Refine Search

New Search

Results: 38

(searched for: Target Tracking Using Reinforcement Learning and Neural Networks)
Save to Scifeed
Page of 1
Articles per Page
by
Show export options
  Select all
Panbo He, , Kaijun Liu, Neal N. Xiong
Transactions on Petri Nets and Other Models of Concurrency XV pp 226-235; doi:10.1007/978-3-030-74717-6_24

The publisher has not yet granted permission to display this abstract.
Yuxiang Yang, , Dongdong Wang, Shunli Zhang, Qi Yu, Liqiang Wang
Published: 1 April 2021
Neurocomputing; doi:10.1016/j.neucom.2021.03.118

The publisher has not yet granted permission to display this abstract.
Fufeng Qiao
Published: 9 March 2021
PLOS ONE, Volume 16; doi:10.1371/journal.pone.0245259

Abstract:
A DCNN-LSTM (Deep Convolutional Neural Network-Long Short Term Memory) model is proposed to recognize and track table tennis’s real-time trajectory in complex environments, aiming to help the audiences understand competition details and provide a reference for training enthusiasts using computers. Real-time motion features are extracted via deep reinforcement networks. DCNN tracks the recognized objects, and the LSTM algorithm predicts the ball’s trajectory. The model is tested on a self-built video dataset and existing systems and compared with other algorithms to verify its effectiveness. Finally, an overall tactical detection system is built to measure ball rotation and predict ball trajectory. Results demonstrate that in feature extraction, the Deep Deterministic Policy Gradient (DDPG) algorithm has the best performance, with a maximum accuracy rate of 89% and a minimum mean square error of 0.2475. The accuracy of target tracking effect and trajectory prediction is as high as 90%. Compared with traditional methods, the performance of the DCNN-LSTM model based on deep learning is improved by 23.17%. The implemented automatic detection system of table tennis tactical indicators can deal with the problems of table tennis tracking and rotation measurement. It can provide a theoretical foundation and practical value for related research in real-time dynamic detection of balls.
Published: 4 February 2021
Sensors, Volume 21; doi:10.3390/s21041076

Abstract:
Unmanned aerial vehicles (UAVs) have been widely used in search and rescue (SAR) missions due to their high flexibility. A key problem in SAR missions is to search and track moving targets in an area of interest. In this paper, we focus on the problem of Cooperative Multi-UAV Observation of Multiple Moving Targets (CMUOMMT). In contrast to the existing literature, we not only optimize the average observation rate of the discovered targets, but we also emphasize the fairness of the observation of the discovered targets and the continuous exploration of the undiscovered targets, under the assumption that the total number of targets is unknown. To achieve this objective, a deep reinforcement learning (DRL)-based method is proposed under the Partially Observable Markov Decision Process (POMDP) framework, where each UAV maintains four observation history maps, and maps from different UAVs within a communication range can be merged to enhance UAVs’ awareness of the environment. A deep convolutional neural network (CNN) is used to process the merged maps and generate the control commands to UAVs. The simulation results show that our policy can enable UAVs to balance between giving the discovered targets a fair observation and exploring the search region compared with other methods.
Ignace Ransquin, Philippe Chatelain
AIAA Scitech 2021 Forum; doi:10.2514/6.2021-0885

The publisher has not yet granted permission to display this abstract.
Jezuina Koroveshi, Ana Ktona
European Journal of Engineering and Technology Research, Volume 6, pp 48-54; doi:10.24018/ejers.2021.6.1.2316

Abstract:
Target tracking is a process that may find applications in different domains such as video surveillance, robot navigation and human computer interaction. In this work we have considered the problem of tracking a moving object in a multi agent environment. The environment is a rectangular space bounded by walls. The first agent is the target and it moves randomly in the space. The second agent should follow the target, keeping as close as possible without crashing with it. It uses sensors to detect the position of the target. The sensor readings give the distance and the angle from the target. We use reinforcement learning to train the tracker to detect any change in the movement of the target and stay within a certain range from it. Reinforcement learning is a form of machine learning in which the agent learns by interacting with the environment. By doing so, for each action taken, the agent receives a reward from the environment, which is used to determine positive or negative behaviour. The goal of the agent is to maximise the total reward received during the interaction. This form of machine learning has applications in different areas, such as: game solving with the most known game being AlphaGO; robotics, for design of hard-to engineer behaviours; traffic light control, personalized recommendations, etc. The sensor readings may have continuous values, making a very large state space. We approximate the value function using neural networks and use different reward functions for learning the best policy.
Brittany Moore, Sheng Khang,
Frontiers in Behavioral Neuroscience, Volume 14; doi:10.3389/fnbeh.2020.541920

Abstract:
Reward modulation is represented in the motor cortex (M1) and could be used to implement more accurate decoding models to improve brain-computer interfaces (BCIs; Zhao et al., 2018). Analyzing trial-to-trial noise-correlations between neural units in the presence of rewarding (R) and non-rewarding (NR) stimuli adds to our understanding of cortical network dynamics. We utilized Pearson’s correlation coefficient to measure shared variability between simultaneously recorded units (32–112) and found significantly higher noise-correlation and positive correlation between the populations’ signal- and noise-correlation during NR trials as compared to R trials. This pattern is evident in data from two non-human primates (NHPs) during single-target center out reaching tasks, both manual and action observation versions. We conducted a mean matched noise-correlation analysis to decouple known interactions between event-triggered firing rate changes and neural correlations. Isolated reward discriminatory units demonstrated stronger correlational changes than units unresponsive to reward firing rate modulation, however, the qualitative response was similar, indicating correlational changes within the network as a whole can serve as another information channel to be exploited by BCIs that track the underlying cortical state, such as reward expectation, or attentional modulation. Reward expectation and attention in return can be utilized with reinforcement learning (RL) towards autonomous BCI updating.
Emilio Capo, Daniele Loiacono
2020 IEEE Symposium Series on Computational Intelligence (SSCI) pp 2327-2334; doi:10.1109/ssci47803.2020.9308138

The publisher has not yet granted permission to display this abstract.
, Yu Lu, Xiaocheng Liu,
Published: 1 December 2020
Ocean Engineering, Volume 217; doi:10.1016/j.oceaneng.2020.107704

The publisher has not yet granted permission to display this abstract.
Guohong Xiong, Lu Dong
2020 Chinese Automation Congress (CAC) pp 2682-2686; doi:10.1109/cac51589.2020.9326946

The publisher has not yet granted permission to display this abstract.
Published: 3 November 2020
Applied Sciences, Volume 10; doi:10.3390/app10217780

Abstract:
In this study, we present a novel tracking system, in which the tracking accuracy can be considerably enhanced by state prediction. Accordingly, we present a new Q-learning-based reinforcement method, augmented by Wang–Landau sampling. In the proposed method, reinforcement learning is used to predict a target configuration for the subsequent frame, while Wang–Landau sampler balances the exploitation and exploration degrees of the prediction. Our method can adapt to control the randomness of policy, using statistics on the number of visits in a particular state. Thus, our method considerably enhances conventional Q-learning algorithm performance, which also enhances visual tracking performance. Numerical results demonstrate that our method substantially outperforms other state-of-the-art visual trackers and runs in realtime because our method contains no complicated deep neural network architectures.
Eli A. Meirom, Haggai Maron, Shie Mannor, Gal Chechik
Published: 11 October 2020
by ArXiv
Abstract:
We consider the problem of monitoring and controlling a partially-observed dynamic process that spreads over a graph. This problem naturally arises in contexts such as scheduling virus tests or quarantining individuals to curb a spreading epidemic; detecting fake news spreading on online networks by manually inspecting posted articles; and targeted marketing where the objective is to encourage the spread of a product. Curbing the spread and constraining the fraction of infected population becomes challenging when only a fraction of the population can be tested or quarantined. To address this challenge, we formulate this setup as a sequential decision problem over a graph. In face of an exponential state space, combinatorial action space and partial observability, we design RLGN, a novel tractable Reinforcement Learning (RL) scheme to prioritize which nodes should be tested, using Graph Neural Networks (GNNs) to rank the graph nodes. We evaluate this approach in three types of social-networks: community-structured, preferential attachment, and based on statistics from real cellular tracking. RLGN consistently outperforms all baselines in our experiments. It suggests that prioritizing tests using RL on temporal graphs can increase the number of healthy people by $25\%$ and contain the epidemic $30\%$ more often than supervised approaches and $2.5\times$ more often than non-learned baselines using the same resources.
Published: 1 May 2020
Pattern Recognition, Volume 101; doi:10.1016/j.patcog.2019.107188

The publisher has not yet granted permission to display this abstract.
Published: 18 April 2020
Sensors, Volume 20; doi:10.3390/s20082320

Abstract:
Counter-drone technology by using artificial intelligence (AI) is an emerging technology and it is rapidly developing. Considering the recent advances in AI, counter-drone systems with AI can be very accurate and efficient to fight against drones. The time required to engage with the target can be less than other methods based on human intervention, such as bringing down a malicious drone by a machine-gun. Also, AI can identify and classify the target with a high precision in order to prevent a false interdiction with the targeted object. We believe that counter-drone technology with AI will bring important advantages to the threats coming from some drones and will help the skies to become safer and more secure. In this study, a deep reinforcement learning (DRL) architecture is proposed to counter a drone with another drone, the learning drone, which will autonomously avoid all kind of obstacles inside a suburban neighborhood environment. The environment in a simulator that has stationary obstacles such as trees, cables, parked cars, and houses. In addition, another non-malicious third drone, acting as moving obstacle inside the environment was also included. In this way, the learning drone is trained to detect stationary and moving obstacles, and to counter and catch the target drone without crashing with any other obstacle inside the neighborhood. The learning drone has a front camera and it can capture continuously depth images. Every depth image is part of the state used in DRL architecture. There are also scalar state parameters such as velocities, distances to the target, distances to some defined geofences and track, and elevation angles. The state image and scalars are processed by a neural network that joints the two state parts into a unique flow. Moreover, transfer learning is tested by using the weights of the first full-trained model. With transfer learning, one of the best jump-starts achieved higher mean rewards (close to 35 more) at the beginning of training. Transfer learning also shows that the number of crashes during training can be reduced, with a total number of crashed episodes reduced by 65%, when all ground obstacles are included.
Junjie Wang, Xiaohong Su, Lingling Zhao, Jun Zhang
Frontiers in Bioengineering and Biotechnology, Volume 8; doi:10.3389/fbioe.2020.00298

Abstract:
Accurate target detection and association are vital for the development of reliable target tracking, especially for cell tracking based on microscopy images due to the similarity of cells. We propose a deep reinforcement learning method to associate the detected targets between frames. According to the dynamic model of each target, the cost matrix is produced by conjointly considering various features of targets and then used as the input of a neural network. The proposed neural network is trained using reinforcement learning to predict a distribution over the association solution. Furthermore, we design a residual convolutional neural network that results in more efficient learning. We validate our method on two applications: the multiple target tracking simulation and the ISBI cell tracking. The results demonstrate that our approach based on reinforcement learning techniques could effectively track targets following different motion patterns and show competitive results.
Dong-Yong Lee, Yong-Hun Cho, Dae-Hong Min, In-Kwon Lee
2020 IEEE Conference on Virtual Reality and 3D User Interfaces (VR) pp 155-163; doi:10.1109/vr46266.2020.1581309443724

The publisher has not yet granted permission to display this abstract.
Dong-Yong Lee, Yong-Hun Cho, Dae-Hong Min, In-Kwon Lee
2020 IEEE Conference on Virtual Reality and 3D User Interfaces (VR) pp 155-163; doi:10.1109/vr46266.2020.00034

The publisher has not yet granted permission to display this abstract.
Domain Adaptation for Visual Understanding; doi:10.1007/978-3-030-30671-7

The publisher has not yet granted permission to display this abstract.
Satpreet H. Singh, Floris Van Breugel, Rajesh P. N. Rao, Bingni W. Brunton
The 2020 Conference on Artificial Life; doi:10.1162/isal_a_00321

Abstract:
The ability to track odor plumes in dynamic environments is critical for flying insects following attractive odors to localize food or mates. This remarkable tracking behavior requires multimodal integration of odor, vision, and wind sensing, is robust to variations in plume statistics and wind speeds, and can often be performed over large distances. Therefore, it is challenging to study in confined experimental settings. Here we describe ongoing work to explore the space of policies effective to accomplish plume tracking, leveraging the reproducibility and interpretability of artificial agents trained in biologically motivated simulations. Specifically, we trained neural-network (NN) agents with deep reinforcement learning to locate the source of a patchy simulated plume, while varying their capacity to store past sensory stimuli. We analyzed the behavior of trained agents by inspecting successful trajectories. We then interrogated the input-output maps learned by the NNs, uncovering interpretable differences in control strategies introduced by varying sensory memory. We believe that our simulation-based approach can generate novel testable hypotheses to guide the development of targeted neuroethological experiments, as well as provide a pathway towards a mechanistic understanding of the key multimodal computations required for plume tracking. The ability to track odor plumes in dynamic environments is critical for flying insects following attractive odors to localize food or mates. This remarkable tracking behavior requires multimodal integration of odor, vision, and wind sensing, is robust to variations in plume statistics and wind speeds, and can often be performed over large distances. Therefore, it is challenging to study in confined experimental settings. Here we describe ongoing work to explore the space of policies effective to accomplish plume tracking, leveraging the reproducibility and interpretability of artificial agents trained in biologically motivated simulations. Specifically, we trained neural-network (NN) agents with deep reinforcement learning to locate the source of a patchy simulated plume, while varying their capacity to store past sensory stimuli. We analyzed the behavior of trained agents by inspecting successful trajectories. We then interrogated the input-output maps learned by the NNs, uncovering interpretable differences in control strategies introduced by varying sensory memory. We believe that our simulation-based approach can generate novel testable hypotheses to guide the development of targeted neuroethological experiments, as well as provide a pathway towards a mechanistic understanding of the key multimodal computations required for plume tracking. The ability to track odor plumes in dynamic environments is critical for flying insects following attractive odors to localize food or mates. This remarkable tracking behavior requires multimodal integration of odor, vision, and wind sensing, is robust to variations in plume statistics and wind speeds, and can often be performed over large distances. Therefore, it is challenging to study in confined experimental settings. Here we describe ongoing work to explore the space of policies effective to accomplish plume tracking, leveraging the reproducibility and interpretability of artificial agents trained in biologically motivated simulations. Specifically, we trained neural-network (NN) agents with deep reinforcement learning to locate the source of a patchy simulated plume, while varying their capacity to store past sensory stimuli. We analyzed the behavior of trained agents by inspecting successful trajectories. We then interrogated the input-output maps learned by the NNs, uncovering interpretable differences in control strategies introduced by varying sensory memory. We believe that our simulation-based approach can generate novel testable hypotheses to guide the development of targeted neuroethological experiments, as well as provide a pathway towards a mechanistic understanding of the key multimodal computations required for plume tracking.
Chen Chen, Hsieh-Yu Li, Audelia G. Dharmawan, Khairuldanial Ismail, Xiang Liu, U-Xuan Tan
2019 IEEE International Conference on Robotics and Biomimetics (ROBIO) pp 1121-1126; doi:10.1109/robio49542.2019.8961517

The publisher has not yet granted permission to display this abstract.
Xiao Long Wei, Xiang Lin Huang, Tao Lu, Ge Ge Song
2019 4th International Conference on Robotics and Automation Engineering (ICRAE) pp 130-134; doi:10.1109/icrae48301.2019.9043821

The publisher has not yet granted permission to display this abstract.
Stephen Grossberg
Frontiers in Computational Neuroscience, Volume 13; doi:10.3389/fncom.2019.00036

Abstract:
This article develops a model of how reactive and planned behaviors interact in real time. Controllers for both animals and animats need reactive mechanisms for exploration, and learned plans to efficiently reach goal objects once an environment becomes familiar. The SOVEREIGN model embodied these capabilities, and was tested in a 3D virtual reality environment. Neural models have characterized important adaptive and intelligent processes that were not included in SOVEREIGN. A major research program is summarized herein by which to consistently incorporate them into an enhanced model called SOVEREIGN2. Key new perceptual, cognitive, cognitive-emotional, and navigational processes require feedback networks which regulate resonant brain states that support conscious experiences of seeing, feeling, and knowing. Also included are computationally complementary processes of the mammalian neocortical What and Where processing streams, and homologous mechanisms for spatial navigation and arm movement control. These include: Unpredictably moving targets are tracked using coordinated smooth pursuit and saccadic movements. Estimates of target and present position are computed in the Where stream, and can activate approach movements. Motion cues can elicit orienting movements to bring new targets into view. Cumulative movement estimates are derived from visual and vestibular cues. Arbitrary navigational routes are incrementally learned as a labeled graph of angles turned and distances traveled between turns. Noisy and incomplete visual sensor data are transformed into representations of visual form and motion. Invariant recognition categories are learned in the What stream. Sequences of invariant object categories are stored in a cognitive working memory, whereas sequences of movement positions and directions are stored in a spatial working memory. Stored sequences trigger learning of cognitive and spatial/motor sequence categories or plans, also called list chunks, which control planned decisions and movements toward valued goal objects. Predictively successful list chunk combinations are selectively enhanced or suppressed via reinforcement learning and incentive motivational learning. Expected vs. unexpected event disconfirmations regulate these enhancement and suppressive processes. Adaptively timed learning enables attention and action to match task constraints. Social cognitive joint attention enables imitation learning of skills by learners who observe teachers from different spatial vantage points.
Dongdong Li, Yangliu Kuai, Gongjian Wen, Li Liu
2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) pp 592-600; doi:10.1109/cvprw.2019.00085

The publisher has not yet granted permission to display this abstract.
, Bing Bai, , , Yun Fu
Published: 5 December 2018
by IEEE
IEEE Transactions on Image Processing, Volume 28, pp 2331-2341; doi:10.1109/tip.2018.2885238

The publisher has not yet granted permission to display this abstract.
Seongheon Lee, Taemin Shim, Sungjoong Kim, Junwoo Park, Kyungwoo Hong, Hyochoong Bang
2018 International Conference on Unmanned Aircraft Systems (ICUAS) pp 108-114; doi:10.1109/icuas.2018.8453315

The publisher has not yet granted permission to display this abstract.
Sangdoo Yun, Jongwon Choi, Youngjoon Yoo, Kimin Yun,
IEEE Transactions on Neural Networks and Learning Systems, Volume 29, pp 2239-2252; doi:10.1109/tnnls.2018.2801826

The publisher has not yet granted permission to display this abstract.
Wonchul Kim, Taewan Kim, Jonggu Lee, H. Jin Kim
2017 11th Asian Control Conference (ASCC) pp 1046-1050; doi:10.1109/ascc.2017.8287315

The publisher has not yet granted permission to display this abstract.
Peng Cui, , Yintao Wang
Published: 12 September 2017
Journal of Advanced Transportation, Volume 2017, pp 1-11; doi:10.1155/2017/6716820

Abstract:
Autonomous robots need to be recharged and exchange information with the host through docking in the long-distance tasks. Therefore, feasible path is required in the docking process to guide the robot and adjust its pose. However, when there are unknown obstacles in the work area, it becomes difficult to determine the feasible path for docking. This paper presents a reactive path planning approach named Dubins-APF (DAPF) to solve the path planning problem for docking in unknown environment with obstacles. In this proposed approach the Dubins curves are combined with the designed obstacle avoidance potential field to plan the feasible path. Firstly, an initial path is planned and followed according to the configurations of the robot and the docking station. Then when the followed path is evaluated to be infeasible, the intermediate configuration is calculated as well as the replanned path based on the obstacle avoidance potential field. The robot will be navigated to the docking station with proper pose eventually via the DAPF approach. The proposed DAPF approach is efficient and does not require the prior knowledge about the environment. Simulation results are given to validate the effectiveness and feasibility of the proposed approach.1. IntroductionNowadays autonomous mobile robots such as autonomous drones, autonomous underwater vehicles, and automated vehicles are widely used in the complex environment and undertake the dangerous and heavy tasks [1–8]. However, their durations are constrained by their limited battery capacities and data storage spaces [9]. To solve this problem docking stations are designed and deployed in the work areas of the robots to maintain them in practical applications [10, 11]. To ensure the safety and the success in the docking process, a feasible path is required to be planned firstly [12]. This is because, for one thing, the area for docking may be unknown in advance or dynamically changing and there may be static and moving obstacles that threat the safety of the robots [12, 13]. For another, the final configuration (pose and velocity) of the robot should be adjusted properly to avoid the impact with the docking station. The kinematic characters of the robot should be considered as well for path tracking and energy saving [14–16].In recent years, a variety of approaches have been proposed to solve the path planning problem in unknown environment. Among them a biologically inspired neural network approach is presented in [17] to plan path for the autonomous underwater vehicle in unknown two-dimension (2D) environment, which is achieved via updating the environment maps according to Dempster-Shafer theory in steps. In [18] the online path planning problem with prescribed target in environment with unknown obstacles is considered and the neural networks trained by the reinforcement learning approach are adopted to solve this problem. In [19] the rapidly exploring random trees star () algorithm is employed to plan the path for autonomous underwater vehicles, where the mutual information between the scalar field model and observations is used to improve the path planning result. In [20] the path planning problem for household robot in unknown environment is considered and solved by the modified artificial potential field (APF) method based on the motion characteristics of household animals. In [21] the collision-free path planning for autonomous container truck is achieved via utilizing the improved ant colony optimization (ACO) algorithm. In this algorithm the local path is generated according to the selected local target which is determined and updated by the rolling window approach. In [22] a dynamic planning algorithm is presented to determine the collision-free path for the mobile terrestrial robot in unknown environment. In this algorithm the local objectives are determined by the genetic algorithm (GA) and the optimum routes are generated dynamically towards the global object. It can be concluded from the efforts above that the efficiency of the path planning approach is considered emphatically in the unknown environment. Meanwhile the reactive frameworks are adaptive in solving the path planning problem in the unknown environment in spite of the fact that different various local path planning approaches such as modified bioinspired method, RRT algorithm, and APF method are adopted. However, the smoothness of the planned path is rarely considered in these efforts as well as the pose and velocity of the robot, which is critical in the docking tasks [23].In this paper, a reactive path planning approach named Dubins-APF is proposed to solve the path planning problem for docking in unknown environment based on combination of the Dubins curves and the APF approach. Dubins curves have been proved to be the optimal paths with minimal turning radius that connect two points with prescribed poses in 2D space [24, 25]. However, if there are obstacles in the environment for docking, it is difficult to determine the feasible path based on Dubins curves [26, 27]. Since the APF approach is efficient in obstacle avoidance, in this paper it is combined with the Dubins curves to determine the feasible path for docking [28–31]. The proposed path planning approach works in a reactive mode which is described as follows. Once the planned path is infeasible, feasible intermediate configurations are determined based on the obstacle avoidance potential field according to the configurations of the detected obstacles. Then feasible Dubins curves are generated as the replanned path based on the intermediate configuration. Through implementing this path planning-replanning strategy continually, the DAPF algorithm will solve the path planning problem for docking.The main contributions of this paper are as follows:(1)The geometrical approach to determine the 3D Dubins curves is proposed in this paper. It can be utilized to generate the docking path and evaluate the feasibility of the planned path as well.(2)The conception of combining the advantages of the Dubins curves and the artificial potential field is proposed and implemented in this paper to improve the quality of the docking path while avoiding obstacles.The structure of this paper is presented as follows. The path planning problem for docking and the notion of the DAPF approach are introduced in Section 1. The problem statement and the path generation approach with 3D Dubins curves are described in Section 2. In Section 3 the Dubins-APF path planning approach is proposed and illustrated in detail. The simulation result and discussion about the DAPF approach are presented in Section 4. Some conclusions and future works are provided in Section 5.2. Path Planning with 3D Dubins Curves2.1. Problem StatementThe path planning problem for docking is to determine a collision-free path to connect the initial position of the autonomous robot and the docking station with prescribed poses under certain constraints of the robot [28]. In this paper the docking station with unidirectional entrance is considered and it is assumed to be static. The position of the docking station is written as and the direction of its entrance is expressed as . Meanwhile the velocity of the autonomous robot is assumed be constant and its position and velocity are written as and , where . The turning ability of the robot is assumed to be limited and the minimal turning radius is written as . Hence to avoid collision with the docking station, the final position and velocity of the robot should be close to and . Additionally the work area for docking is assumed to be unknown in advance, which means the autonomous robot can only acquire the environment information within its sensor range .2.2. 3D Dubins CurvesThe Dubins curve only consists of two kinds of segments which are the circle () segment and the straight-line () segment, where the radius of the segment is equal to the minimal turning radius of the robot and this curve is smooth at the intersections of the adjacent segments. In 2D environment the optimality of the two Dubins curves has been proved and the curve or the curve is reckoned as the shortest path [24]. However, the determination of the Dubins curve becomes complex in 3D environment due to the increase of the dimensionality, which means not all the segments of the Dubins curves are coplanar [32, 33]. Therefore considering the efficiency of its application in path planning, a 3D Dubins curve determination approach is presented in this paper based on the geometric characters of the typical curve.The typical curve is shown in Figure 1. The initial configuration of position and pose (green vector) and the final configuration (red vector) are presented as and and the feasible path that connects them is a 3D Dubins curve which consists of one segment and two segments. The segments (red arc) and (green arc) are two circular arcs with centers and and radii and , respectively. They are connected by the segment (blue straight line) with intersections and separately. To indicate the coplanarity of the Dubins curve, two auxiliary lines and are drawn as Figure 1 shows. has on it and parallels with and has on it and parallels with . They intersect with the elongation of the segment at and , respectively. Based on the spatial relations of , , and the segment, the segments of the curve can be divided into two planes determined by and the segment and and the segment separately. The intersection line of these two planes is the segment.Figure 1: The illustration of the typical 3D curve.Inspired by the geometric relations of these two planes, the determination process of the 3D Dubins curve is implemented as follows. The intersections of , , and the elongation of are presented aswhere and are nonzero constants. A special case of (1) should be noticed where either or is infinite. It means that the corresponding segment is semicircular arc and this case will be discussed later. Once and a
Sangdoo Yun, Jongwon Choi, Youngjoon Yoo, Kimin Yun, Jin Young Choi
2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) pp 1349-1358; doi:10.1109/cvpr.2017.148

The publisher has not yet granted permission to display this abstract.
Karan K. Budhraja, Tim Oates
2017 IEEE Congress on Evolutionary Computation (CEC) pp 67-76; doi:10.1109/cec.2017.7969297

The publisher has not yet granted permission to display this abstract.
, Jacob Leavitt, Dylan Stahl, Chris Landry, Philip Corlett
Published: 1 May 2017
Biological Psychiatry, Volume 81; doi:10.1016/j.biopsych.2017.02.606

The publisher has not yet granted permission to display this abstract.
Mathieu D'acremont,
Published: 4 February 2016
Cerebral Cortex, Volume 26, pp 1818-1830; doi:10.1093/cercor/bhw013

Abstract:
Large-scale human interaction through, for example, financial markets causes ceaseless random changes in outcome variability, producing frequent and salient outliers that render the outcome distribution more peaked than the Gaussian distribution, and with longer tails. Here, we study how humans cope with this evolutionary novel leptokurtic noise, focusing on the neurobiological mechanisms that allow the brain, 1) to recognize the outliers as noise and 2) to regulate the control necessary for adaptive response. We used functional magnetic resonance imaging, while participants tracked a target whose movements were affected by leptokurtic noise. After initial overreaction and insufficient subsequent correction, participants improved performance significantly. Yet, persistently long reaction times pointed to continued need for vigilance and control. We ran a contrasting treatment where outliers reflected permanent moves of the target, as in traditional mean-shift paradigms. Importantly, outliers were equally frequent and salient. There, control was superior and reaction time was faster. We present a novel reinforcement learning model that fits observed choices better than the Bayes-optimal model. Only anterior insula discriminated between the 2 types of outliers. In both treatments, outliers initially activated an extensive bottom-up attention and belief network, followed by sustained engagement of the fronto-parietal control network.
Cheng-Tao Chung, Cheng-Yu Tsai, Hsiang-Hung Lu, Chia-Hsiang Liu, Hung-Yi Lee, Lin-Shan Lee, Chung Cheng-Tao, Tsai Cheng-Yu, Lu Hsiang-Hung, Liu Chia-Hsiang, et al.
2015 IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU) pp 245-251; doi:10.1109/asru.2015.7404801

The publisher has not yet granted permission to display this abstract.
Jung-Il Park, Sung-Ho Suh, Umirov Ulugbek
Journal of the Korean Society for Precision Engineering, Volume 29, pp 79-86; doi:10.7736/kspe.2012.29.1.079

The publisher has not yet granted permission to display this abstract.
Shingo Nakamura, Shuji Hashimoto
The 2011 International Joint Conference on Neural Networks pp 2465-2470; doi:10.1109/ijcnn.2011.6033539

The publisher has not yet granted permission to display this abstract.
, Yu-Te Su, Shao-Wei Lai, Jhen-Jia Hu
IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), Volume 41, pp 736-748; doi:10.1109/tsmcb.2010.2089978

The publisher has not yet granted permission to display this abstract.
Feng Zhou, Deyun Zhou, Geng Yu
2008 Congress on Image and Signal Processing, Volume 4, pp 73-77; doi:10.1109/cisp.2008.236

The publisher has not yet granted permission to display this abstract.
M. Winter, G. Metta, G. Sandini
Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium, Volume 4, pp 539-542 vol.4; doi:10.1109/ijcnn.2000.860827

The publisher has not yet granted permission to display this abstract.
Page of 1
Articles per Page
by
Show export options
  Select all
Back to Top Top