This study has successfully integrated an advanced laser drilling machine (LDM) and a photo-elastic (PE) system to measure the residual stress of acrylic substrates after drilling manufacturing. The full-field distribution of fringe denotes the strain concentration of the specimen after hole drilling, and we collect massive data based on a self-built system of measuring the photo-elasticity effect of the samples. The work uses the self-developed PE system, contains parameters during manufacturing, and yields quantitative sensor fusion results promising the preventative maintenance of the LDM. The diagnostic maintenance system can achieve this through signal processing and artificial intelligence algorithms. In particular, edge computing architectures can effectively diagnose faults in real-time. The appearance of contours caused by machining is traditionally measured with a surface profile meter, but now we are experimenting with measuring shapes and residual stresses through photo-elasticity. This paper introduces the sapphire substrate for the emerging material and compares its properties with those of acrylic specimens. We have constructed a 5G experimental field and verified the developed architecture and methods, and well-developed technologies have been promoted in the industry. The measured results and data can cooperate with upcoming 5G communication and utilize the advantages of enhanced mobile broadband (eMBB), massive machine type transmissions (mMTC), and ultra-reliable and low latency communications (URLLC). This work applies the domain knowledge of PE and the advantage of 5G technology, providing a diagnostic maintenance system for the laser drilling machine.
{"title":"In-situ Measurement of Residual Stress and Thermal Deformation of Substrate after Laser Drilling","authors":"Cheng-Lun Kan, Han-San Xie, Chao-Ching Ho, Ching-Yuan Chang","doi":"10.1109/MESA55290.2022.10004403","DOIUrl":"https://doi.org/10.1109/MESA55290.2022.10004403","url":null,"abstract":"This study has successfully integrated an advanced laser drilling machine (LDM) and a photo-elastic (PE) system to measure the residual stress of acrylic substrates after drilling manufacturing. The full-field distribution of fringe denotes the strain concentration of the specimen after hole drilling, and we collect massive data based on a self-built system of measuring the photo-elasticity effect of the samples. The work uses the self-developed PE system, contains parameters during manufacturing, and yields quantitative sensor fusion results promising the preventative maintenance of the LDM. The diagnostic maintenance system can achieve this through signal processing and artificial intelligence algorithms. In particular, edge computing architectures can effectively diagnose faults in real-time. The appearance of contours caused by machining is traditionally measured with a surface profile meter, but now we are experimenting with measuring shapes and residual stresses through photo-elasticity. This paper introduces the sapphire substrate for the emerging material and compares its properties with those of acrylic specimens. We have constructed a 5G experimental field and verified the developed architecture and methods, and well-developed technologies have been promoted in the industry. The measured results and data can cooperate with upcoming 5G communication and utilize the advantages of enhanced mobile broadband (eMBB), massive machine type transmissions (mMTC), and ultra-reliable and low latency communications (URLLC). This work applies the domain knowledge of PE and the advantage of 5G technology, providing a diagnostic maintenance system for the laser drilling machine.","PeriodicalId":410029,"journal":{"name":"2022 18th IEEE/ASME International Conference on Mechatronic and Embedded Systems and Applications (MESA)","volume":"84 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121247159","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-28DOI: 10.1109/mesa55290.2022.10004453
{"title":"MESA2022 Book of Abstracts","authors":"","doi":"10.1109/mesa55290.2022.10004453","DOIUrl":"https://doi.org/10.1109/mesa55290.2022.10004453","url":null,"abstract":"","PeriodicalId":410029,"journal":{"name":"2022 18th IEEE/ASME International Conference on Mechatronic and Embedded Systems and Applications (MESA)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134078351","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
For indoor security robots, face recognition is an important ability. However, face recognition is suffered from the limitation by environment uncertainties, the factors including perceptual aliasing, occlusion, illumination changes and significant viewpoint changes. These uncertainties will affect the recognition accuracy and processing time, which will cause the security concerns. This paper proposes a convolutional neural networks-based face recognition system for the mobile robots to perform visual perception and control tasks. The trained model proposed in this paper (i.e., FaceNet) is compared and tested against two different algorithms, VGGNet and AlexNet. With image streaming, images are transferred to the cloud for GPU computing. In addition, the Cartographer SLAM algorithms is used for the indoor simultaneous localization and mapping. The experimental results show that the accuracy of proposed face recognition system under the conditions of four different illumination is 88%, which proves the feasibility of the method. Through the cloud GPU, the local computation and processing time can be reduced. The established mobile robot system can perform the indoor navigating and simultaneous localization and mapping.
{"title":"Deep Learning based Face Recognition for Security Robot","authors":"Min-Fan Ricky Lee, Yun-Min Huang, Jiaqian Sun, Xuerong Chen, Tingting Huang","doi":"10.1109/MESA55290.2022.10004482","DOIUrl":"https://doi.org/10.1109/MESA55290.2022.10004482","url":null,"abstract":"For indoor security robots, face recognition is an important ability. However, face recognition is suffered from the limitation by environment uncertainties, the factors including perceptual aliasing, occlusion, illumination changes and significant viewpoint changes. These uncertainties will affect the recognition accuracy and processing time, which will cause the security concerns. This paper proposes a convolutional neural networks-based face recognition system for the mobile robots to perform visual perception and control tasks. The trained model proposed in this paper (i.e., FaceNet) is compared and tested against two different algorithms, VGGNet and AlexNet. With image streaming, images are transferred to the cloud for GPU computing. In addition, the Cartographer SLAM algorithms is used for the indoor simultaneous localization and mapping. The experimental results show that the accuracy of proposed face recognition system under the conditions of four different illumination is 88%, which proves the feasibility of the method. Through the cloud GPU, the local computation and processing time can be reduced. The established mobile robot system can perform the indoor navigating and simultaneous localization and mapping.","PeriodicalId":410029,"journal":{"name":"2022 18th IEEE/ASME International Conference on Mechatronic and Embedded Systems and Applications (MESA)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130994128","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-28DOI: 10.1109/MESA55290.2022.10004461
Zheng-Jie Huang, Wei-Hao Lu, Brijesh Patel, Po-Yan Chiu, Tz-Yu Yang, Hao Jian Tong, V. Bučinskas, M. Greitans, P. Lin
In this era of automation, image processing is an indispensable part of computer vision. Many computer vision approaches in the industry depend on a relatively bright environment. Under low light source conditions, the distribution of image information is too concentrated in specific intensity ranges due to the color factor of the subject itself, resulting in noise and contrast loss. Enhancing contrast is a crucial step in improving the quality of the image and showing visible details. This study proposes a method based on a convolutional neural network (CNN), using the pixel difference between paired images, called a motion matrix, as an annotation for low-contrast images. The image's motion vector is predicted after the neural network model has been trained to produce the low-contrast enhanced image. Then, the proposed model is compared with the Low-Light image Enhancement (LLNet), Multi-Scale Retinex Color Restoration (MSRCR), and Fuzzy Automatic Cluster Enhancement (FACE) approaches. The effectiveness of the proposed method was further evaluated by comparing several quality indicators, including peak signal-to-noise ratio, structural similarity, root-mean-square-error, root-mean-square-contrast and computation time efficiency.
{"title":"Convolutional Neural Network-based Image Restoration (CNNIR)","authors":"Zheng-Jie Huang, Wei-Hao Lu, Brijesh Patel, Po-Yan Chiu, Tz-Yu Yang, Hao Jian Tong, V. Bučinskas, M. Greitans, P. Lin","doi":"10.1109/MESA55290.2022.10004461","DOIUrl":"https://doi.org/10.1109/MESA55290.2022.10004461","url":null,"abstract":"In this era of automation, image processing is an indispensable part of computer vision. Many computer vision approaches in the industry depend on a relatively bright environment. Under low light source conditions, the distribution of image information is too concentrated in specific intensity ranges due to the color factor of the subject itself, resulting in noise and contrast loss. Enhancing contrast is a crucial step in improving the quality of the image and showing visible details. This study proposes a method based on a convolutional neural network (CNN), using the pixel difference between paired images, called a motion matrix, as an annotation for low-contrast images. The image's motion vector is predicted after the neural network model has been trained to produce the low-contrast enhanced image. Then, the proposed model is compared with the Low-Light image Enhancement (LLNet), Multi-Scale Retinex Color Restoration (MSRCR), and Fuzzy Automatic Cluster Enhancement (FACE) approaches. The effectiveness of the proposed method was further evaluated by comparing several quality indicators, including peak signal-to-noise ratio, structural similarity, root-mean-square-error, root-mean-square-contrast and computation time efficiency.","PeriodicalId":410029,"journal":{"name":"2022 18th IEEE/ASME International Conference on Mechatronic and Embedded Systems and Applications (MESA)","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132742133","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-28DOI: 10.1109/MESA55290.2022.10004459
Ryan H. Kim, H. Min
In the status quo, traffic control systems operate on predetermined patterns and instructions devised from past data. While this method functions effectively for traffic under normal conditions, it becomes heavily congested and inefficient during instances of high traffic, which leads to a multitude of temporal, economic, health, and environmental harms. However, by combining traditional traffic controllers with modern technologies such as Internet of Things devices and computer vision, these issues can be effectively addressed. This research presents a novel, affordable Artificial Intelligence of Things traffic control system that enables accurate real-time vehicle detection and signal control. This work is split into two sections: (1) an AIoT physical system that can scan traffic conditions in real-time and (2) a realistic traffic simulator with a custom optimization algorithm. Combined, this research provides up to 35% greater throughput, 50% reduced waiting time, and 50% reduction in greenhouse gas emission reductions in comparison to nonoptimized algorithms used in the status quo. The implementation of this work leads to various temporal, economic, environmental, and health benefits; in addition to providing comparable emission reduction as the complete replacement of all internal combustion engine vehicles with battery electric vehicles, while significantly reducing vehicle travel time, systems installation time, and cost.
{"title":"ARCTO: AIoT System for Reducing Carbon Emissions Using Traffic Optimization","authors":"Ryan H. Kim, H. Min","doi":"10.1109/MESA55290.2022.10004459","DOIUrl":"https://doi.org/10.1109/MESA55290.2022.10004459","url":null,"abstract":"In the status quo, traffic control systems operate on predetermined patterns and instructions devised from past data. While this method functions effectively for traffic under normal conditions, it becomes heavily congested and inefficient during instances of high traffic, which leads to a multitude of temporal, economic, health, and environmental harms. However, by combining traditional traffic controllers with modern technologies such as Internet of Things devices and computer vision, these issues can be effectively addressed. This research presents a novel, affordable Artificial Intelligence of Things traffic control system that enables accurate real-time vehicle detection and signal control. This work is split into two sections: (1) an AIoT physical system that can scan traffic conditions in real-time and (2) a realistic traffic simulator with a custom optimization algorithm. Combined, this research provides up to 35% greater throughput, 50% reduced waiting time, and 50% reduction in greenhouse gas emission reductions in comparison to nonoptimized algorithms used in the status quo. The implementation of this work leads to various temporal, economic, environmental, and health benefits; in addition to providing comparable emission reduction as the complete replacement of all internal combustion engine vehicles with battery electric vehicles, while significantly reducing vehicle travel time, systems installation time, and cost.","PeriodicalId":410029,"journal":{"name":"2022 18th IEEE/ASME International Conference on Mechatronic and Embedded Systems and Applications (MESA)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115773104","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-28DOI: 10.1109/MESA55290.2022.10004470
Yuxuan Cao, Zhen Zhang
Laser induced forward transfer (LIFT) technique is widely considered as an efficient way to transfer print and assemble micro-sized semiconductor devices between wafers at a high speed. In this paper we propose a novel noncontact selective laser transfer printing and assembly method based on thermal releasing tape and infrared lasers. Thermal analysis and simulation are conducted to illustrate the mechanism of transfer printing. The experiment results of microchip arrays printing demonstrate that the proposed method has high precision, full selectivity, and compatibility of flexible transfer printing. The proposed method shows great potential to become a very competitive method for transfer printing of micro-devices.
{"title":"Noncontact selective laser transfer printing and assembly of micro-sized semiconductor devices","authors":"Yuxuan Cao, Zhen Zhang","doi":"10.1109/MESA55290.2022.10004470","DOIUrl":"https://doi.org/10.1109/MESA55290.2022.10004470","url":null,"abstract":"Laser induced forward transfer (LIFT) technique is widely considered as an efficient way to transfer print and assemble micro-sized semiconductor devices between wafers at a high speed. In this paper we propose a novel noncontact selective laser transfer printing and assembly method based on thermal releasing tape and infrared lasers. Thermal analysis and simulation are conducted to illustrate the mechanism of transfer printing. The experiment results of microchip arrays printing demonstrate that the proposed method has high precision, full selectivity, and compatibility of flexible transfer printing. The proposed method shows great potential to become a very competitive method for transfer printing of micro-devices.","PeriodicalId":410029,"journal":{"name":"2022 18th IEEE/ASME International Conference on Mechatronic and Embedded Systems and Applications (MESA)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129534715","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Along with the aging society, the elderly population increases. Most non-disabled elderly prefer to age in their comfortable homes. To support such home care for the elderly, continuous real-time monitoring of all this and early warning in the event of an unexpected event are beneficial. Current monitoring systems, such as wearable sensors or webcams, could monitor the activity of elderly people and support their independent living. However, it malfunctions when the elderly do not wear wearable sensors; the webcam has privacy concerns. The study proposes a novel intelligent system to monitor the daily life of the elderly and to notify anomalies in real time. Millimeter-wave (mmWave) radar, machine learning, and self-comparison method were adopted to implement such a system. A data-driven self-comparison scheme is proposed to reduce false alarms. Clinical data from 73 seniors (58 males; mean age and standard deviation 71.7 ± 7.4 years; 15 females; 70.8 ± 7.8 years) were collected in the hospital for the training of the sleep prediction model. Five older solidary volunteers attended the data collection at their home for indoor tracking and sleep monitoring. The experimental results revealed that the proposed system could achieve a false alarm rate below 5%. The findings of the study may serve as a guide for the research and development of non-invasive sensing systems for the care of elderly adults at home.
{"title":"Detecting Anomalies of Daily Living of the Elderly Using Radar and Self-Comparison Method","authors":"Fu-Kuei Chen, You-Kwang Wang, Hsin-Piao Lin, Chien-Yu Chen, Shu-Ming Yeh, Ching-Yu Wang","doi":"10.1109/MESA55290.2022.10004481","DOIUrl":"https://doi.org/10.1109/MESA55290.2022.10004481","url":null,"abstract":"Along with the aging society, the elderly population increases. Most non-disabled elderly prefer to age in their comfortable homes. To support such home care for the elderly, continuous real-time monitoring of all this and early warning in the event of an unexpected event are beneficial. Current monitoring systems, such as wearable sensors or webcams, could monitor the activity of elderly people and support their independent living. However, it malfunctions when the elderly do not wear wearable sensors; the webcam has privacy concerns. The study proposes a novel intelligent system to monitor the daily life of the elderly and to notify anomalies in real time. Millimeter-wave (mmWave) radar, machine learning, and self-comparison method were adopted to implement such a system. A data-driven self-comparison scheme is proposed to reduce false alarms. Clinical data from 73 seniors (58 males; mean age and standard deviation 71.7 ± 7.4 years; 15 females; 70.8 ± 7.8 years) were collected in the hospital for the training of the sleep prediction model. Five older solidary volunteers attended the data collection at their home for indoor tracking and sleep monitoring. The experimental results revealed that the proposed system could achieve a false alarm rate below 5%. The findings of the study may serve as a guide for the research and development of non-invasive sensing systems for the care of elderly adults at home.","PeriodicalId":410029,"journal":{"name":"2022 18th IEEE/ASME International Conference on Mechatronic and Embedded Systems and Applications (MESA)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130672929","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Emission quantification from small uncrewed aircraft systems (sUAS) are of high interest due to their low cost and flexibility to measure site level emissions. However, understanding the emission performance of the current/future methodologies, typically, requires many field experiments in a variety of weather conditions, affecting the repeatability and making it costly to pursue. In this work, we utilize MOABS/DT to explore single and multi-sUAS emission quantification methods for estimating emissions in three source release scenarios and plume meandering conditions.
{"title":"Single and Multi-sUAS Based Emission Quantification Performance Assessment Using MOABS/DT: A Simulation Case Study","authors":"Derek Hollenbeck, Demitrius Zulevicl, Yangquan Chen","doi":"10.1109/MESA55290.2022.10004398","DOIUrl":"https://doi.org/10.1109/MESA55290.2022.10004398","url":null,"abstract":"Emission quantification from small uncrewed aircraft systems (sUAS) are of high interest due to their low cost and flexibility to measure site level emissions. However, understanding the emission performance of the current/future methodologies, typically, requires many field experiments in a variety of weather conditions, affecting the repeatability and making it costly to pursue. In this work, we utilize MOABS/DT to explore single and multi-sUAS emission quantification methods for estimating emissions in three source release scenarios and plume meandering conditions.","PeriodicalId":410029,"journal":{"name":"2022 18th IEEE/ASME International Conference on Mechatronic and Embedded Systems and Applications (MESA)","volume":"104 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126342177","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-28DOI: 10.1109/MESA55290.2022.10004454
Y. Yeh, Wei-Cheng Wang, Rongping Chen
To aim at the collision-free navigation framework for multi-robot systems in dynamic environment, this work develops a two-layer methodology to implement the obstacle avoidance for multiple robots. A global planner is introduced to construct the global plan at the high layer. Combining the artificial potential field with the pure-pursuit algorithm, a low-layer local planner is designed to generate the control commands for tracking the waypoints obtained from the global plan. Moreover, the rolling windows method, the obstacle filter, and the multi-robot coordination strategy are also introduced to enhance the robustness of the proposed approach implemented on practical robots. Both simulations and experimental results are presented to verify the feasibility of the proposed method.
{"title":"Collision-Free Navigation for Multiple Robots in Dynamic Environment","authors":"Y. Yeh, Wei-Cheng Wang, Rongping Chen","doi":"10.1109/MESA55290.2022.10004454","DOIUrl":"https://doi.org/10.1109/MESA55290.2022.10004454","url":null,"abstract":"To aim at the collision-free navigation framework for multi-robot systems in dynamic environment, this work develops a two-layer methodology to implement the obstacle avoidance for multiple robots. A global planner is introduced to construct the global plan at the high layer. Combining the artificial potential field with the pure-pursuit algorithm, a low-layer local planner is designed to generate the control commands for tracking the waypoints obtained from the global plan. Moreover, the rolling windows method, the obstacle filter, and the multi-robot coordination strategy are also introduced to enhance the robustness of the proposed approach implemented on practical robots. Both simulations and experimental results are presented to verify the feasibility of the proposed method.","PeriodicalId":410029,"journal":{"name":"2022 18th IEEE/ASME International Conference on Mechatronic and Embedded Systems and Applications (MESA)","volume":"47 3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125150146","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-28DOI: 10.1109/MESA55290.2022.10004448
Di An, Yangquan Chen
It is widely known that the remaining capacity of any lithium polymer (Li-Po) rechargeable battery is hard to know precisely in real time. Battery management systems (BMS) are used to precisely monitor battery health including state of charge (SOC) and the remaining capacity. But, BMS is usually limited by its size, power consumption, and compatibility, which could potentially have a negative impact on the battery powered mission such as long distance drone flights. Therefore, in this study, we present a new approach for (Li-Po) battery capacity sensing using a miniature millimeter Wave radar array in real-time. We assessed our contactless battery capacity sensing method with a classifier algorithm using labeled data collected from real battery discharging load circuit experiments. According to the results, our technique achieved 98.8% classification accuracy across eight different battery capacity levels. The machine learning algorithm is computationally light and easily implementable on edge computing platforms such as the Raspberry Pi. This work confirms that it is feasible to sense the real-time remaining capacity of Li-Po batteries that can lead to a capacity-aware cognitive battery management system.
{"title":"A Miniature Millimeter-Wave Radar Based Contactless Lithium Polymer Battery Capacity Sensing with Edge Artificial Intelligence","authors":"Di An, Yangquan Chen","doi":"10.1109/MESA55290.2022.10004448","DOIUrl":"https://doi.org/10.1109/MESA55290.2022.10004448","url":null,"abstract":"It is widely known that the remaining capacity of any lithium polymer (Li-Po) rechargeable battery is hard to know precisely in real time. Battery management systems (BMS) are used to precisely monitor battery health including state of charge (SOC) and the remaining capacity. But, BMS is usually limited by its size, power consumption, and compatibility, which could potentially have a negative impact on the battery powered mission such as long distance drone flights. Therefore, in this study, we present a new approach for (Li-Po) battery capacity sensing using a miniature millimeter Wave radar array in real-time. We assessed our contactless battery capacity sensing method with a classifier algorithm using labeled data collected from real battery discharging load circuit experiments. According to the results, our technique achieved 98.8% classification accuracy across eight different battery capacity levels. The machine learning algorithm is computationally light and easily implementable on edge computing platforms such as the Raspberry Pi. This work confirms that it is feasible to sense the real-time remaining capacity of Li-Po batteries that can lead to a capacity-aware cognitive battery management system.","PeriodicalId":410029,"journal":{"name":"2022 18th IEEE/ASME International Conference on Mechatronic and Embedded Systems and Applications (MESA)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123177475","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：