Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981618
Cézar B. Lemos, P. Farias, Eduardo F. Simas Filho, A. Conceicao
Efficient object detection is important for automatic manufacturing systems applications. This work proposes the use of deep learning neural networks for vision-based additive manufactured object recognition. Three deep learning based object detection architectures (SSD300, SSD512 and Faster R-CNN) are applied for detection of parts manufactured on a 3D printer. The object detection information is used to feed a vision-based robotic grasping task, as part of a robotic assisted additive manufacturing system. Transfer learning is applied and a high detection efficiency is achieved for the considered dataset.
{"title":"Convolutional Neural Network Based Object Detection for Additive Manufacturing","authors":"Cézar B. Lemos, P. Farias, Eduardo F. Simas Filho, A. Conceicao","doi":"10.1109/ICAR46387.2019.8981618","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981618","url":null,"abstract":"Efficient object detection is important for automatic manufacturing systems applications. This work proposes the use of deep learning neural networks for vision-based additive manufactured object recognition. Three deep learning based object detection architectures (SSD300, SSD512 and Faster R-CNN) are applied for detection of parts manufactured on a 3D printer. The object detection information is used to feed a vision-based robotic grasping task, as part of a robotic assisted additive manufacturing system. Transfer learning is applied and a high detection efficiency is achieved for the considered dataset.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"9 1","pages":"420-425"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82431593","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 : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981629
Victor Souza, Alan F. P. Tavares, C. Quiroz, P. Kurka
Surveillance of overhead power line installations can be conveniently addressed using unmanned aerial vehicles (UAV). UAV are robotic platforms able to perform sophisticated tasks such as autonomous flight based on visual information. In this paper, we propose a novel solution to the problem of following a power line autonomously based on monocular vision. The method uses Deep Neural Networks (DNN) and the Hough transform to successfully discern power line images from environmental information, which is an essential result to accomplish fully autonomous vision-based navigation. A simulated navigation test demonstrates the efficiency of the proposed method, in the special condition of following right-angled changes of direction, which is a known restriction in many navigation methods reported in literature. The design of the proposed method is modular and can be incorporated in navigation strategies for automatic surveillance applications.
{"title":"Monocular vision navigation for aerial surveillance of power lines based on Deep Neural Networks and Hough transform","authors":"Victor Souza, Alan F. P. Tavares, C. Quiroz, P. Kurka","doi":"10.1109/ICAR46387.2019.8981629","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981629","url":null,"abstract":"Surveillance of overhead power line installations can be conveniently addressed using unmanned aerial vehicles (UAV). UAV are robotic platforms able to perform sophisticated tasks such as autonomous flight based on visual information. In this paper, we propose a novel solution to the problem of following a power line autonomously based on monocular vision. The method uses Deep Neural Networks (DNN) and the Hough transform to successfully discern power line images from environmental information, which is an essential result to accomplish fully autonomous vision-based navigation. A simulated navigation test demonstrates the efficiency of the proposed method, in the special condition of following right-angled changes of direction, which is a known restriction in many navigation methods reported in literature. The design of the proposed method is modular and can be incorporated in navigation strategies for automatic surveillance applications.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"54 1","pages":"414-419"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87523203","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 : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981602
Arne Sachtler, Korbinian Nottensteiner, M. Kassecker, A. Albu-Schäffer
Future manufacturing systems will have to allow frequent conversion of production processes. In order to prevent a surge in setup times and the cost involved, we suggest refraining from specialized part feeders, fixture units and manual calibration routines. To this effect we regard autonomous object registration as corner stone to lower the manual calibration effort. In this work, we propose a framework for autonomous object registration in robotic workcells using a combination of visual and touch-based sensing. Vision systems in process automation often require very defined conditions to produce reliable pose estimates. Therefore, we combine data from a vision system with touch-based sensing in order to benefit from both sensing modalities and reduce the requirements on the vision system. We use a particle filter approach in order to estimate the pose of individual features and show how to retrieve the overall pose of an object in the workcell from the feature estimates. The observed feature distribution is used to autonomously trigger dedicated actions for touch-based probing with a lightweight robotic arm in order to increase the accuracy. In particular, we describe the detection of circular features and validated the framework in experiments with our robotic assembly system.
{"title":"Combined Visual and Touch-based Sensing for the Autonomous Registration of Objects with Circular Features","authors":"Arne Sachtler, Korbinian Nottensteiner, M. Kassecker, A. Albu-Schäffer","doi":"10.1109/ICAR46387.2019.8981602","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981602","url":null,"abstract":"Future manufacturing systems will have to allow frequent conversion of production processes. In order to prevent a surge in setup times and the cost involved, we suggest refraining from specialized part feeders, fixture units and manual calibration routines. To this effect we regard autonomous object registration as corner stone to lower the manual calibration effort. In this work, we propose a framework for autonomous object registration in robotic workcells using a combination of visual and touch-based sensing. Vision systems in process automation often require very defined conditions to produce reliable pose estimates. Therefore, we combine data from a vision system with touch-based sensing in order to benefit from both sensing modalities and reduce the requirements on the vision system. We use a particle filter approach in order to estimate the pose of individual features and show how to retrieve the overall pose of an object in the workcell from the feature estimates. The observed feature distribution is used to autonomously trigger dedicated actions for touch-based probing with a lightweight robotic arm in order to increase the accuracy. In particular, we describe the detection of circular features and validated the framework in experiments with our robotic assembly system.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"4 1","pages":"426-433"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90148655","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 : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981606
B. Nemec, Mihael Simonič, T. Petrič, A. Ude
Fast deployment of robot tasks requires appropriate tools that enable efficient reuse of existing robot control policies. Learning from Demonstration (LfD) is a popular tool for the intuitive generation of robot policies, but the issue of how to address the adaptation of existing policies has not been properly addressed yet. In this work, we propose an incremental LfD framework that efficiently solves the above-mentioned issue. It has been implemented and tested on a number of popular collaborative robots, including Franka Emika Panda, Universal Robot UR10, and KUKA LWR 4.
{"title":"Incremental Policy Refinement by Recursive Regression and Kinesthetic Guidance","authors":"B. Nemec, Mihael Simonič, T. Petrič, A. Ude","doi":"10.1109/ICAR46387.2019.8981606","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981606","url":null,"abstract":"Fast deployment of robot tasks requires appropriate tools that enable efficient reuse of existing robot control policies. Learning from Demonstration (LfD) is a popular tool for the intuitive generation of robot policies, but the issue of how to address the adaptation of existing policies has not been properly addressed yet. In this work, we propose an incremental LfD framework that efficiently solves the above-mentioned issue. It has been implemented and tested on a number of popular collaborative robots, including Franka Emika Panda, Universal Robot UR10, and KUKA LWR 4.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"8 1","pages":"344-349"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85306676","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 : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981612
Y. Kita, Ryuichi Takase, Tatsuya Komuro, Norihiko Kato, N. Kita
This paper proposes a method to detect and localize pallets on shelves using a wide-angle camera mounted on a vehicle several meters away along the shelves. Although wide-angle observation is a significant advantage in navigation, objects appear small and largely distorted in the observed image, thus creating a severe disadvantage for the detection and localization of objects. To solve this problem, we propose reprojecting wide-angle views onto a three-dimensional (3D) plane to obtain a good projection image for object detection, in which pallets appear in the same shape and size regardless of the location in the image. Since the image projected on the plane coplanar to the pallet's front surface satisfy the above condition, we approximate the projection plane using the front surface of the shelves with pallets on it. Owing to the constant appearance of the pallet on the projected image, it can be robustly detected using the matching with a fixed template. When a pallet is detected, it is three-dimensionally localized from the equation of the 3D plane of the projection plane and its 2D location in the projected image. Experiments using empirically captured scenes demonstrated the detection and localization ability of the proposed method. Despite small area in the original wide-angle view, pallets of interest were discriminatively detected from others and localized with acceptable accuracy.
{"title":"Detection and localization of pallets on shelves using a wide-angle camera","authors":"Y. Kita, Ryuichi Takase, Tatsuya Komuro, Norihiko Kato, N. Kita","doi":"10.1109/ICAR46387.2019.8981612","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981612","url":null,"abstract":"This paper proposes a method to detect and localize pallets on shelves using a wide-angle camera mounted on a vehicle several meters away along the shelves. Although wide-angle observation is a significant advantage in navigation, objects appear small and largely distorted in the observed image, thus creating a severe disadvantage for the detection and localization of objects. To solve this problem, we propose reprojecting wide-angle views onto a three-dimensional (3D) plane to obtain a good projection image for object detection, in which pallets appear in the same shape and size regardless of the location in the image. Since the image projected on the plane coplanar to the pallet's front surface satisfy the above condition, we approximate the projection plane using the front surface of the shelves with pallets on it. Owing to the constant appearance of the pallet on the projected image, it can be robustly detected using the matching with a fixed template. When a pallet is detected, it is three-dimensionally localized from the equation of the 3D plane of the projection plane and its 2D location in the projected image. Experiments using empirically captured scenes demonstrated the detection and localization ability of the proposed method. Despite small area in the original wide-angle view, pallets of interest were discriminatively detected from others and localized with acceptable accuracy.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"47 1","pages":"785-792"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91136279","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 : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981639
Kenny A. Q. Caldas, V. Grassi
The purpose of this work is the development of a nonlinear model-based predictive controller for Eco-cruise in autonomous ground vehicles. Eco-driving consists of a group of strategies adopted by a driver aiming to reduce fuel consumption and improvement of safety and comfort levels during a trip. Through the use of latitude-longitude information and a GPS module, the predictive controller can calculate a sequence of control input to smooth the vehicle's acceleration and braking along the route in critical parts, such as uphills, downhills and curves, following the speed limits of each road. This is accomplished by predictions based on the mathematical model of the vehicle and estimation of gasoline expenditure. The chosen optimizer algorithm is called C/GMRES, where its main advantage from the traditional methods is that the solution of the optimal problem does not require iterative searches, which greatly reduces the computational burden, allowing a real time implementation. Another advantage of the proposed method is that it only requires publicly available data to obtain the controller parameters. The Eco-cruise NMPC was implemented in simulation and compared with a cruise controller. The obtained results were considered satisfactory and showed the predictive controller's potential to reduce fuel consumption in autonomous vehicles.
{"title":"Eco-cruise NMPC Control for Autonomous Vehicles","authors":"Kenny A. Q. Caldas, V. Grassi","doi":"10.1109/ICAR46387.2019.8981639","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981639","url":null,"abstract":"The purpose of this work is the development of a nonlinear model-based predictive controller for Eco-cruise in autonomous ground vehicles. Eco-driving consists of a group of strategies adopted by a driver aiming to reduce fuel consumption and improvement of safety and comfort levels during a trip. Through the use of latitude-longitude information and a GPS module, the predictive controller can calculate a sequence of control input to smooth the vehicle's acceleration and braking along the route in critical parts, such as uphills, downhills and curves, following the speed limits of each road. This is accomplished by predictions based on the mathematical model of the vehicle and estimation of gasoline expenditure. The chosen optimizer algorithm is called C/GMRES, where its main advantage from the traditional methods is that the solution of the optimal problem does not require iterative searches, which greatly reduces the computational burden, allowing a real time implementation. Another advantage of the proposed method is that it only requires publicly available data to obtain the controller parameters. The Eco-cruise NMPC was implemented in simulation and compared with a cruise controller. The obtained results were considered satisfactory and showed the predictive controller's potential to reduce fuel consumption in autonomous vehicles.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"46 1","pages":"356-361"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90824590","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 : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981667
Alexandre C. Horn, P. Pinheiro, C. B. Silva, A. A. Neto, Paulo L. J. Drews-Jr
This paper presents a study on the configuration of propulsion systems for multirotor-like hybrid vehicles that travel in both, aerial and aquatic medium, hereafter called Hybrid Unmanned Aerial Underwater Vehicles (HUAUVs). Following, we evaluate the impact of diameter and shape of propellers, and the medium in which the robot moves, on thrust generation and energy consumption. Sets of counter-rotating propellers (CRPs) and single propellers propulsion systems were experimentally evaluated in air and under the water. One of our conclusions is that aerial counter-rotating propellers, used in the state-of-the-art hybrid vehicles, are relatively inefficient in the air, and even more in the water, and this energy loss must be taken into account during the platform project. Furthermore, we also demonstrate that aquatic screwlike propellers are more efficient in the water than aerial ones, equally contradicting the current literature. Moreover, since aquatic propellers provide more thrust with smaller diameters, they improve our capacity to miniaturize hybrid vehicles.
{"title":"A Study on Configuration of Propellers for Multirotor-like Hybrid Aerial-Aquatic Vehicles","authors":"Alexandre C. Horn, P. Pinheiro, C. B. Silva, A. A. Neto, Paulo L. J. Drews-Jr","doi":"10.1109/ICAR46387.2019.8981667","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981667","url":null,"abstract":"This paper presents a study on the configuration of propulsion systems for multirotor-like hybrid vehicles that travel in both, aerial and aquatic medium, hereafter called Hybrid Unmanned Aerial Underwater Vehicles (HUAUVs). Following, we evaluate the impact of diameter and shape of propellers, and the medium in which the robot moves, on thrust generation and energy consumption. Sets of counter-rotating propellers (CRPs) and single propellers propulsion systems were experimentally evaluated in air and under the water. One of our conclusions is that aerial counter-rotating propellers, used in the state-of-the-art hybrid vehicles, are relatively inefficient in the air, and even more in the water, and this energy loss must be taken into account during the platform project. Furthermore, we also demonstrate that aquatic screwlike propellers are more efficient in the water than aerial ones, equally contradicting the current literature. Moreover, since aquatic propellers provide more thrust with smaller diameters, they improve our capacity to miniaturize hybrid vehicles.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"70 1","pages":"173-178"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75324119","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 : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981666
I. J. P. B. Franco, T. T. Ribeiro, A. Conceicao
Images captured by visual sensors, such as cameras, with the goal of performing image based control, require processing for the extraction of useful information in the presence of imperfection of objects of the scene and restrictive environmental conditions. The problem of path following encounters these inconveniences, more precisely in the detection of the marks that represent the path to be followed. Handling faults along the path on non-homogeneous floors and extracting parameters, such as visual pose and curvature, accurately, are some of the difficulties encountered. In this article, a system of detection and extraction of parameters for the path following problem based on NMPC (Nonlinear Model Predictive Control), using computer vision techniques is proposed. To remedy the above-mentioned problems, the visual path is approximated by a quadratic function. The algorithm proposed here was embedded in Husky UGV (Unmanned Ground Vehicle) robot and compared with the original approach. Experimental results demonstrate the superiority of the proposed new algorithm.
{"title":"A Novel Approach for Parameter Extraction of an NMPC-based Visual Follower Model","authors":"I. J. P. B. Franco, T. T. Ribeiro, A. Conceicao","doi":"10.1109/ICAR46387.2019.8981666","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981666","url":null,"abstract":"Images captured by visual sensors, such as cameras, with the goal of performing image based control, require processing for the extraction of useful information in the presence of imperfection of objects of the scene and restrictive environmental conditions. The problem of path following encounters these inconveniences, more precisely in the detection of the marks that represent the path to be followed. Handling faults along the path on non-homogeneous floors and extracting parameters, such as visual pose and curvature, accurately, are some of the difficulties encountered. In this article, a system of detection and extraction of parameters for the path following problem based on NMPC (Nonlinear Model Predictive Control), using computer vision techniques is proposed. To remedy the above-mentioned problems, the visual path is approximated by a quadratic function. The algorithm proposed here was embedded in Husky UGV (Unmanned Ground Vehicle) robot and compared with the original approach. Experimental results demonstrate the superiority of the proposed new algorithm.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"17 1","pages":"117-122"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73023077","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 : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981609
T. C. Santos, D. Wolf
Lane change is one of the most common maneuvers in traffic. It is a lateral movement which can affect the trajectory of other vehicles, with the possibility of collision. Although the field of autonomous vehicles has been growing and the repertoire of methods used for autonomous navigation is extensive, lane-changing still one of the causes of flow reduction and accidents. Researches usually address the lane change problem from the perspective of a single-agent. On the other hand, multi-agent systems based on interaction among vehicles, equipped with wireless communication technologies, can improve traffic efficiency and safety. This paper presents a multi-agent automated conflict resolution to collision avoidance method based on Probability Collectives to perform coordinated maneuvers in a fully decentralized manner using communication. We used SUMO to simulate lane-changing situations in a two-lane highway scenario, where the vehicles had to do a lane-merging maneuver. We evaluate the time for all vehicles of the platoon complete a lane-merging, vehicle flow per second, maximum string length, and the platoon's average speed. We compared the SUMO centralized lane change method with a fully decentralized PC method. The results show that the PC method closely approximates to the centralized SUMO method which uses all data for its decision making.
{"title":"Automated Conflict Resolution of Lane Change Utilizing Probability Collectives","authors":"T. C. Santos, D. Wolf","doi":"10.1109/ICAR46387.2019.8981609","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981609","url":null,"abstract":"Lane change is one of the most common maneuvers in traffic. It is a lateral movement which can affect the trajectory of other vehicles, with the possibility of collision. Although the field of autonomous vehicles has been growing and the repertoire of methods used for autonomous navigation is extensive, lane-changing still one of the causes of flow reduction and accidents. Researches usually address the lane change problem from the perspective of a single-agent. On the other hand, multi-agent systems based on interaction among vehicles, equipped with wireless communication technologies, can improve traffic efficiency and safety. This paper presents a multi-agent automated conflict resolution to collision avoidance method based on Probability Collectives to perform coordinated maneuvers in a fully decentralized manner using communication. We used SUMO to simulate lane-changing situations in a two-lane highway scenario, where the vehicles had to do a lane-merging maneuver. We evaluate the time for all vehicles of the platoon complete a lane-merging, vehicle flow per second, maximum string length, and the platoon's average speed. We compared the SUMO centralized lane change method with a fully decentralized PC method. The results show that the PC method closely approximates to the centralized SUMO method which uses all data for its decision making.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"14 1","pages":"623-628"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73263749","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 : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981626
Juan Rodriguez, Davinson Castano Cano
Localization proves still a challenging task in robotics, with Visual-SLAM being a both powerful and popular solution. In Ground-SLAM, a branch of Visual-SLAM that uses groundplane views, there is still a lot to explore but the tools for testing new methods need improvement. In this paper, we present a novel framework for simulating the motion of Visual-SLAM methods over a given groundplane. It can also be used to benchmark different methods on a given dataset. We present the functioning of our system, as well as the requirements for a similar system to work. In order to showcase the capabilities of SimSLAM 2D, we present a test case of different trajectories on a smooth concrete terrain comparing the performance of two methods, StreetMap and ORB-SLAM. We also test StreetMap in localization only mode with the purpose of showing the visualization tools. We thereby show the ease of simulating and testing a Visual-SLAM method on a 2D environment in our framework, which represents a useful tool when developing new Visual-SLAM methods.
{"title":"SimSLAM 2D: A Simulation Framework for Testing and Benchmarking of two-dimensional Visual-SLAM Methods","authors":"Juan Rodriguez, Davinson Castano Cano","doi":"10.1109/ICAR46387.2019.8981626","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981626","url":null,"abstract":"Localization proves still a challenging task in robotics, with Visual-SLAM being a both powerful and popular solution. In Ground-SLAM, a branch of Visual-SLAM that uses groundplane views, there is still a lot to explore but the tools for testing new methods need improvement. In this paper, we present a novel framework for simulating the motion of Visual-SLAM methods over a given groundplane. It can also be used to benchmark different methods on a given dataset. We present the functioning of our system, as well as the requirements for a similar system to work. In order to showcase the capabilities of SimSLAM 2D, we present a test case of different trajectories on a smooth concrete terrain comparing the performance of two methods, StreetMap and ORB-SLAM. We also test StreetMap in localization only mode with the purpose of showing the visualization tools. We thereby show the ease of simulating and testing a Visual-SLAM method on a 2D environment in our framework, which represents a useful tool when developing new Visual-SLAM methods.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"151 1","pages":"141-147"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84344877","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：