Pub Date : 2018-11-01DOI: 10.1109/M2VIP.2018.8600822
T. Jacobs
In the last decades, omnidirectional mobile robot platforms have entered many areas of industrial automation and daily life. While platforms with universal and Mecanum wheels are quite common, standard wheels have the advantage of robustness against uneven ground and slippage. This paper gives an overview on recent developments in the design of omnidirectional undercarriages with standard wheels and categorizes mechanical designs according to their holonomy. Furthermore, requirements for different applications are discussed and simple evaluation criteria are proposed that can be used to compare designs.
{"title":"Omnidirectional Robot Undercarriages with Standard Wheels - A Survey","authors":"T. Jacobs","doi":"10.1109/M2VIP.2018.8600822","DOIUrl":"https://doi.org/10.1109/M2VIP.2018.8600822","url":null,"abstract":"In the last decades, omnidirectional mobile robot platforms have entered many areas of industrial automation and daily life. While platforms with universal and Mecanum wheels are quite common, standard wheels have the advantage of robustness against uneven ground and slippage. This paper gives an overview on recent developments in the design of omnidirectional undercarriages with standard wheels and categorizes mechanical designs according to their holonomy. Furthermore, requirements for different applications are discussed and simple evaluation criteria are proposed that can be used to compare designs.","PeriodicalId":365579,"journal":{"name":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130726950","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 : 2018-11-01DOI: 10.1109/M2VIP.2018.8600821
J. Schutte, J. Leveneur, Xiaowen Yuan, J. Potgieter
Plasma derived cross-linking of polymer surfaces is an established technique within industry. This study investigated how the utilization and proximity of an atmospheric plasma generating device affects the ultimate tensile strength and surface quality of 3D printed Acrylonitrile Butadiene Styrene. The study made use of a linear stage device to accurately modify the distance between parts and devices. We aimed to evaluate the potential of atmospheric plasma treatment to improve the mechanical performance of 3D printing systems. 3D printed samples were generated by a Tier Time UPBox Fused deposition modelling printer. Thirty ASTM D638 Type IV dog-bone samples were generated. They were grouped into 6 categories corresponding to varied heights from the generated plasma, namely 36 mm (Sample 1), 30 mm (Sample 2), 24 mm (Sample 3), 18 mm (Sample 4), 12 mm (Sample 5) as well as an unprocessed Benchmark Sample. Optimal ultimate tensile strength occurred at 24 mm being 19.01 MPa whilst the surface quality of the samples experienced little change.
{"title":"Evaluation of the effects of corona discharge plasma exposure proximity to Fused Deposition Modelling 3D Printed Acrylonitrile Butadiene Styrene","authors":"J. Schutte, J. Leveneur, Xiaowen Yuan, J. Potgieter","doi":"10.1109/M2VIP.2018.8600821","DOIUrl":"https://doi.org/10.1109/M2VIP.2018.8600821","url":null,"abstract":"Plasma derived cross-linking of polymer surfaces is an established technique within industry. This study investigated how the utilization and proximity of an atmospheric plasma generating device affects the ultimate tensile strength and surface quality of 3D printed Acrylonitrile Butadiene Styrene. The study made use of a linear stage device to accurately modify the distance between parts and devices. We aimed to evaluate the potential of atmospheric plasma treatment to improve the mechanical performance of 3D printing systems. 3D printed samples were generated by a Tier Time UPBox Fused deposition modelling printer. Thirty ASTM D638 Type IV dog-bone samples were generated. They were grouped into 6 categories corresponding to varied heights from the generated plasma, namely 36 mm (Sample 1), 30 mm (Sample 2), 24 mm (Sample 3), 18 mm (Sample 4), 12 mm (Sample 5) as well as an unprocessed Benchmark Sample. Optimal ultimate tensile strength occurred at 24 mm being 19.01 MPa whilst the surface quality of the samples experienced little change.","PeriodicalId":365579,"journal":{"name":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134289749","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}
Many artificial muscles have been reported in recent years, in addition to the most popular pneumatic artificial muscles, there are many other artificial muscles that based on smart materials or composite materials. Artificial muscles usually have compliant structure, light weight, or high power-to-weight ratio. However, they also expose many defects that caused by their structure or material, moreover they are not developed in line with muscle mechanics. Therefore, based on Hill muscle model, we proposed a novel tendon-sheath artificial muscle with slender structure and light weight, it can meet the requirements of robotic applications like exoskeleton. Moreover, a small clamper driven by electromagnet is developed in this paper for the switching control of muscle activation and inactivation. The clamper is designed for clamping wire rope with diameter of 1-2mm which refers to the inner tendon of the tendon-sheath mechanism, and the maximum clamping force is 800N. The clamping performance of the clamper was tested on an experiment platform. The experimental results illustrated that the small clamper can clamp the moving wire rope effectively. On average, it takes 230ms for the small clamper from receiving the control signal to starting clamping, and another 150ms is taken for fully clamping.
{"title":"Development of A Small Clamper for Tendon-sheath Artificial Muscle","authors":"Qi Zhang, Xiaopeng Shen, Xingsong Wang, Mengqian Tian, Mingxing Yang, Kaiwei Ma, Qingcong Wu","doi":"10.1109/M2VIP.2018.8600846","DOIUrl":"https://doi.org/10.1109/M2VIP.2018.8600846","url":null,"abstract":"Many artificial muscles have been reported in recent years, in addition to the most popular pneumatic artificial muscles, there are many other artificial muscles that based on smart materials or composite materials. Artificial muscles usually have compliant structure, light weight, or high power-to-weight ratio. However, they also expose many defects that caused by their structure or material, moreover they are not developed in line with muscle mechanics. Therefore, based on Hill muscle model, we proposed a novel tendon-sheath artificial muscle with slender structure and light weight, it can meet the requirements of robotic applications like exoskeleton. Moreover, a small clamper driven by electromagnet is developed in this paper for the switching control of muscle activation and inactivation. The clamper is designed for clamping wire rope with diameter of 1-2mm which refers to the inner tendon of the tendon-sheath mechanism, and the maximum clamping force is 800N. The clamping performance of the clamper was tested on an experiment platform. The experimental results illustrated that the small clamper can clamp the moving wire rope effectively. On average, it takes 230ms for the small clamper from receiving the control signal to starting clamping, and another 150ms is taken for fully clamping.","PeriodicalId":365579,"journal":{"name":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","volume":"74 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133071227","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 : 2018-11-01DOI: 10.1109/M2VIP.2018.8600841
Y. Miao, Zhisheng Zhang, F. Jia, Min Dai
The main purpose of this paper is to build an embedded platform based on TMS320VC5509A processor, and finally realizes the recognition of snore and controls pillow change the height to relieve snoring symptoms. The whole embedded hardware platform collects the sounds through the microphone module, and completes data storage, data processing and data interaction through other peripherals. After a series of pre-processing operations, short-time energy and short-time zero crossing rate dual threshold detection is selected as endpoint recognition algorithm, MFCC(Mel Frequency Cepstral Coefficient) is selected as feature extraction and KNN(k-nearest neighbors algorithm) is selected as recognition algorithm. The experimental results show that this hardware system can run well and the design is reasonable. At the same time, it can also achieve a good accuracy in snore analysis and recognition.
本文的主要目的是构建一个基于TMS320VC5509A处理器的嵌入式平台，最终实现对打鼾的识别，并控制枕头的高度变化来缓解打鼾症状。整个嵌入式硬件平台通过麦克风模块采集声音，并通过其他外设完成数据存储、数据处理和数据交互。经过一系列预处理操作，选择短时间能量和短时间过零率双阈值检测作为端点识别算法，选择Mel Frequency Cepstral Coefficient (MFCC)作为特征提取算法，选择KNN(k近邻算法)作为识别算法。实验结果表明，该硬件系统运行良好，设计合理。同时，在鼾声分析和识别方面也能达到很好的准确率。
{"title":"Treatment pillow for relieving snoring symptoms based on snore recognition","authors":"Y. Miao, Zhisheng Zhang, F. Jia, Min Dai","doi":"10.1109/M2VIP.2018.8600841","DOIUrl":"https://doi.org/10.1109/M2VIP.2018.8600841","url":null,"abstract":"The main purpose of this paper is to build an embedded platform based on TMS320VC5509A processor, and finally realizes the recognition of snore and controls pillow change the height to relieve snoring symptoms. The whole embedded hardware platform collects the sounds through the microphone module, and completes data storage, data processing and data interaction through other peripherals. After a series of pre-processing operations, short-time energy and short-time zero crossing rate dual threshold detection is selected as endpoint recognition algorithm, MFCC(Mel Frequency Cepstral Coefficient) is selected as feature extraction and KNN(k-nearest neighbors algorithm) is selected as recognition algorithm. The experimental results show that this hardware system can run well and the design is reasonable. At the same time, it can also achieve a good accuracy in snore analysis and recognition.","PeriodicalId":365579,"journal":{"name":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122425756","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 : 2018-11-01DOI: 10.1109/M2VIP.2018.8600915
Karim Ali, Fabian Baumeister, A. Verl
In current calibration and compensation methods for gear backlash of industrial robots, only stationary effects are considered. Alternatively, expensive sensors are used for absolute position control of robots. This paper investigates the dynamic behavior of gear backlash and the measurability of its influences at a KUKA KR500–3 MT robot in the context of online gear backlash compensation.
{"title":"Measurability of the dynamic behavior of gear backlash at industrial robots","authors":"Karim Ali, Fabian Baumeister, A. Verl","doi":"10.1109/M2VIP.2018.8600915","DOIUrl":"https://doi.org/10.1109/M2VIP.2018.8600915","url":null,"abstract":"In current calibration and compensation methods for gear backlash of industrial robots, only stationary effects are considered. Alternatively, expensive sensors are used for absolute position control of robots. This paper investigates the dynamic behavior of gear backlash and the measurability of its influences at a KUKA KR500–3 MT robot in the context of online gear backlash compensation.","PeriodicalId":365579,"journal":{"name":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132169831","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 : 2018-11-01DOI: 10.1109/M2VIP.2018.8600844
Florian Jaensch, A. Csiszar, Christian Scheifele, A. Verl
In the engineering phase of modern manufacturing systems, simulation-based methods and tools have been established to face the increasing demands on time-efficiency and profitability. In the application of these simulation solutions, model-based digital twins are created, as multi-domain simulation models to describe the behavior of the manufacturing system. During the production process, a data-driven digital twin arises in the context of industry 4.0 based on an increasing networking and new cloud technologies. Recent developments in machine learning of fer new possibilities in conjunction with the digital twin. These range from data-based learning of models to learning control logic of complex systems. This paper proposes a combined model-based and data-driven concept of a digital twin. It shows how to use machine learning in connection with these models, in order to archive faster development times of manufacturing systems.
{"title":"Digital Twins of Manufacturing Systems as a Base for Machine Learning","authors":"Florian Jaensch, A. Csiszar, Christian Scheifele, A. Verl","doi":"10.1109/M2VIP.2018.8600844","DOIUrl":"https://doi.org/10.1109/M2VIP.2018.8600844","url":null,"abstract":"In the engineering phase of modern manufacturing systems, simulation-based methods and tools have been established to face the increasing demands on time-efficiency and profitability. In the application of these simulation solutions, model-based digital twins are created, as multi-domain simulation models to describe the behavior of the manufacturing system. During the production process, a data-driven digital twin arises in the context of industry 4.0 based on an increasing networking and new cloud technologies. Recent developments in machine learning of fer new possibilities in conjunction with the digital twin. These range from data-based learning of models to learning control logic of complex systems. This paper proposes a combined model-based and data-driven concept of a digital twin. It shows how to use machine learning in connection with these models, in order to archive faster development times of manufacturing systems.","PeriodicalId":365579,"journal":{"name":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128646644","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 : 2018-11-01DOI: 10.1109/M2VIP.2018.8600912
Fatma Haouas, B. Solaiman, Z. B. Dhiaf, A. Hamouda
In this paper, we present a new method for automatic mass function estimation and focal elements selection as a fundamental step to apply the Dempster-Shafer Theory (DST). The idea is to use the centroids and membership distributions obtained by applying the Fuzzy-C-Means algorithm (FCM) to define the mass function. The proposed method allows finding composite focal elements that represent the highest uncertainty and ambiguity. Experiments were conduced on multi-spectral and multi-temporal images for the purpose of change detection by integrating the proposed method of mass function estimation in a process of post-classification by DST. The proposed system of change detection is characterised by a multi-level of imperfection handling where the ambiguity is modelled firstly by FCM, then the uncertainty and the imprecision are handled in the step of mass function estimation. The effectiveness of the proposed methodology is demonstrated by applying it to find transformed region within two landsat images where we obtained high rates of classifications.
{"title":"Automatic mass function estimation based Fuzzy-C-Means algorithm for remote sensing images change detection","authors":"Fatma Haouas, B. Solaiman, Z. B. Dhiaf, A. Hamouda","doi":"10.1109/M2VIP.2018.8600912","DOIUrl":"https://doi.org/10.1109/M2VIP.2018.8600912","url":null,"abstract":"In this paper, we present a new method for automatic mass function estimation and focal elements selection as a fundamental step to apply the Dempster-Shafer Theory (DST). The idea is to use the centroids and membership distributions obtained by applying the Fuzzy-C-Means algorithm (FCM) to define the mass function. The proposed method allows finding composite focal elements that represent the highest uncertainty and ambiguity. Experiments were conduced on multi-spectral and multi-temporal images for the purpose of change detection by integrating the proposed method of mass function estimation in a process of post-classification by DST. The proposed system of change detection is characterised by a multi-level of imperfection handling where the ambiguity is modelled firstly by FCM, then the uncertainty and the imprecision are handled in the step of mass function estimation. The effectiveness of the proposed methodology is demonstrated by applying it to find transformed region within two landsat images where we obtained high rates of classifications.","PeriodicalId":365579,"journal":{"name":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125300770","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 : 2018-11-01DOI: 10.1109/M2VIP.2018.8600839
Matthias Milan Strljic, Daniella Brovkina, T. Korb, O. Riedel, A. Lechler
In this paper, a work in progress system architecture is presented, which aims to implement a continuous production chain for collective cloud manufacturing. Current problems for collective production in the areas of cloud manufacturing and computer-aided process planning are listed, which have a particular impact on the individuality and innovative ability of SMEs. To solve these problems, an architectural concept using software containers and an evolutionary needs model is presented. The focus here is on making the decision to collaborate on providing a service in regards to a global demand-model and extending the external commissioning from a single component to the entire production of a product as a demand. Finally, an innovative concept for production process planning for assembly lines using model-based systems engineering approach is presented as an evaluation scenario.
{"title":"Platform architecture concept for the composition of collective cloud manufacturing*","authors":"Matthias Milan Strljic, Daniella Brovkina, T. Korb, O. Riedel, A. Lechler","doi":"10.1109/M2VIP.2018.8600839","DOIUrl":"https://doi.org/10.1109/M2VIP.2018.8600839","url":null,"abstract":"In this paper, a work in progress system architecture is presented, which aims to implement a continuous production chain for collective cloud manufacturing. Current problems for collective production in the areas of cloud manufacturing and computer-aided process planning are listed, which have a particular impact on the individuality and innovative ability of SMEs. To solve these problems, an architectural concept using software containers and an evolutionary needs model is presented. The focus here is on making the decision to collaborate on providing a service in regards to a global demand-model and extending the external commissioning from a single component to the entire production of a product as a demand. Finally, an innovative concept for production process planning for assembly lines using model-based systems engineering approach is presented as an evaluation scenario.","PeriodicalId":365579,"journal":{"name":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","volume":"135 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127347931","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 : 2018-11-01DOI: 10.1109/M2VIP.2018.8600840
Zhong Hoo Chau, Ishara Paranawithana, Liangjing Yang, U-Xuan Tan
There are many approaches to plant cell segmentation, but there is no established method to segment plant cell for a portable, USB-powered optical microscope. Existing methods leverage on sophisticated microscope such as confocal laser scanning microscope or electron microscope may not be applicable for a portable setup. Staining of plant cell specimens, in order to improve visibility of boundaries, might affect the plant cell and also requires additional preparation work prior to acquisition which could be infeasible for on-the-fly applications. Conventional plant cell segmentation using watershed transform often results in over-segmentation, hindering the effectiveness of the method. Hence, we propose a thresholding method based on Otsu’s method, to retain majority of the image information to improve the success rate of the cell segmentation. The method is implemented on a leaf cellular image acquired from freshwater weed elodea. The region identified by the improved watershed transform can be further processed to locate the centroids of the cells. We experimented our method on images filled fully with plant cells and filled partially with plant cells. We also studied the impact of boundary definition of the image to our method.
{"title":"Plant Cell Segmentation with Adaptive Thresholding","authors":"Zhong Hoo Chau, Ishara Paranawithana, Liangjing Yang, U-Xuan Tan","doi":"10.1109/M2VIP.2018.8600840","DOIUrl":"https://doi.org/10.1109/M2VIP.2018.8600840","url":null,"abstract":"There are many approaches to plant cell segmentation, but there is no established method to segment plant cell for a portable, USB-powered optical microscope. Existing methods leverage on sophisticated microscope such as confocal laser scanning microscope or electron microscope may not be applicable for a portable setup. Staining of plant cell specimens, in order to improve visibility of boundaries, might affect the plant cell and also requires additional preparation work prior to acquisition which could be infeasible for on-the-fly applications. Conventional plant cell segmentation using watershed transform often results in over-segmentation, hindering the effectiveness of the method. Hence, we propose a thresholding method based on Otsu’s method, to retain majority of the image information to improve the success rate of the cell segmentation. The method is implemented on a leaf cellular image acquired from freshwater weed elodea. The region identified by the improved watershed transform can be further processed to locate the centroids of the cells. We experimented our method on images filled fully with plant cells and filled partially with plant cells. We also studied the impact of boundary definition of the image to our method.","PeriodicalId":365579,"journal":{"name":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116819058","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 : 2018-07-22DOI: 10.1109/M2VIP.2018.8600838
Malik Aqeel Anwar, A. Raychowdhury
We present NavREn-Rl, an approach to NAVigate an unmanned aerial vehicle in an indoor Real ENvironment via end-to-end reinforcement learning (RL). A suitable reward function is designed keeping in mind the cost and weight constraints for micro drone with minimum number of sensing modalities. Collection of small number of expert data and knowledge based data aggregation is integrated into the RL process to aid convergence. Experimentation is carried out on a Parrot AR drone in different indoor arenas and the results are compared with other baseline technologies. We demonstrate how the drone successfully avoids obstacles and navigates across different arenas. Video of the drone navigating using the proposed approach can be seen at https://youtu.be/yOTkTHUPNVY
{"title":"NavREn-Rl: Learning to fly in real environment via end-to-end deep reinforcement learning using monocular images","authors":"Malik Aqeel Anwar, A. Raychowdhury","doi":"10.1109/M2VIP.2018.8600838","DOIUrl":"https://doi.org/10.1109/M2VIP.2018.8600838","url":null,"abstract":"We present NavREn-Rl, an approach to NAVigate an unmanned aerial vehicle in an indoor Real ENvironment via end-to-end reinforcement learning (RL). A suitable reward function is designed keeping in mind the cost and weight constraints for micro drone with minimum number of sensing modalities. Collection of small number of expert data and knowledge based data aggregation is integrated into the RL process to aid convergence. Experimentation is carried out on a Parrot AR drone in different indoor arenas and the results are compared with other baseline technologies. We demonstrate how the drone successfully avoids obstacles and navigates across different arenas. Video of the drone navigating using the proposed approach can be seen at https://youtu.be/yOTkTHUPNVY","PeriodicalId":365579,"journal":{"name":"2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123455021","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：