detecting real-time anomalies [16]. Antonini, et al [29] proposed an end-to-end adaptive and configurable anomaly detection system using the Tiny-MLOps methodology. They designed a TinyML pipeline considering industrial environments from data sampling, feature extraction, model training, and inference to anomaly event notifications.…”

Section: Related Workmentioning

confidence: 99%

“…To use the anomaly detection model in a real-world manufacturing environment, it is necessary to consider optimizing the model for optimal performance, such as hyperparameter estimation and determining the threshold for abnormal sections. The convenience of redeploying the anomaly detection model, learned from new data, should also be considered [11,29]. Traditionally, real-time service models are operated by data operations organizations, which are well suited to re-modify models or operate realtime models (see Fig.…”

Section: E Service Model Developmentmentioning

confidence: 99%

Enhanced Anomaly Detection in Manufacturing Processes Through Hybrid Deep Learning Techniques

Lee,

Kim,

Kim

2023

IEEE Access

View full text Add to dashboard Cite

Smart factory systems have been introduced to prevent the decline in overall equipment effectiveness caused by the presence of defects within factory manufacturing equipment. In this context, it is crucial to predict process downtime using manufacturing equipment data and take preemptive actions. However, anomaly detection models for preemptive actions have limitations in labeling anomaly alarms. Moreover, from real-time data collection to model deployment, it needs to consider the model-based service implementation for stakeholders. Our research develops a hybrid deep learning-based anomaly detection model that does not require data labeling. The advantage of this model stems from its ability to identify anomalous patterns via the reconstruction of the sequential progression inherent in the data. According to our experimental findings, the proposed hybrid model demonstrated superior efficacy compared to alternative anomaly detection algorithms. By preemptively predicting downtime within the manufacturing process, it contributes to enhanced production efficiency. Furthermore, we develop an anomaly detection system based on a real-time service framework from the data collection step to improve the activation of smart factories. This research contributes to the literature on the monitoring and management of anomaly detection in smart factories. It also has practical implications for the manufacturing industry by recommending efficiency measures for smart factories to reduce downtime in manufacturing processes and improve product quality.

show abstract

“…There are also existing research papers [ 20 , 21 , 22 , 23 ] that explore the topics of TinyML and IoT in a manner similar to this paper.…”

Section: Related Workmentioning

confidence: 99%

Study of the Impact of Data Compression on the Energy Consumption Required for Data Transmission in a Microcontroller-Based System

Piątkowski,

Puślecki,

Walkowiak

2023

Sensors

View full text Add to dashboard Cite

As the number of Internet of Things (IoT) devices continues to rise dramatically each day, the data generated and transmitted by them follow similar trends. Given that a significant portion of these embedded devices operate on battery power, energy conservation becomes a crucial factor in their design. This paper aims to investigate the impact of data compression on the energy consumption required for data transmission. To achieve this goal, we conduct a comprehensive study using various transmission modules in a severely resource-limited microcontroller-based system designed for battery power. Our study evaluates the performance of several compression algorithms, conducting a detailed analysis of computational and memory complexity, along with performance metrics. The primary finding of our study is that by carefully selecting an algorithm for compressing different types of data before transmission, a significant amount of energy can be saved. Moreover, our investigation demonstrates that for a battery-powered embedded device transmitting sensor data based on the STM32F411CE microcontroller, the recommended transmission module is the nRF24L01+ board, as it requires the least amount of energy to transmit one byte of data. This module is most effective when combined with the LZ78 algorithm for optimal energy and time efficiency. In the case of image data, our findings indicate that the use of the JPEG algorithm for compression yields the best results. Overall, our research underscores the importance of selecting appropriate compression algorithms tailored to specific data types, contributing to enhanced energy efficiency in IoT devices.

show abstract

An Adaptable and Unsupervised TinyML Anomaly Detection System for Extreme Industrial Environments

Cited by 28 publications

References 41 publications

Image-Based Insect Counting Embedded in E-traps that Learn Without Manual Image Annotation and Self-Dispose Captured Insects

Image-Based Insect Counting Embedded in E-traps that Learn Without Manual Image Annotation and Self-Dispose Captured Insects

Enhanced Anomaly Detection in Manufacturing Processes Through Hybrid Deep Learning Techniques

Study of the Impact of Data Compression on the Energy Consumption Required for Data Transmission in a Microcontroller-Based System

Contact Info

Product

Resources

About