In this paper a high-speed vision system for table tennis robot is designed. The system architecture is designed firstly. Then a ball detection algorithm is proposed to improve the image processing speed up to 200 fra...
详细信息
In this paper a high-speed vision system for table tennis robot is designed. The system architecture is designed firstly. Then a ball detection algorithm is proposed to improve the image processing speed up to 200 frames per second. Besides, a novel ball trajectory reconstruction algorithm is established, in which the cameras in the stereo vision system needn't be synchronized. Experimental results verify the effectiveness and accuracy of the proposed method.
This paper describes a recognition method for the initial and end points of lap joints based on the image processing techniques. The edge detection technique is used in order to find the edges of the image. And the im...
详细信息
This paper describes a recognition method for the initial and end points of lap joints based on the image processing techniques. The edge detection technique is used in order to find the edges of the image. And the image processing techniques are employed in order to link the broken edges and find T-junctions of the image which are determined as candidates for the initial and end points of the weld seam. The T-junctions that match the conditions of angles checking and coordinate comparing will be detected as initial and end points of weld seam. The proposed method mainly insists of the 4 steps namely 1) smoothing the image and extracting edges of the image by using Canny operator 2)linking broken edges 3)detecting T-junctions and 4)analysing the initial and end points of weld seams. The experimental results showed that the initial and end points of the lap joints can be precisely recognized using the methods proposed in this paper.
In this paper, we develop a data-driven iterative adaptive dynamic programming algorithm to learn offline the approximate optimal control of unknown discrete-time nonlinear systems. We do not use a model network to id...
ISBN:
(纸本)9781479914821
In this paper, we develop a data-driven iterative adaptive dynamic programming algorithm to learn offline the approximate optimal control of unknown discrete-time nonlinear systems. We do not use a model network to identify the unknown system, but utilize the available offline data to learn the approximate optimal control directly. First, the data-driven iterative adaptive dynamic programming algorithm is presented with a convergence analysis. Then, the error bounds for this algorithm are provided considering the approximation errors of function approximation structures. To implement the developed algorithm, two neural networks are used to approximate the state-action value function and the control policy. Finally, two simulation examples are given to demonstrate the effectiveness of the developed algorithm.
In this paper, we introduce a metric learning approach for the classification process in the recognition procedure for P300 waves in electroencephalographic (EEG) signals. We show that the accuracy of support machine ...
详细信息
ISBN:
(纸本)9781479946037
In this paper, we introduce a metric learning approach for the classification process in the recognition procedure for P300 waves in electroencephalographic (EEG) signals. We show that the accuracy of support machine vector (SVM) classification is significantly improved by learning a similarity metric from the training data instead of using the default Euclidean metric. The effectiveness of the algorithm is validated through experiments on the dataset II of the brain-computer interface (BCI) Competition III (P300 speller).
In Still-to-Video (S2V)face recognition, only a few high resolution images are enrolled for each subject, while the probe is videos of complex variations. As faces present distinct characteristics under different scen...
详细信息
In Still-to-Video (S2V)face recognition, only a few high resolution images are enrolled for each subject, while the probe is videos of complex variations. As faces present distinct characteristics under different scenarios, recognition in the original space is obviously inefficient. In this paper, we propose a novel discriminant analysis method to learn separate mappings for different scenarios (still, video), and further pursue a common discriminant space based on these mappings. Concretely, by modeling each video as a set of local models, we form the scenario-oriented mapping learning as an Image-Model discriminant analysis framework. The learning objective is formulated by incorporating the intra-class compactness and inter-class separability for good discrimination. Moreover, a weighted learning scheme is introduced to concentrate on the discriminating information of the most confusing samples and then further enhance the performance. Experiments on the COX-S2V dataset demonstrate the effectiveness of the proposed method.
The paper investigates a path planning and following algorithm of observing targets for an underwater vehicle-manipulator system(UVMS). The algorithm including the operational task assignment and path following for th...
详细信息
The paper investigates a path planning and following algorithm of observing targets for an underwater vehicle-manipulator system(UVMS). The algorithm including the operational task assignment and path following for the UVMS is proposed. The operational task assignment consists of path planning and task assignment. Path planning provides planned trajectory distance between any two targets for task assignment, and task assignment produces a suitable task order for the minimum of the UVMS sailing distance. In path planning, the shortest path considering UVMS nonholonomic constraints-Dubins curve, is employed. It is implemented by matrix transformation. Then a digraph is generated by marking the shortest distance between any two targets. Hence the task assignment problem is considered as a traveling salesman problem, which can be achieved by the genetic algorithm. Finally a path-following guidance method is employed. Simulation results show the effectiveness of the proposed method.
An online reinforcement learning algorithm is proposed in this paper to directly utilizes online data efficiently for continuous deterministic systems without system parameters. The dependence on some specific approxi...
详细信息
ISBN:
(纸本)9781479945511
An online reinforcement learning algorithm is proposed in this paper to directly utilizes online data efficiently for continuous deterministic systems without system parameters. The dependence on some specific approximation structures is crucial to limit the wide application of online reinforcement learning algorithms. We utilize the online data directly with the kd-tree technique to remove this limitation. Moreover, we design the algorithm in the Probably Approximately Correct principle. Two examples are simulated to verify its good performance.
Percutaneous coronary intervention is the gold standard to coronary diseases in the past decades due to much less trauma and quick recovery. However, due to the traits of minimal invasiveness, clinicians have to defea...
详细信息
ISBN:
(纸本)9781424479276
Percutaneous coronary intervention is the gold standard to coronary diseases in the past decades due to much less trauma and quick recovery. However, due to the traits of minimal invasiveness, clinicians have to defeat the difficulties in eye-hand coordination during the procedure, which also makes it a non-trivial task in the catheterization lab. The computer-aided surgical simulation is designed to provide a reliable tool for the early stage of the training of the procedure. In this simulation system, the surface model of the vessels contribute the major part in the virtual anatomic environment. On the other hand, heavy interactions between the virtual surgical tools and the model surface occur during the training. In order to achieve acceptable performances, the patient-specific vessel surface model needs further process to adapt to this situation. We proposed in this paper an approach to optimize the meshes that consist the surface model with its application in consideration. The connectivity of the surface model is firstly checked. Next a smooth processing is applied without modifying the geometry of the largest-connected surface. Then the quantities of the polygons consisting the model surface are eliminated both dramatically and appropriately. The resultant surface model is applied in the validation test interacting with the virtual guidewire.
Recently, the development of minimally invasive surgery has led to the emergence of the new research area: minimally invasive vascular surgery simulation. The goal of the surgery simulation is to provide a new way to ...
详细信息
Recently, the development of minimally invasive surgery has led to the emergence of the new research area: minimally invasive vascular surgery simulation. The goal of the surgery simulation is to provide a new way to enable the trainees to obtain the core skills of the techniques. One of the key component of the simulation for core skills training in minimally invasive vascular surgery is how to develop 3D real-time computer-based virtual environments. In this paper, we present the virtual environments targeting to simulate the motion of a catheter/guide wire with complex 3D vascular models. Experimental results show that the 3D virtual environments are effective and promising.
In this paper, we propose a novel pan-tilt control approach for visual tracking. The approach uses the target image feature to control the pan-tilt device following the target with random movement. In order to improve...
详细信息
In this paper, we propose a novel pan-tilt control approach for visual tracking. The approach uses the target image feature to control the pan-tilt device following the target with random movement. In order to improve the performance of pan-tilt camera platform, active disturbance rejection control (ADRC) technique is applied to the tracking control of pantilt camera. The approach uses two ADRC controllers working in parallel for a pan-tilt camera platform which can be fixed on an autonomous mobile robot. This kind of control scheme does not need the precise mathematical model of pan-tilt camera tracking system and also not need the movement state of the target. In ADRC technique, many uncertain factors, estimated and compensated by extended state observer (ESO), are treated as system disturbance in real time. Then, the original nonlinear control system is reduced to a cascade integral form, which is easy to design 1-order ADRC controllers for the system. The tracking system uses a basic mean shift tracker, which calculates the error between the target image position and the center of the image in each frame for the controllers. Experimental results on a mobile robot show that the pan-tilt camera can track the moving target efficiently and stable.
暂无评论