{"title":"A Mobile Interactive Robot for Social Distancing in Hospitals","authors":"M. Virgolin, M. Bellone, Krister Wolff, M. Wahde","doi":"10.1109/IRC52146.2021.00020","DOIUrl":"https://doi.org/10.1109/IRC52146.2021.00020","url":null,"abstract":"We introduce the multimodal interactive mobile robot ISOLDE, intended for use in hospitals, with the primary aim of helping healthcare staff to maintain social distancing during pandemics, such as the ongoing Covid-19 pandemic. ISOLDE also addresses the growing concern related to the use of black box models in artificial intelligence, especially in situations involving high-stakes decisions. Thus, ISOLDE’s interactive capabilities have been implemented using a fully interpretable dialogue manager, making it easy to monitor and, if needed, correct the robot’s actions, even for a non-expert. A use case is presented (in a laboratory setting) in which the robot successfully interacts with healthcare staff to carry out a requested transportation and delivery task, and also measuring a patient’s temperature.","PeriodicalId":439358,"journal":{"name":"2021 Fifth IEEE International Conference on Robotic Computing (IRC)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123451969","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}
Mrinal Senapati, Bhaskar Anand, A. Thakur, Harshal Verma, P. Rajalakshmi
{"title":"Object Detection and Segmentation using LiDAR-Camera Fusion for Autonomous Vehicle","authors":"Mrinal Senapati, Bhaskar Anand, A. Thakur, Harshal Verma, P. Rajalakshmi","doi":"10.1109/IRC52146.2021.00029","DOIUrl":"https://doi.org/10.1109/IRC52146.2021.00029","url":null,"abstract":"The Light detection and ranging (LiDAR) sensor plays a crucial role in perceiving the environment for an autonomous vehicle. But, in many scenarios LiDAR is unable to capture important information, for example, traffic light signals. This kind of scenario can be avoided by using camera images with LiDAR data. But, the system will not work effectively, if there is no proper calibration and synchronization between camera images and LiDAR data. In this paper, we have demonstrated a system, where objects are synchronously detected and segmented in both images and LiDAR data from KITTI datasets. Currently, the system is working in real-time using Robot Operating System (ROS) and can process up to 10 frames of image and point cloud data per second.","PeriodicalId":439358,"journal":{"name":"2021 Fifth IEEE International Conference on Robotic Computing (IRC)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129525495","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}
{"title":"Integration of Scene Image and Conversational Text for Human-Robot Dialogue","authors":"Hao-Yi Wang, Jhih-Yuan Huang, Wei-Po Lee","doi":"10.1109/IRC52146.2021.00013","DOIUrl":"https://doi.org/10.1109/IRC52146.2021.00013","url":null,"abstract":"In recent years, it becomes a trend to build social dialogue systems to achieve human-robot interaction. While current conversational systems mostly focus on the dialogue utterances, visual context also plays an important role in determining the relevance of the system responses to the dialogue. In this study, we develop an integrated approach to investigate how to achieve such a visual-language task. Our approach takes both scene images and dialogue utterances into account and trains a neural model to generate machine responses of natural language. We have conducted a series of experiments to evaluate the presented approach, and the results confirm its usefulness and effectiveness.","PeriodicalId":439358,"journal":{"name":"2021 Fifth IEEE International Conference on Robotic Computing (IRC)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127821360","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}
{"title":"A knowledge base for robots to model the real-world as a hypergraph","authors":"J. Sijs, J. Fletcher","doi":"10.1109/IRC52146.2021.00027","DOIUrl":"https://doi.org/10.1109/IRC52146.2021.00027","url":null,"abstract":"Robotic systems operating in the real world would benefit from a clear semantic model to understand the real world. Such semantics are typically captured in a knowledge structure. The speed at which robots need to update their real-world knowledge (because of new sensory observations) prevented the use of complex knowledge structures, due to which they typically rely on hierarchical structures and triples. However, the shift from automated to autonomous robotics implied that more knowledge about the real-world is necessary, thereby requiring many work-arounds to maintain such a hierarchical knowledge base. To remove such work-arounds, this article adopts the richer hypergraph model for creating an actual knowledge base implemented on a real robotic system. Experiments show that the use of a hypergraph, without any work-arounds, still allows real-time updates of the knowledge-base with actual information extracted from the robot’s sensors.","PeriodicalId":439358,"journal":{"name":"2021 Fifth IEEE International Conference on Robotic Computing (IRC)","volume":"99 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121577621","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}
{"title":"Applying a Virtual Tool Center Point Calibration and Position Teaching System of Robotic Arm to Automatically Detect Displays","authors":"C. Mai, C. K. Huang, H. Hocheng","doi":"10.1109/IRC52146.2021.00034","DOIUrl":"https://doi.org/10.1109/IRC52146.2021.00034","url":null,"abstract":"The precision of robotic arms is affected by the accuracy of the locations of tools and workpieces. To ensure the precision of robotic arms, they must be calibrated with the locations of tools and workpieces. To date, calibration methods for tools typically rely on recognition of physical characteristic points. Consequently, these methods cannot be applied to situations in which no physical tool center point is present such as noncontact detection or spray coating. To solve this problem, an automatic calibration and robot position teaching system for virtual tool center points was proposed, and the system was applied for display detection. The system utilized triangulation ranging and a visual servo for calibration and robot position teaching, solving the problem of noncontact applications of virtual tool center points, and eliminating reliance on manpower for robot position teaching. If the automatic teaching method is used, operator experience is unnecessary. Therefore, training or recruitment of skilled manufacturing workers is unnecessary, improving the smart manufacturing process and benefitting industry.","PeriodicalId":439358,"journal":{"name":"2021 Fifth IEEE International Conference on Robotic Computing (IRC)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128452753","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}
Christopher Allred, M. Russell, Mario Harper, J. Pusey
{"title":"Improving Methods for Multi-Terrain Classification Beyond Visual Perception","authors":"Christopher Allred, M. Russell, Mario Harper, J. Pusey","doi":"10.1109/IRC52146.2021.00022","DOIUrl":"https://doi.org/10.1109/IRC52146.2021.00022","url":null,"abstract":"Terrain classification in mixed-surface unstructured environments is key for safe navigation, energy efficiency, and anticipating motion volatility. This is particularly true for dynamically moving legged platforms which are highly impacted by foot ground interactions. This research demonstrates terrain classification using a long short-term memory (LSTM) model trained on actuator time series data, particularly the difference in center-of-pressure (COP) and leg forces. The LSTM COPForce model showed a 97.5% accuracy in classification on three outdoor surfaces with small amounts of data and no additional sensors.","PeriodicalId":439358,"journal":{"name":"2021 Fifth IEEE International Conference on Robotic Computing (IRC)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130796659","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}
{"title":"Hierarchical Task and Motion Planning through Deep Reinforcement Learning","authors":"Abdullah Al Redwan Newaz, Tauhidul Alam","doi":"10.1109/IRC52146.2021.00023","DOIUrl":"https://doi.org/10.1109/IRC52146.2021.00023","url":null,"abstract":"Task and motion planning (TAMP) integrates the generation of high-level tasks in a discrete space and the execution of low-level actions in a continuous space. Such planning integration is susceptible to uncertainties and computationally challenging as low-level actions should be verified for high-level tasks. Therefore, this paper presents a hierarchical task and motion planning method under uncertainties. We utilize Markov Decision Processes (MDPs) to model task and motion planning in a stochastic environment. The motion planner handles motion uncertainty and leverages physical constraints to synthesize an optimal low-level control policy for a single robot to generate motions in continuous action and state spaces. Given the optimal control policy for multiple homogeneous robots, the task planner synthesizes an optimal high-level tasking policy in discrete task and state spaces addressing both task and motion uncertainties. Our optimal tasking and control policies are synthesized through deep reinforcement learning algorithms. The performance of our method is validated in realistic physics-based simulations with two quadrotors transporting objects in a warehouse setting.","PeriodicalId":439358,"journal":{"name":"2021 Fifth IEEE International Conference on Robotic Computing (IRC)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126140663","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}
{"title":"Learning More Complex Actions with Deep Reinforcement Learning","authors":"Chenxi Wang, Youtian Du, Shengyuan Xie, Yongdi Lu","doi":"10.1109/IRC52146.2021.00028","DOIUrl":"https://doi.org/10.1109/IRC52146.2021.00028","url":null,"abstract":"How to learn more complex actions based on the predefined atomic actions is a significant problem for improving the intelligence of machines. In this paper, we present a novel approach called action space expanding Q-learning (ASE-Q) to the learning of more complex actions. We first introduce two types of formation of the complex actions; then, we design an action combination network to dynamically learn the complex actions a Q-network to evaluate action-value of the complex actions. Learning more complex actions can speed up the process that an agent accomplishes a task and improve its capacity in learning new actions.","PeriodicalId":439358,"journal":{"name":"2021 Fifth IEEE International Conference on Robotic Computing (IRC)","volume":"71 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116965084","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}
{"title":"Vibration Control for Pivoting by Robot Hand Equipped with CAVS and FingerVision","authors":"Yoshiyuki Suzuki, Akihiko Yamaguchi, Seita Nojiri, Tetsuyou Watanabe, Koichi Hashimoto","doi":"10.1109/IRC52146.2021.00010","DOIUrl":"https://doi.org/10.1109/IRC52146.2021.00010","url":null,"abstract":"We explore in-hand manipulation, especially pivoting, with a robot hand equipped with a contact area variable surface (CAVS) and a vision-based tactile sensor FingerVision. CAVS is a skin for robots where the surface friction coefficient passively changes according to the external force; during holding objects with a small force, the friction coefficient is low, while the friction coefficient becomes higher when increasing the holding force. FingerVision provides multimodal sensations, including force and slip distributions, and orientation of grasped objects. On our robot hand, CAVS is embedded on the surface of the fingertips, and FingerVision is installed under the CAVS skin. With this hand, we expect a gentle in-hand manipulation of objects where we properly control objects by slipping them on purpose to achieve a desired operation, however an adequate control method for such operations has been unclear. This paper focuses on the control methods of the CAVS+FingerVision hand for pivoting objects as an example of in-hand manipulation. Through the empirical comparisons, we found that a vibration control of fingertips achieves the best accuracy and outperforms a simple fuzzy control, and CAVS outperforms a usual flat skin.","PeriodicalId":439358,"journal":{"name":"2021 Fifth IEEE International Conference on Robotic Computing (IRC)","volume":"151 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116252863","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}
{"title":"An Improved Approach for Inverse Kinematics and Motion Planning of an Industrial Robot Manipulator with Reinforcement Learning","authors":"J. Weber, M. Schmidt","doi":"10.1109/IRC52146.2021.00009","DOIUrl":"https://doi.org/10.1109/IRC52146.2021.00009","url":null,"abstract":"Robot manipulators have become very popular in many application areas in the last decades. Typically, simplified robot models are used to apply machine learning algorithms, so the particular challenges of using real robots are not taken into account (e.g., joint angle constraints). This work contributes a new approach for controlling manipulators that solves both the inverse kinematics problem and the path planning problem.The paper uses a Deep Deterministic Policy Gradient (DDPG) agent to learn the motion of a robot manipulator designed for industrial use. It introduces a new state space to reliably learn general motions with user-selectable start and target positions with a high success rate. A new reward function improves the robot’s path to the target position. In addition, the learned motions result in densely occupied paths, so that the presented learning approach can also be used for path planning of robot manipulators.The new approach outperforms other learned state-of-the-art approaches for solving both the problems of inverse kinematics and of path planning of arm-like robots in terms of success rate and mean error. The approach was applied to a physical robot device and can be easily transferred to other path planning problems.","PeriodicalId":439358,"journal":{"name":"2021 Fifth IEEE International Conference on Robotic Computing (IRC)","volume":"122 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133434217","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}