Hyper Spectral Image Super-Resolution(HSI-SR) has gained significant attention in recent years due to its potential applications.However,the challenge of obtaining high-resolution hyperspectral images is compounded by...Hyper Spectral Image Super-Resolution(HSI-SR) has gained significant attention in recent years due to its potential applications.However,the challenge of obtaining high-resolution hyperspectral images is compounded by limitations in sensor resolution and the high dimensionality of spectral data.Traditional approaches,including interpolation-based methods and sparse representation techniques,often struggle to capture the intricate spectral-spatial dependencies in hyperspectral images.To address these limitations,this study proposes a Hadamard Self-Attention Network(HSAN) for fusing a High-resolution Multispectral Image(Hr-MSI) and a Low-resolution Hyper Spectral Image(Lr-HSI),achieving HSI-SR for obtaining a High-resolution Hyper Spectral Image(Hr-HSI).The core of HSAN is a new Hadamard self-attention mechanism that can be more efficient than traditional dot-product attention because it avoids matrix multiplications and softmax operations.Considering that deep learning-based data fusion typically entails a significant computational and storage burden,this new approach can be integrated with convolutional layers to form an unsupervised lightweight network,which significantly reduces dependence on computational resources.Experimental results across four datasets validate the effectiveness and advantages of HSAN,compared with state-of-the-art approaches.The source code will be available at http://gffzz188fe103f8f1460asqpckb6kwcvww6fqq.ffgz.tsg.suse.edu.cn/zxnhkdm/HSAN.展开更多
Adaptive gaits for legged robots often requires force sensors installed on foot-tips,however impact,temperature or humidity can affect or even damage those sensors.Efforts have been made to realize indirect force esti...Adaptive gaits for legged robots often requires force sensors installed on foot-tips,however impact,temperature or humidity can affect or even damage those sensors.Efforts have been made to realize indirect force estimation on the legged robots using leg structures based on planar mechanisms.Robot Octopus III is a six-legged robot using spatial parallel mechanism(UP-2UPS) legs.This paper proposed a novel method to realize indirect force estimation on walking robot based on a spatial parallel mechanism.The direct kinematics model and the inverse kinematics model are established.The force Jacobian matrix is derived based on the kinematics model.Thus,the indirect force estimation model is established.Then,the relation between the output torques of the three motors installed on one leg to the external force exerted on the foot tip is described.Furthermore,an adaptive tripod static gait is designed.The robot alters its leg trajectory to step on obstacles by using the proposed adaptive gait.Both the indirect force estimation model and the adaptive gait are implemented and optimized in a real time control system.An experiment is carried out to validate the indirect force estimation model.The adaptive gait is tested in another experiment.Experiment results show that the robot can successfully step on a 0.2 m-high obstacle.This paper proposes a novel method to overcome obstacles for the six-legged robot using spatial parallel mechanism legs and to avoid installing the electric force sensors in harsh environment of the robot's foot tips.展开更多
基金National Natural Science Foundation of China(No. 42571463)Macao Young Scholars Program(No. AM2023033)+1 种基金Shaanxi Province Youth Science and Technology Star Program(No. 2024ZCKJXX-115)Natural Science Foundation of Shaanxi Province(No.2025JC-YBMS-257)。
摘要Hyper Spectral Image Super-Resolution(HSI-SR) has gained significant attention in recent years due to its potential applications.However,the challenge of obtaining high-resolution hyperspectral images is compounded by limitations in sensor resolution and the high dimensionality of spectral data.Traditional approaches,including interpolation-based methods and sparse representation techniques,often struggle to capture the intricate spectral-spatial dependencies in hyperspectral images.To address these limitations,this study proposes a Hadamard Self-Attention Network(HSAN) for fusing a High-resolution Multispectral Image(Hr-MSI) and a Low-resolution Hyper Spectral Image(Lr-HSI),achieving HSI-SR for obtaining a High-resolution Hyper Spectral Image(Hr-HSI).The core of HSAN is a new Hadamard self-attention mechanism that can be more efficient than traditional dot-product attention because it avoids matrix multiplications and softmax operations.Considering that deep learning-based data fusion typically entails a significant computational and storage burden,this new approach can be integrated with convolutional layers to form an unsupervised lightweight network,which significantly reduces dependence on computational resources.Experimental results across four datasets validate the effectiveness and advantages of HSAN,compared with state-of-the-art approaches.The source code will be available at http://gffzz188fe103f8f1460asqpckb6kwcvww6fqq.ffgz.tsg.suse.edu.cn/zxnhkdm/HSAN.
基金Supported by National Basic Research Program of China(973 Program,Grant No.2013CB035501)Research Fund of the State Key Lab of MSV of China(Grant No.MSV201208)
摘要Adaptive gaits for legged robots often requires force sensors installed on foot-tips,however impact,temperature or humidity can affect or even damage those sensors.Efforts have been made to realize indirect force estimation on the legged robots using leg structures based on planar mechanisms.Robot Octopus III is a six-legged robot using spatial parallel mechanism(UP-2UPS) legs.This paper proposed a novel method to realize indirect force estimation on walking robot based on a spatial parallel mechanism.The direct kinematics model and the inverse kinematics model are established.The force Jacobian matrix is derived based on the kinematics model.Thus,the indirect force estimation model is established.Then,the relation between the output torques of the three motors installed on one leg to the external force exerted on the foot tip is described.Furthermore,an adaptive tripod static gait is designed.The robot alters its leg trajectory to step on obstacles by using the proposed adaptive gait.Both the indirect force estimation model and the adaptive gait are implemented and optimized in a real time control system.An experiment is carried out to validate the indirect force estimation model.The adaptive gait is tested in another experiment.Experiment results show that the robot can successfully step on a 0.2 m-high obstacle.This paper proposes a novel method to overcome obstacles for the six-legged robot using spatial parallel mechanism legs and to avoid installing the electric force sensors in harsh environment of the robot's foot tips.