Pub Date : 2021-04-16DOI: 10.1109/ICCRE51898.2021.9435671
Don Daven Christopher Trapal, Bryan Chia Chee Leong, Haw Wen Ng, J. Zhong, S. Srigrarom, Teng Hooi Chan
Detection and tracking of small and fast moving aerial targets especially drones has got attention nowadays. This paper focuses on the vision-based technique using images taken from observing cameras. In real life situations, the target drone objects maybe are hidden in the cluttered background such as trees shadow or foliage, rows of buildings, and other kind of scenery that will hinder the clear indication of the drones. The object is further confused by the presence of its shadow and reflection from the water or glass wall reflection. For vision-based objection, the clarity and the ambiguity of the target images in the video stream are the key for effective and successful target detection and tracking. Here, we present the improvement by mitigating the effect of cluttered background, shadow and water reflection to the target images. We applied the schemes to make the drone more visible and more clear. We also implemented super resolution to increase the image resolution for more precise detection and tracking. As a result, the target drone could be detected and tracked throughout the sample clips. The comparative tracking results using DCF are presented. Likewise, we applied the water removal scheme to eliminate the reflection to avoid confusion to the tracker. With this, just the correct drone targets were detected and tracked. Overall, the drones could be detected and tracked all the way, as long as they appeared in the camera scene.
{"title":"Improvement of Vision-based Drone Detection and Tracking by Removing Cluttered Background, Shadow and Water Reflection with Super Resolution","authors":"Don Daven Christopher Trapal, Bryan Chia Chee Leong, Haw Wen Ng, J. Zhong, S. Srigrarom, Teng Hooi Chan","doi":"10.1109/ICCRE51898.2021.9435671","DOIUrl":"https://doi.org/10.1109/ICCRE51898.2021.9435671","url":null,"abstract":"Detection and tracking of small and fast moving aerial targets especially drones has got attention nowadays. This paper focuses on the vision-based technique using images taken from observing cameras. In real life situations, the target drone objects maybe are hidden in the cluttered background such as trees shadow or foliage, rows of buildings, and other kind of scenery that will hinder the clear indication of the drones. The object is further confused by the presence of its shadow and reflection from the water or glass wall reflection. For vision-based objection, the clarity and the ambiguity of the target images in the video stream are the key for effective and successful target detection and tracking. Here, we present the improvement by mitigating the effect of cluttered background, shadow and water reflection to the target images. We applied the schemes to make the drone more visible and more clear. We also implemented super resolution to increase the image resolution for more precise detection and tracking. As a result, the target drone could be detected and tracked throughout the sample clips. The comparative tracking results using DCF are presented. Likewise, we applied the water removal scheme to eliminate the reflection to avoid confusion to the tracker. With this, just the correct drone targets were detected and tracked. Overall, the drones could be detected and tracked all the way, as long as they appeared in the camera scene.","PeriodicalId":382619,"journal":{"name":"2021 6th International Conference on Control and Robotics Engineering (ICCRE)","volume":"147 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129073900","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 : 2021-04-16DOI: 10.1109/ICCRE51898.2021.9435683
J. Parnklang, Peerapat Aroonrua
The designing of mosquitoes counting system instrument is presented in this work. The mosquitoes that were counted died in order not to measure duplicate counting data. As soon as the input source counting machine can detect the mosquito, the single trigger signal is transmitted to the IOT system to interrupt the server immediately. The number of real mosquito is not transmitting to the IOT but only a signal to interrupt the server.The server records the number of the interrupt signal with real-time clock. Then the interrupt information will be further handled. The front end counting machine consist of the high voltage generate with the suitable voltage value and electrode distance for the required mosquitoes size.The low trigger pulse signals of the mosquitoes killed by high voltage are sending to the controller unit. Immediately, interrupt counting signal of the number of mosquitoes is sent to the big stream data collection on IOT system by the time stamp technique. Form the measurement results, 10 live sample mosquitoes in a limited space box to fly though the counting machine show that the count results are 100 %correct count.
{"title":"Real Time Display on IoT Mosquitoes Count Detection","authors":"J. Parnklang, Peerapat Aroonrua","doi":"10.1109/ICCRE51898.2021.9435683","DOIUrl":"https://doi.org/10.1109/ICCRE51898.2021.9435683","url":null,"abstract":"The designing of mosquitoes counting system instrument is presented in this work. The mosquitoes that were counted died in order not to measure duplicate counting data. As soon as the input source counting machine can detect the mosquito, the single trigger signal is transmitted to the IOT system to interrupt the server immediately. The number of real mosquito is not transmitting to the IOT but only a signal to interrupt the server.The server records the number of the interrupt signal with real-time clock. Then the interrupt information will be further handled. The front end counting machine consist of the high voltage generate with the suitable voltage value and electrode distance for the required mosquitoes size.The low trigger pulse signals of the mosquitoes killed by high voltage are sending to the controller unit. Immediately, interrupt counting signal of the number of mosquitoes is sent to the big stream data collection on IOT system by the time stamp technique. Form the measurement results, 10 live sample mosquitoes in a limited space box to fly though the counting machine show that the count results are 100 %correct count.","PeriodicalId":382619,"journal":{"name":"2021 6th International Conference on Control and Robotics Engineering (ICCRE)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130224821","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 : 2021-04-16DOI: 10.1109/ICCRE51898.2021.9435736
Baoshuang Ge, Hai Zhang, Dapeng Wang, Hongliang Huang
Dead-reckoning plays an important role in mobile robot’s navigation. Inertial devices as the core sensors for dead-reckoning can obtain the complete navigation states. However, the accuracy of these navigation states suffers from a decline over time. Thus, the alignment of inertial device must be performed before navigation. The initial states of robot and the biases of inertial devices are the basic parameters that need to be aligned. Usually, the initialization of robot’s position and velocity can be accomplished through external measurements, while attitude initialization relies on special alignment algorithms. In this paper, an effective alignment algorithm is designed for a mobile robot used in practical applications, in which both the robot’s initial attitude and inertial devices’ biases are considered. The measurements collected from a barometer and an odometer are used for alignment. The experimental results show that the proposed method can serve as an effective alignment for mobile robot’s navigation.
{"title":"An Online Alignment Algorithm for Inertial Devices Applied to Mobile Robot Navigation","authors":"Baoshuang Ge, Hai Zhang, Dapeng Wang, Hongliang Huang","doi":"10.1109/ICCRE51898.2021.9435736","DOIUrl":"https://doi.org/10.1109/ICCRE51898.2021.9435736","url":null,"abstract":"Dead-reckoning plays an important role in mobile robot’s navigation. Inertial devices as the core sensors for dead-reckoning can obtain the complete navigation states. However, the accuracy of these navigation states suffers from a decline over time. Thus, the alignment of inertial device must be performed before navigation. The initial states of robot and the biases of inertial devices are the basic parameters that need to be aligned. Usually, the initialization of robot’s position and velocity can be accomplished through external measurements, while attitude initialization relies on special alignment algorithms. In this paper, an effective alignment algorithm is designed for a mobile robot used in practical applications, in which both the robot’s initial attitude and inertial devices’ biases are considered. The measurements collected from a barometer and an odometer are used for alignment. The experimental results show that the proposed method can serve as an effective alignment for mobile robot’s navigation.","PeriodicalId":382619,"journal":{"name":"2021 6th International Conference on Control and Robotics Engineering (ICCRE)","volume":"148 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134232069","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 : 2021-04-16DOI: 10.1109/ICCRE51898.2021.9435662
Sheung-Lai Lo, K. Leung, Y. Leung
A multiple sensors platform is a commonly-used method for robotic application which can capture different types of environmental data. However, the calibration process through finding the geometric relationships is needed in order to integrate different sensors’ data together in the same coordination system representation. In this study, we proposed and implemented a simple and yet accurate calibration method for estimating the extrinsic parameters in a transformation matrix representation between more than one sensor with 3D Point-Cloud data. Our calibration method uses both Kinect and LiDAR Point-Cloud scans based on a cuboid as a calibration object, which obtains efficient information with high-level feature points. Such extrinsic parameters can be distinguished with a single shot of capturing the Point-Cloud frame. To verify whether or not our method is practical in a real-life scenario, we evaluated the method and obtained the data using a homemade robot. With our proposed calibration methods, it enables a possibility to exchange the data between LiDAR and Kinect simply and robustly.
{"title":"Evaluating the LiDAR and Kinect Calibration Methods and Application","authors":"Sheung-Lai Lo, K. Leung, Y. Leung","doi":"10.1109/ICCRE51898.2021.9435662","DOIUrl":"https://doi.org/10.1109/ICCRE51898.2021.9435662","url":null,"abstract":"A multiple sensors platform is a commonly-used method for robotic application which can capture different types of environmental data. However, the calibration process through finding the geometric relationships is needed in order to integrate different sensors’ data together in the same coordination system representation. In this study, we proposed and implemented a simple and yet accurate calibration method for estimating the extrinsic parameters in a transformation matrix representation between more than one sensor with 3D Point-Cloud data. Our calibration method uses both Kinect and LiDAR Point-Cloud scans based on a cuboid as a calibration object, which obtains efficient information with high-level feature points. Such extrinsic parameters can be distinguished with a single shot of capturing the Point-Cloud frame. To verify whether or not our method is practical in a real-life scenario, we evaluated the method and obtained the data using a homemade robot. With our proposed calibration methods, it enables a possibility to exchange the data between LiDAR and Kinect simply and robustly.","PeriodicalId":382619,"journal":{"name":"2021 6th International Conference on Control and Robotics Engineering (ICCRE)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133974731","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 : 2021-04-16DOI: 10.1109/ICCRE51898.2021.9435682
J. Parnklang, Thitirat Kasomkul, Chutikarn Wiangwong, Natcha Pibanpaknitee
The design and model of greenhouse daytime constant light source spectrum from sunlight and RGB LED controlled is present in this title. According to the light color requirement that the special plant need to grow are different in each type of plant. When the requirement of plant are known, we have designed constant frequency spectrum of light in the greenhouse throughout the daytime. The design system is based on solar light which is difference in the color and intensity of light on the daytime. Light frequency adjustment system interferes with the solar spectrum in the greenhouse. The system stops interfering when the optical spectrum is the programed designed wavelength of light. The light spectrum of each color (red, blue and green) measured is normalized to realize the proportion of each color that need to intervene. The example of a plant used in a system test is a cannabis an economic crop in which each pre-harvest period requires a different spectrum ratio color of light. Other variables of plant growth be controlled as well. The sunlight that can pass through the outdoor greenhouse has already been filtered from UV and IR rays. The experimental results show the higher grow rate of the marijuana that plant in the adjust spectrum greenhouse is 12%. We also found that light spectral-optimized cannabis get a number leaf density greater than 30% compare to being grown in an uncontrolled nearby outdoor greenhouse.
{"title":"Daytime Solar Greenhouse Plant Light Spectum Control System","authors":"J. Parnklang, Thitirat Kasomkul, Chutikarn Wiangwong, Natcha Pibanpaknitee","doi":"10.1109/ICCRE51898.2021.9435682","DOIUrl":"https://doi.org/10.1109/ICCRE51898.2021.9435682","url":null,"abstract":"The design and model of greenhouse daytime constant light source spectrum from sunlight and RGB LED controlled is present in this title. According to the light color requirement that the special plant need to grow are different in each type of plant. When the requirement of plant are known, we have designed constant frequency spectrum of light in the greenhouse throughout the daytime. The design system is based on solar light which is difference in the color and intensity of light on the daytime. Light frequency adjustment system interferes with the solar spectrum in the greenhouse. The system stops interfering when the optical spectrum is the programed designed wavelength of light. The light spectrum of each color (red, blue and green) measured is normalized to realize the proportion of each color that need to intervene. The example of a plant used in a system test is a cannabis an economic crop in which each pre-harvest period requires a different spectrum ratio color of light. Other variables of plant growth be controlled as well. The sunlight that can pass through the outdoor greenhouse has already been filtered from UV and IR rays. The experimental results show the higher grow rate of the marijuana that plant in the adjust spectrum greenhouse is 12%. We also found that light spectral-optimized cannabis get a number leaf density greater than 30% compare to being grown in an uncontrolled nearby outdoor greenhouse.","PeriodicalId":382619,"journal":{"name":"2021 6th International Conference on Control and Robotics Engineering (ICCRE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129491230","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 : 2021-04-16DOI: 10.1109/ICCRE51898.2021.9435726
Renxu Zhang, Chengleng Han
In order to explore the working principle of the arm torsion Electromagnetic Active Suspension (ITEAS), this paper studies the three functions of height adjustment, stiffness adjustment and damping adjustment based on the basic structure of ITEAS. Firstly, the mathematical model of height control of ITEAS is built according to the working principle, and then the ITEAS is established according to the tire mass and the motion form of the body Based on the two degree of freedom 1/4 suspension model, the stiffness active adjustment model and damping active adjustment model are built respectively, and verified in the MATLAB / Simulink module. The results show that ITEAS can effectively realize the stiffness active adjustment and damping active adjustment according to the road excitation.
{"title":"Research on the Principle of Arm Torsion Electromagnetic Active Suspensison","authors":"Renxu Zhang, Chengleng Han","doi":"10.1109/ICCRE51898.2021.9435726","DOIUrl":"https://doi.org/10.1109/ICCRE51898.2021.9435726","url":null,"abstract":"In order to explore the working principle of the arm torsion Electromagnetic Active Suspension (ITEAS), this paper studies the three functions of height adjustment, stiffness adjustment and damping adjustment based on the basic structure of ITEAS. Firstly, the mathematical model of height control of ITEAS is built according to the working principle, and then the ITEAS is established according to the tire mass and the motion form of the body Based on the two degree of freedom 1/4 suspension model, the stiffness active adjustment model and damping active adjustment model are built respectively, and verified in the MATLAB / Simulink module. The results show that ITEAS can effectively realize the stiffness active adjustment and damping active adjustment according to the road excitation.","PeriodicalId":382619,"journal":{"name":"2021 6th International Conference on Control and Robotics Engineering (ICCRE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129740558","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}
The spherical robot was a kind of mobile robot which used the spherical shell to roll to realize its own motion. It had the characteristics of a small turning radius and omnidirectional motion. Besides the structure of the spherical robot was multitudinous, a kind of two degrees of freedom of the oscillating XK-II spherical robot was designed in this paper. Likewise, the motion principle and the structure of the XK-II spherical robot were introduced in this paper. The relation of speed vector projection and the description method of Euler Angle were used in the paper to establish the two inputs and five outputs kinematics equation of the XK-II type spherical robot. Then, the kinematics equation was solved by using the four order-five order Runge-Kutta algorithm, at the same time, the trajectory tracking controller of line, roundness, sinusoid and Lissajous figures were designed by combining PID. Finally, the kinematics model was validated by simulation and real experiment, and the trajectory tracking control of the spherical robot was realized effectively.
{"title":"Research on the Trajectory Tracking Control of a Swinging Spherical Robot","authors":"Guiyang Cui, Haiyan Yang, Kaifeng Xiong, Hua Zhang, Mingming Guo","doi":"10.1109/ICCRE51898.2021.9435651","DOIUrl":"https://doi.org/10.1109/ICCRE51898.2021.9435651","url":null,"abstract":"The spherical robot was a kind of mobile robot which used the spherical shell to roll to realize its own motion. It had the characteristics of a small turning radius and omnidirectional motion. Besides the structure of the spherical robot was multitudinous, a kind of two degrees of freedom of the oscillating XK-II spherical robot was designed in this paper. Likewise, the motion principle and the structure of the XK-II spherical robot were introduced in this paper. The relation of speed vector projection and the description method of Euler Angle were used in the paper to establish the two inputs and five outputs kinematics equation of the XK-II type spherical robot. Then, the kinematics equation was solved by using the four order-five order Runge-Kutta algorithm, at the same time, the trajectory tracking controller of line, roundness, sinusoid and Lissajous figures were designed by combining PID. Finally, the kinematics model was validated by simulation and real experiment, and the trajectory tracking control of the spherical robot was realized effectively.","PeriodicalId":382619,"journal":{"name":"2021 6th International Conference on Control and Robotics Engineering (ICCRE)","volume":"95 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129939815","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 : 2021-04-16DOI: 10.1109/ICCRE51898.2021.9435677
Yingzhe Sun, Qifeng Zhang, Aiqun Zhang
The obstacle avoidance method and its effect of underwater hexapod robot is one of the important aspects of whether the underwater hexapod robot can work smoothly. As a simple and real-time obstacle avoidance method, artificial potential method has been widely concerned. Based on the improved artificial potential method, the motion planning of underwater hexapod robot is studied in this paper. The three-dimensional information of obstacles in the working area is detected by scanning sonar with orthogonal arrangement, and the obstacles are abstracted into a circle to facilitate the construction of obstacle avoidance model. The improved physical model of artificial potential method is established by optimizing the functional equations of gravitational potential field and repulsive potential field, and the real-time crawling path of the underwater hexapod robot is obtained. On this basis, the realtime crawling path of the underwater hexapod robot is constructed for the realization of obstacle avoidance. Finally, the motion planning effect of the underwater hexapod robot is verified by simulation. The simulation results show that the motion planning of the underwater hexapod robot based on the improved artificial potential method can avoid obstacles well. The path generated is smooth and the crawling efficiency is more efficient. It can make the underwater hexapod robot maintain a safe crawling distance and avoid obstacles to reach the end point smoothly, which has good practical application value.
{"title":"Research and Simulation of Motion Planning for Underwater Hexapod Robot Based on Improved Artificial Potential Method","authors":"Yingzhe Sun, Qifeng Zhang, Aiqun Zhang","doi":"10.1109/ICCRE51898.2021.9435677","DOIUrl":"https://doi.org/10.1109/ICCRE51898.2021.9435677","url":null,"abstract":"The obstacle avoidance method and its effect of underwater hexapod robot is one of the important aspects of whether the underwater hexapod robot can work smoothly. As a simple and real-time obstacle avoidance method, artificial potential method has been widely concerned. Based on the improved artificial potential method, the motion planning of underwater hexapod robot is studied in this paper. The three-dimensional information of obstacles in the working area is detected by scanning sonar with orthogonal arrangement, and the obstacles are abstracted into a circle to facilitate the construction of obstacle avoidance model. The improved physical model of artificial potential method is established by optimizing the functional equations of gravitational potential field and repulsive potential field, and the real-time crawling path of the underwater hexapod robot is obtained. On this basis, the realtime crawling path of the underwater hexapod robot is constructed for the realization of obstacle avoidance. Finally, the motion planning effect of the underwater hexapod robot is verified by simulation. The simulation results show that the motion planning of the underwater hexapod robot based on the improved artificial potential method can avoid obstacles well. The path generated is smooth and the crawling efficiency is more efficient. It can make the underwater hexapod robot maintain a safe crawling distance and avoid obstacles to reach the end point smoothly, which has good practical application value.","PeriodicalId":382619,"journal":{"name":"2021 6th International Conference on Control and Robotics Engineering (ICCRE)","volume":"91 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122811984","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 : 2021-04-16DOI: 10.1109/ICCRE51898.2021.9435660
Yuyu Zhao, Shaowu Yang
Visual object tracking plays an important role in military guidance, human-computer interaction, and robot visual navigation. With the high performance and real-time speed, the Siamese network has become popular in recent years, which localizes the target by comparing the similarity of the appearance template and the candidate boxes in the search region. However, the appearance template is only extracted from the current frame, which results in the missing of the target for constantly changing appearance. In this paper, we propose a template extractor that can capture the latest appearance features from the previously predicted templates, named PlasticNet. We take inspiration from the memory mechanism of neuroscience (synaptic plasticity): the connections between neurons will be enhanced when they are stimulated at the same time. We combined it with recurrent networks to realize the PlasticNet. Our method can easily be integrated into existing siamese trackers. Our proposed model is applied in SiamRPN and improved performance. Extensive experiments on OTB2015, VOT2018, VOT2016 datasets demonstrate that our PlasticNet can effectively adapt to appearance changes.
{"title":"PlasticNet: A Memory Template Extractor with Plastic Neural Networks for Object Tracking","authors":"Yuyu Zhao, Shaowu Yang","doi":"10.1109/ICCRE51898.2021.9435660","DOIUrl":"https://doi.org/10.1109/ICCRE51898.2021.9435660","url":null,"abstract":"Visual object tracking plays an important role in military guidance, human-computer interaction, and robot visual navigation. With the high performance and real-time speed, the Siamese network has become popular in recent years, which localizes the target by comparing the similarity of the appearance template and the candidate boxes in the search region. However, the appearance template is only extracted from the current frame, which results in the missing of the target for constantly changing appearance. In this paper, we propose a template extractor that can capture the latest appearance features from the previously predicted templates, named PlasticNet. We take inspiration from the memory mechanism of neuroscience (synaptic plasticity): the connections between neurons will be enhanced when they are stimulated at the same time. We combined it with recurrent networks to realize the PlasticNet. Our method can easily be integrated into existing siamese trackers. Our proposed model is applied in SiamRPN and improved performance. Extensive experiments on OTB2015, VOT2018, VOT2016 datasets demonstrate that our PlasticNet can effectively adapt to appearance changes.","PeriodicalId":382619,"journal":{"name":"2021 6th International Conference on Control and Robotics Engineering (ICCRE)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131817479","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 : 2021-04-16DOI: 10.1109/ICCRE51898.2021.9435666
Seongin Na, Hanlin Niu, B. Lennox, F. Arvin
Pheromone-based communication has been adopted into robotic systems inspired by social insects as an alternative communication strategy, promising for dynamic and complex environments. For this reason, artificial pheromone communication system for robotic systems, especially for swarm robotic systems, has been proposed with diverse mediums such as light and virtual environment. However, the low versatility of each method makes it difficult to utilise the benefits of pheromone-based communication in diverse robotic platforms and environments. In this paper, we proposed PhERS (Pheromone for Every RobotS) framework designed to increase versatility, aiming to boost research and applications of pheromone-based communication in robotics. To validate the framework, we conducted experiments with simulated robots manoeuvred by hand-tuned controller performing navigation and collision avoidance tasks. As another contribution, we proposed a novel Deep Reinforcement Learning (DRL)-based controller for robots utilising pheromones to overcome the limitations of hand-tuned controller. Experiments and observed results demonstrated the feasibility of using the proposed framework in a robotic scenario, showing that DRLbased controller outperforms the baseline hand-tuned controller in a dynamic environment.
受群居昆虫的启发，机器人系统采用了基于信息素的通信作为一种替代通信策略，有望在动态和复杂的环境中应用。因此，针对机器人系统特别是群体机器人系统的人工信息素通信系统被提出了多种介质，如光和虚拟环境。然而，每种方法的低通用性使得难以在不同的机器人平台和环境中利用基于信息素的通信的好处。在本文中，我们提出了PhERS (Pheromone for Every RobotS)框架，旨在增加多功能性，旨在促进基于信息素的机器人通信的研究和应用。为了验证该框架，我们对模拟机器人进行了实验，这些机器人由手动调谐控制器操纵，执行导航和避碰任务。作为另一个贡献，我们提出了一种新的基于深度强化学习(DRL)的机器人控制器，利用信息素来克服手动调谐控制器的局限性。实验和观察结果证明了在机器人场景中使用该框架的可行性，表明基于drl的控制器在动态环境中优于基线手动调谐控制器。
{"title":"Universal Artificial Pheromone Framework with Deep Reinforcement Learning for Robotic Systems","authors":"Seongin Na, Hanlin Niu, B. Lennox, F. Arvin","doi":"10.1109/ICCRE51898.2021.9435666","DOIUrl":"https://doi.org/10.1109/ICCRE51898.2021.9435666","url":null,"abstract":"Pheromone-based communication has been adopted into robotic systems inspired by social insects as an alternative communication strategy, promising for dynamic and complex environments. For this reason, artificial pheromone communication system for robotic systems, especially for swarm robotic systems, has been proposed with diverse mediums such as light and virtual environment. However, the low versatility of each method makes it difficult to utilise the benefits of pheromone-based communication in diverse robotic platforms and environments. In this paper, we proposed PhERS (Pheromone for Every RobotS) framework designed to increase versatility, aiming to boost research and applications of pheromone-based communication in robotics. To validate the framework, we conducted experiments with simulated robots manoeuvred by hand-tuned controller performing navigation and collision avoidance tasks. As another contribution, we proposed a novel Deep Reinforcement Learning (DRL)-based controller for robots utilising pheromones to overcome the limitations of hand-tuned controller. Experiments and observed results demonstrated the feasibility of using the proposed framework in a robotic scenario, showing that DRLbased controller outperforms the baseline hand-tuned controller in a dynamic environment.","PeriodicalId":382619,"journal":{"name":"2021 6th International Conference on Control and Robotics Engineering (ICCRE)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125106058","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：