Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981582
M. Chandarana, Dana Hughes, M. Lewis, K. Sycara, S. Scherer
In Swarm Search and Service (SSS) applications, swarm vehicles are responsible for concurrently searching an area while immediately servicing jobs discovered while searching. Multiple job types may be present in the environment. As vehicles move in and out of the swarm to service jobs, the coverage rate (i.e., area searched by the swarm per time step) changes dynamically to reflect the number of vehicles currently engaged in search. As a result, the arrival rates of jobs also changes dynamically. When planning SSS missions, the resource requirements, such as the swarm size needed to achieve a desired system performance, must be determined. The dynamically changing arrival rates make traditional queuing methods ill-suited to predict the performance of the swarm. This paper presents a hybrid method - Hybrid Model - for predicting the performance of the swarm a priori. It utilizes a Markov model, whose state representation captures the proportion of agents searching or servicing jobs. State-dependent queuing models are used to calculate the state transition function of the Markov states. The model has been developed as a prediction tool to assist mission planners in balancing complex trade-offs, but also provides a basis for optimizing swarm size where cost functions are known. The Hybrid Model is tested in previously considered constant coverage rate scenarios and the results are compared to a previously developed Queuing Model. Additional SSS missions are then simulated and their resulting performance is used to further evaluate the effectiveness of using the Hybrid Model as a prediction tool for swarm performance in more general scenarios with dynamically changing coverage rates.
{"title":"Hybrid Model for A Priori Performance Prediction of Multi-Job Type Swarm Search and Service Missions","authors":"M. Chandarana, Dana Hughes, M. Lewis, K. Sycara, S. Scherer","doi":"10.1109/ICAR46387.2019.8981582","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981582","url":null,"abstract":"In Swarm Search and Service (SSS) applications, swarm vehicles are responsible for concurrently searching an area while immediately servicing jobs discovered while searching. Multiple job types may be present in the environment. As vehicles move in and out of the swarm to service jobs, the coverage rate (i.e., area searched by the swarm per time step) changes dynamically to reflect the number of vehicles currently engaged in search. As a result, the arrival rates of jobs also changes dynamically. When planning SSS missions, the resource requirements, such as the swarm size needed to achieve a desired system performance, must be determined. The dynamically changing arrival rates make traditional queuing methods ill-suited to predict the performance of the swarm. This paper presents a hybrid method - Hybrid Model - for predicting the performance of the swarm a priori. It utilizes a Markov model, whose state representation captures the proportion of agents searching or servicing jobs. State-dependent queuing models are used to calculate the state transition function of the Markov states. The model has been developed as a prediction tool to assist mission planners in balancing complex trade-offs, but also provides a basis for optimizing swarm size where cost functions are known. The Hybrid Model is tested in previously considered constant coverage rate scenarios and the results are compared to a previously developed Queuing Model. Additional SSS missions are then simulated and their resulting performance is used to further evaluate the effectiveness of using the Hybrid Model as a prediction tool for swarm performance in more general scenarios with dynamically changing coverage rates.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"11 1","pages":"714-719"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74674879","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/ICAR46387.2019.8981589
Vojtěch Vonásek, Robert Pěnička, B. Kozlíková
Path planning of 3D solid objects leads to search in a six-dimensional configuration space, which can be solved by sampling-based motion planning. The well-known issue of sampling-based planners is the narrow passage problem, which is caused by the presence of small regions of the configuration space that are difficult to cover by random samples. Guided-based planners cope with this issue by increasing the probability of sampling along an estimated solution (a guiding path). In the case of six-dimensional configuration space, the guiding path needs to be computed in the configuration space rather than in the workspace. Fast computation of guiding paths can be achieved by solving a similar, yet simpler problem, e.g., by reducing the size of the robot. This results in an approximate solution (path) that is assumed to be located near the solution of the original problem. The guided sampling along this approximate solution may, however, fail if the approximate solution is too far from the desired solution. In this paper, we cope with this problem by sampling the configuration space along multiple approximate solutions. The approximate solutions are computed using a proposed iterative process: after a path (solution) is found, it forms a region where the subsequent search is inhibited, which boosts the search of new solutions. The performance of the proposed approach is verified in scenarios with multiple narrow passages and compared with the state-of-the-art planners.
{"title":"Computing multiple guiding paths for sampling-based motion planning","authors":"Vojtěch Vonásek, Robert Pěnička, B. Kozlíková","doi":"10.1109/ICAR46387.2019.8981589","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981589","url":null,"abstract":"Path planning of 3D solid objects leads to search in a six-dimensional configuration space, which can be solved by sampling-based motion planning. The well-known issue of sampling-based planners is the narrow passage problem, which is caused by the presence of small regions of the configuration space that are difficult to cover by random samples. Guided-based planners cope with this issue by increasing the probability of sampling along an estimated solution (a guiding path). In the case of six-dimensional configuration space, the guiding path needs to be computed in the configuration space rather than in the workspace. Fast computation of guiding paths can be achieved by solving a similar, yet simpler problem, e.g., by reducing the size of the robot. This results in an approximate solution (path) that is assumed to be located near the solution of the original problem. The guided sampling along this approximate solution may, however, fail if the approximate solution is too far from the desired solution. In this paper, we cope with this problem by sampling the configuration space along multiple approximate solutions. The approximate solutions are computed using a proposed iterative process: after a path (solution) is found, it forms a region where the subsequent search is inhibited, which boosts the search of new solutions. The performance of the proposed approach is verified in scenarios with multiple narrow passages and compared with the state-of-the-art planners.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"33 1","pages":"374-381"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75090670","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/ICAR46387.2019.8981572
Felipe G. Oliveira, A. A. Neto, P. Borges, M. Campos, D. Macharet
In outdoor field robotics, considering the environmental characteristics is key to improving the efficiency of autonomous navigation. In this context, identifying rough terrain can significantly increase the reliability of operations. This paper addresses the problem of mapping the navigation cost associated with uneven outdoor terrains. We propose an augmented vector field representation obtained only with the use of inertial sensors. The map is determined considering characteristics such as roughness and slope. Experiments were carried out with different robots in real-world environments presenting different terrain characteristics to analyze the quality and efficiency of the mapping process. Results show that the obtained navigation cost maps provide a reliable indication of the ground characteristics of outdoor environments and can be used in the path planning stage.
{"title":"Augmented Vector Field Navigation Cost Mapping using Inertial Sensors","authors":"Felipe G. Oliveira, A. A. Neto, P. Borges, M. Campos, D. Macharet","doi":"10.1109/ICAR46387.2019.8981572","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981572","url":null,"abstract":"In outdoor field robotics, considering the environmental characteristics is key to improving the efficiency of autonomous navigation. In this context, identifying rough terrain can significantly increase the reliability of operations. This paper addresses the problem of mapping the navigation cost associated with uneven outdoor terrains. We propose an augmented vector field representation obtained only with the use of inertial sensors. The map is determined considering characteristics such as roughness and slope. Experiments were carried out with different robots in real-world environments presenting different terrain characteristics to analyze the quality and efficiency of the mapping process. Results show that the obtained navigation cost maps provide a reliable indication of the ground characteristics of outdoor environments and can be used in the path planning stage.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"46 1","pages":"388-393"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81524194","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/ICAR46387.2019.8981610
Sebastian Dingler, H. Burrichter
This paper proposes a novel method to study the error function of ICP based algorithms. With our method, we visualize the multidimensional error function of ICP algorithms in one dimension which allows us to compare and quantify the performance of ICP algorithms in an intuitive and descriptive manner. This is motivated by the fact that there are many ICP variants around and for researchers and engineers it is challenging which algorithm they shall choose. New approaches are often only evaluated based on runtime and accuracy. Our visual method allows to gain further insights beyond those metrics. We demonstrate the capability of our method by applying it to the KITTI LIDAR odometry benchmark. Our experiments show evidence for errors in the ground truth data, difficulties in highway scenarios and prove the power of superior error metrics such as the newly emerged symmetric objective function.
{"title":"A Visual Method to study the Error Function of ICP Algorithms","authors":"Sebastian Dingler, H. Burrichter","doi":"10.1109/ICAR46387.2019.8981610","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981610","url":null,"abstract":"This paper proposes a novel method to study the error function of ICP based algorithms. With our method, we visualize the multidimensional error function of ICP algorithms in one dimension which allows us to compare and quantify the performance of ICP algorithms in an intuitive and descriptive manner. This is motivated by the fact that there are many ICP variants around and for researchers and engineers it is challenging which algorithm they shall choose. New approaches are often only evaluated based on runtime and accuracy. Our visual method allows to gain further insights beyond those metrics. We demonstrate the capability of our method by applying it to the KITTI LIDAR odometry benchmark. Our experiments show evidence for errors in the ground truth data, difficulties in highway scenarios and prove the power of superior error metrics such as the newly emerged symmetric objective function.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"20 1","pages":"278-283"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81968841","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/ICAR46387.2019.8981607
Felipe J. S. Vasconcelos, Iury de Amorim Gaspar Filgueiras, W. Correia
Manipulators are becoming more and more common to perform many tasks, whose accomplishment is often related to the applied control law. Regarding to this issue, control theory is helpful as the proper controller choice may turn the manipulator into a handy tool. Within this context this work presents an automatic tuning method for the Generalized Predictive Controller (GPC) in order to tracking the trajectory of a manipulator end-effector. The strategy employs Particle Swarm Optimization (PSO) to properly determine GPC cost function weights at each iteration that lead to zero error tracking. The proposed controller is compared to classical approaches for three different trajectories with results showing a better performance and tracking accuracy for the proposed approach.
{"title":"Auto-Tuning of GPC weights based on Particle Swarm Optimization applied to a Manipulator End-Effector Trajectory Tracking","authors":"Felipe J. S. Vasconcelos, Iury de Amorim Gaspar Filgueiras, W. Correia","doi":"10.1109/ICAR46387.2019.8981607","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981607","url":null,"abstract":"Manipulators are becoming more and more common to perform many tasks, whose accomplishment is often related to the applied control law. Regarding to this issue, control theory is helpful as the proper controller choice may turn the manipulator into a handy tool. Within this context this work presents an automatic tuning method for the Generalized Predictive Controller (GPC) in order to tracking the trajectory of a manipulator end-effector. The strategy employs Particle Swarm Optimization (PSO) to properly determine GPC cost function weights at each iteration that lead to zero error tracking. The proposed controller is compared to classical approaches for three different trajectories with results showing a better performance and tracking accuracy for the proposed approach.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"16 1","pages":"702-707"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77019223","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/ICAR46387.2019.8981657
Christian Ritter, Shashank Sharma
This work addresses the field of human-robot collaboration, in particular, the hand-guidance of a redundant mobile manipulator. An efficient hand-guidance is related to the improvement of the inertial properties of the end-effector felt by the operator, as these describe the effort to guide the robot. In this regard, the operational space formulation is used to minimize the effective mass in the null space of the hand-guidance task. The minimization is based on a geometric ellipsoid representation and applies a gradient descent method. The approach is tested on the KUKA VALERI mobile manipulator and implemented on a real-time framework, whereby a particular focus is on stability and safety. For this purpose, a power limitation is proposed. Finally, the improvement of the effective mass related to a specific direction is discussed.
{"title":"Hand-Guidance of a Mobile Manipulator Using Online Effective Mass Optimization","authors":"Christian Ritter, Shashank Sharma","doi":"10.1109/ICAR46387.2019.8981657","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981657","url":null,"abstract":"This work addresses the field of human-robot collaboration, in particular, the hand-guidance of a redundant mobile manipulator. An efficient hand-guidance is related to the improvement of the inertial properties of the end-effector felt by the operator, as these describe the effort to guide the robot. In this regard, the operational space formulation is used to minimize the effective mass in the null space of the hand-guidance task. The minimization is based on a geometric ellipsoid representation and applies a gradient descent method. The approach is tested on the KUKA VALERI mobile manipulator and implemented on a real-time framework, whereby a particular focus is on stability and safety. For this purpose, a power limitation is proposed. Finally, the improvement of the effective mass related to a specific direction is discussed.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"21 1","pages":"192-197"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80321817","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/ICAR46387.2019.8981641
Victor R. F. Miranda, L. Mozelli, A. A. Neto, G. Freitas
The longitudinal trajectory tracking for small Unmanned Ground Vehicles (UGVs) subject to real-world disturbances is addressed, by considering an anisotropic ground friction and the variation of its own mass during transportation. A methodology was proposed to perform the design of the controller in the Proportional-Integral-Derivative (PID) format, which is robust to parametric uncertainties and minimizes output disturbances. Next, a small ground platform for load delivery tasks has been developed, equipped with low-cost navigation sensors and onboard computers. Finally, simulated and real-world experiments illustrate the effectiveness of this application under different scenarios, by combining distinct conditions of friction, payload, and terrain profiles.
{"title":"On the Robust Longitudinal Trajectory Tracking for Load Transportation Vehicles on Uneven Terrains","authors":"Victor R. F. Miranda, L. Mozelli, A. A. Neto, G. Freitas","doi":"10.1109/ICAR46387.2019.8981641","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981641","url":null,"abstract":"The longitudinal trajectory tracking for small Unmanned Ground Vehicles (UGVs) subject to real-world disturbances is addressed, by considering an anisotropic ground friction and the variation of its own mass during transportation. A methodology was proposed to perform the design of the controller in the Proportional-Integral-Derivative (PID) format, which is robust to parametric uncertainties and minimizes output disturbances. Next, a small ground platform for load delivery tasks has been developed, equipped with low-cost navigation sensors and onboard computers. Finally, simulated and real-world experiments illustrate the effectiveness of this application under different scenarios, by combining distinct conditions of friction, payload, and terrain profiles.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"34 1","pages":"320-325"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88834295","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/ICAR46387.2019.8981614
Mohammad Asif, Andreas Daasch, Hendrik Unger, M. Schultalbers
Low cost depth cameras and advancements in the field of deep learning have paved the way to precisely estimate 3D hand pose using a single depth camera. However, to accurately estimate the pose one has to detect the hands in the scene and track them over consecutive frames. In this paper, we propose a voting based system to track and estimate the 3D pose of a human hand. Based upon Robot Operating System (ROS), it comprises a hand segmentation stage, a clustering stage, a voting stage, a validation stage and a pose estimation stage. The final output is the 3D pose which is then used by a robot to follow the human hand.
{"title":"Voting Based System for Robust 3D Hand Pose Estimation and Tracking","authors":"Mohammad Asif, Andreas Daasch, Hendrik Unger, M. Schultalbers","doi":"10.1109/ICAR46387.2019.8981614","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981614","url":null,"abstract":"Low cost depth cameras and advancements in the field of deep learning have paved the way to precisely estimate 3D hand pose using a single depth camera. However, to accurately estimate the pose one has to detect the hands in the scene and track them over consecutive frames. In this paper, we propose a voting based system to track and estimate the 3D pose of a human hand. Based upon Robot Operating System (ROS), it comprises a hand segmentation stage, a clustering stage, a voting stage, a validation stage and a pose estimation stage. The final output is the 3D pose which is then used by a robot to follow the human hand.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"1 1","pages":"248-253"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84721607","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/ICAR46387.2019.8981576
Min Wang, H. Voos
Current commercial UAVs are to a large extent remotely piloted by amateur human pilots. Due to lack of teleoperation experience or skills, they often drive the UAVs into collision. Therefore, in order to ensure safety of the UAV as well as its surroundings, it is necessary for the UAV to boast the capability of detecting emergency situation and acting on its own when facing imminent threat. However, the majority of UAVs currently available in the market are not equipped with such capability. To fill in the gap, in this paper we present a complete sense-and-avoid solution for assisting unskilled pilots in ensuring a safe flight. Particularly, we propose a novel nonlinear vehicle control system which takes into account of sensor characteristics, an emergency evaluation policy and a novel optimization-based avoidance control strategy. The effectiveness of the proposed approach is demonstrated and validated in simulation with multiple moving objects.
{"title":"Safer UAV Piloting: A Robust Sense-and-Avoid Solution for Remotely Piloted Quadrotor UAVs in Complex Environments","authors":"Min Wang, H. Voos","doi":"10.1109/ICAR46387.2019.8981576","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981576","url":null,"abstract":"Current commercial UAVs are to a large extent remotely piloted by amateur human pilots. Due to lack of teleoperation experience or skills, they often drive the UAVs into collision. Therefore, in order to ensure safety of the UAV as well as its surroundings, it is necessary for the UAV to boast the capability of detecting emergency situation and acting on its own when facing imminent threat. However, the majority of UAVs currently available in the market are not equipped with such capability. To fill in the gap, in this paper we present a complete sense-and-avoid solution for assisting unskilled pilots in ensuring a safe flight. Particularly, we propose a novel nonlinear vehicle control system which takes into account of sensor characteristics, an emergency evaluation policy and a novel optimization-based avoidance control strategy. The effectiveness of the proposed approach is demonstrated and validated in simulation with multiple moving objects.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"25 1","pages":"529-534"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85425886","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/ICAR46387.2019.8981630
L. Rosset, Monika Florek-Jasinska, M. Suppa, M. Roa
Vision and proprioception are traditional sources of information for robotic grasping, but they are insufficient to achieve a stable grasp without slippage or without applying an excessive force on the object. Tactile sensors can aid in this problem by providing spatial and temporal data on the contact between fingertips and object. In this work, tactile fingertip sensors are used to detect slippage through two separate methods: the first, using principles inspired by human tactile sensing, and the second, by using a convolutional neural network trained with suitably labeled test samples. To perform a fair comparison of the methods, two evaluations are performed using a test bench and a pick-and-place robotic application. Results show promising use of the model-based method to avoid translational slippage, as it was able to consistently keep objects from slipping without overloading the grasp. Limitations of both model- and learning-based approaches are identified and discussed.
{"title":"Experimental study on model- vs. learning-based slip detection","authors":"L. Rosset, Monika Florek-Jasinska, M. Suppa, M. Roa","doi":"10.1109/ICAR46387.2019.8981630","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981630","url":null,"abstract":"Vision and proprioception are traditional sources of information for robotic grasping, but they are insufficient to achieve a stable grasp without slippage or without applying an excessive force on the object. Tactile sensors can aid in this problem by providing spatial and temporal data on the contact between fingertips and object. In this work, tactile fingertip sensors are used to detect slippage through two separate methods: the first, using principles inspired by human tactile sensing, and the second, by using a convolutional neural network trained with suitably labeled test samples. To perform a fair comparison of the methods, two evaluations are performed using a test bench and a pick-and-place robotic application. Results show promising use of the model-based method to avoid translational slippage, as it was able to consistently keep objects from slipping without overloading the grasp. Limitations of both model- and learning-based approaches are identified and discussed.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"105 1","pages":"493-500"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80639527","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：