The null space of the Jacobian for the redundant manipulators can be utilized to reach dual control goals and optimal effects. For the flexible base redundant manipulator system, the reaction force from the motion of ...
The null space of the Jacobian for the redundant manipulators can be utilized to reach dual control goals and optimal effects. For the flexible base redundant manipulator system, the reaction force from the motion of the manipulator should be carefully handled while the end effector is tracking a desired trajectory. Different optimal goals are discussed, including the minimization of base deflection force/torque and base energy dissipation path. Considering the base reaction dynamic, the reaction force/torque is treated as the input of the base and the deflection energy is treated as the output, an integrated optimal goal is built and solved based on the LQR control to get an optimal feedback for the closed-loop dynamics. We discuss the different optimal goals and implement trough the redefined acceleration. We test the different control laws on the three- link manipulators which mounted on a three-DoF flexible base. Simulation results demonstrate that optimal goal in LQR form can be advantageous in terms of flexible base manipulator.
Conflict Based Search(CBS) is used for multi-agent Pathfinding (MAPF) to enable each Agent to reach the target node. The CBS algorithm uses the heuristic algorithm A* search to calculate the MAPF solution, and the pat...
Conflict Based Search(CBS) is used for multi-agent Pathfinding (MAPF) to enable each Agent to reach the target node. The CBS algorithm uses the heuristic algorithm A* search to calculate the MAPF solution, and the path planning uses forward search, which cannot explore the path of the unknown region. On this basis, this paper proposes that, in the case of unknown map and changing environment at any time, when encountering new obstacles, the information obtained from previous search should be used without completely replanning the path. Because of the idea of incremental programming, the number of reprogramming times and the number of affected nodes can be reduced. The optimized algorithm consumes less time and memory, and improves the efficiency of path planning.
This paper presents a real-time, pixelwise method to generate grasp synthesis based on fully convolutional netural networks (FCN). Our proposed Attention Grasping Network (AGN) applies a novel attention mechanism to r...
This paper presents a real-time, pixelwise method to generate grasp synthesis based on fully convolutional netural networks (FCN). Our proposed Attention Grasping Network (AGN) applies a novel attention mechanism to robotic grasp detection, which automatically learns to focus on salient features of the input image. The model with attention mechnisms can compensate for the loss of detail information in standard FCN, which increases the sensitivity of the model and accuracy of prediction. In addition,in order to ensure a real-time grasp and save computing resources, the light-weight AGN model predicts the position and angle of grasping point. Our method only takes 22ms to execute the grasp detection pipeline on a GPU-equipped computer, and achieves 97.8% accuracy on Cornell Grasping Dataset.
In the entire transportation system, Railway transportation plays a very important role. However, the current test method is mainly manual inspection, labor intensity, high risk. In recent years, with the rapid develo...
详细信息
ISBN:
(数字)9781728107707
ISBN:
(纸本)9781728107714
In the entire transportation system, Railway transportation plays a very important role. However, the current test method is mainly manual inspection, labor intensity, high risk. In recent years, with the rapid development of MEMS technology, wireless network sensors have been widely used. Sensors play an important role in fusing with big data. Sensors are becoming smaller and smarter. The application of wireless network sensor provides a new idea for the safety monitoring of railway transportation. In view of the current backwardness of railway transportation detection methods and high cost of manual maintenance, this study builds a multi-mode wireless sensor platform for railway monitoring based on MEMS sensors. The platform is mainly composed of a power supply system, a monitoring system and wireless sensor network. As the core of the system energy, the power supply system makes full use of solar energy resources, reduces system cost. And the platform achieves the goal of long-term railway monitoring without external DC power supply. The railway generates acceleration information when a train passes by. Wireless network sensors monitor the status of the railway by collecting acceleration information. The monitoring platform realizes real-time transmission of monitoring information without external power supply. And The intelligent monitoring and maintenance of railway transportation system are realized. The experimental results show that the platform can reduce the cost of manual monitoring and maintenance, and has important practical significance for railway maintenance.
In this paper, we propose a novel pose optimizer which can be inserted into either supervised or unsupervised end-to-end visual odometry for the purpose of local pose optimization. The pose optimizer is an analogue of...
In this paper, we propose a novel pose optimizer which can be inserted into either supervised or unsupervised end-to-end visual odometry for the purpose of local pose optimization. The pose optimizer is an analogue of the pose graph optimization used in traditional VSLAM algorithms. Local pose optimization is performed by an attention-based neural network which iteratively refines the predicted pose estimates of an image snippet. Instead of complicated graph convolutional network, the attention mechanism based on geometric consistency of trajectory constraint is utilized because pose features whose spatial distribution is not important can be flattened to vectors and then processed. The pose optimizer is aimed at improving pose estimation accuracy by redistributing errors of pose estimates. Quantitative and qualitative evaluation of the proposed approach on the KITTI Odometry dataset [1] is presented to demonstrate its effectiveness in improving pose estimation accuracy and minimizing pose drift.
Hand gesture recognition has become the focus of researchers lately because of its manifold applications in various fields. Leap Motion (LM) is a device to obtain useful and accurate information of the hand action, wh...
详细信息
ISBN:
(数字)9781728107707
ISBN:
(纸本)9781728107714
Hand gesture recognition has become the focus of researchers lately because of its manifold applications in various fields. Leap Motion (LM) is a device to obtain useful and accurate information of the hand action, which is suitable for collecting the three-dimensional (3D) human hand gesture. In this paper, a novel framework which consists of an incremental learning (IL) algorithm without deep structure is proposed and applied to hand gestures classification that explicitly aimed to the LM data. The same datasets are used to train the proposed framework and the conventional Long Short Term Memory Recurrent Neural Network (LSTM-RNN). Due to the structural advantage of the proposed model, the recognition performance is improved distinctly in robustness and training time than the LSTM network. Moreover, convincing experiment results are given to illustrate that the solution is more efficient in static gesture classification.
Compared with traditional 2D image processing, 3D point cloud processing has become a hot technology in the industry, but there are few researches applied to nondestructive testing of curved workpieces. This article a...
Compared with traditional 2D image processing, 3D point cloud processing has become a hot technology in the industry, but there are few researches applied to nondestructive testing of curved workpieces. This article aims at the non-destructive testing of curved workpieces, using 3D point cloud and robotic arm for water immersion ultrasonic non-destructive testing. Aiming at the 3D point cloud with many outliers, using the commonly used filter in two-dimensional images-the guide filter, the algorithm is improved and used for the filtering and downsampling of the 3D point cloud, and the adjustment of the curved surface workpiece and the robot arm's pose is introduced. Finally, the workpiece was scanned by a robotic arm with a water immersion ultrasonic nondestructive testing system to prove the feasibility of the experiment and improve the detection efficiency of traditional nondestructive testing technology.
Model-Free Adaptive control (MFAC) is a new data-driven control method, which depends only on the input/output (I/O) measurement data rather than the mathematical model information of the actual controlled system. The...
详细信息
ISBN:
(纸本)9781728102634
Model-Free Adaptive control (MFAC) is a new data-driven control method, which depends only on the input/output (I/O) measurement data rather than the mathematical model information of the actual controlled system. The SISO MFAC based on the compact-form dynamic linearization (SISO-CFMFAC) is a promising approach to control the SISO nonlinear systems. However, the parameters in the SISO-CFMFAC should be tuned carefully before being put into use. Unfortunately, so far the parameter tuning of SISO-CFMFAC is still a laborious, time-consuming and cost-consuming work. In this paper, a novel parameter self-tuning approach of SISO-CFMFAC based on back propagation Neural Network with System Error set as input (SISO-CFMFAC-NNSE) is proposed, and then verified by using a typical time-varying nonlinear SISO system. Results show that the proposed controller named SISO-CFMFAC-NNSE can achieve better control stability and accuracy than the existing controller of SISO-CFMFAC.
Recent years have witnessed rapid progress in target tracking. To track a moving target for mobile robots, however, both performance and speed of the algorithm are indispensable. This paper proposes a dual model fusio...
Recent years have witnessed rapid progress in target tracking. To track a moving target for mobile robots, however, both performance and speed of the algorithm are indispensable. This paper proposes a dual model fusion strategy to improve target tracking drift. Among them, the spatiotemporal context model in middle-level feature space (MFSTC) is utilized to ameliorate target tracking effect when illumination or appearance changes, the mean shift based on 3D back projection (MS3D) is fused to allow the algorithm to tackle occlusion and deformation. Tracking controller based on visual servo is also designed for mobile robots. We validate the efficiency of the proposed fusion model on the university of Birmingham RGB-D Tracking Benchmark (BTB) and show that our approach compares favorably with the state-of-the-arts, mobile robots using our approach can track targets robustly under various challenging scenes.
暂无评论