New satellite-based 6G for the Internet of Things (IoT) is expected to provide complete global coverage and support fully transparent services. In addition, a large number of low Earth orbit (LEO) satellites are to be...
详细信息
New satellite-based 6G for the Internet of Things (IoT) is expected to provide complete global coverage and support fully transparent services. In addition, a large number of low Earth orbit (LEO) satellites are to be deployed to connect IoT devices (actuators and sensors) that are beyond terrestrial network coverage. However, there are the two major LEO satellites' hindering issues: 1) spectrum inefficiency leading to high cost and 2) continuous motions of satellites that limit the contact time to approximately 10 min which results in frequent handovers, link budget limitations, and high Doppler effects. This article discusses design approaches and principles that allow us to develop a cost-effective intelligent data-aided satellite communication and control framework for LEO networks by employing key features of 6G multiconnectivity, distributed sensing, and machine learning algorithms. Our ideas are evaluated with preliminary analytic modeling and simulation results.
Accurate modeling of system dynamics is crucial for achieving high-performance planning and control of robotic systems. Although existing data-driven approaches represent a promising approach for modeling dynamics, th...
详细信息
ISBN:
(纸本)9798350377712;9798350377705
Accurate modeling of system dynamics is crucial for achieving high-performance planning and control of robotic systems. Although existing data-driven approaches represent a promising approach for modeling dynamics, their accuracy is limited to a short prediction horizon, overlooking the impact of compounding prediction errors over longer prediction horizons. Strategies to mitigate these cumulative errors remain underexplored. To bridge this gap, in this paper, we study the key design choices for efficiently learning long-horizon prediction dynamics for quadrotors. Specifically, we analyze the impact of multiple architectures, historical data, and multi-step loss formulation. We show that sequential modeling techniques showcase their advantage in minimizing compounding errors compared to other types of solutions. Furthermore, we propose a novel decoupled dynamics learning approach, which further simplifies the learning process while also enhancing the approach modularity. Extensive experiments and ablation studies on real-world quadrotor data demonstrate the versatility and precision of the proposed approach. Our outcomes offer several insights and methodologies for enhancing long-term predictive accuracy of learned quadrotor dynamics for planning and control.
Due to the scarcity of fault samples and the weakness of processing higher-order interactive information, the most existing intelligence methods fail to achieve the optimal effect in fault diagnosis. To address these ...
详细信息
ISBN:
(纸本)9798350321050
Due to the scarcity of fault samples and the weakness of processing higher-order interactive information, the most existing intelligence methods fail to achieve the optimal effect in fault diagnosis. To address these problems, a time-frequency hypergraph neural network-based fault diagnosis method is proposed. In the proposed network, the limited data is initially segmented using the sliding window mechanism to obtain a set of time-domain signal instances. Additionally, the Fast Fourier Transform (FFT) is applied to each signal instance to extract corresponding frequency-domain signals, so as to capture more fault-sensitive features. Subsequently, a two-layer convolutional neural network is used to extract fault-attention features from both the time and frequency domain signals. Also, in order to reduce computational complexity, the time-frequency domain features are adaptively stacked based on a self-attention mechanism. Furthermore, a feature similarity graph is constructed for the time-frequency domain features using a k-nearest neighbor algorithm. This graph is then input into the hypergraph neural network (HGNN) to obtain the final diagnosis results. One comparative experiment shows that the proposed method not only mitigates the performance degradation caused by limited samples and noisy environments, but also effectively leverages the higher-order interaction information among nodes in the hypergraph.
This article presents an off-policy tracking control scheme for the continuous-time nonaffine yaw channel of uncrewed aerial vehicle helicopter. First, the article constructs an affine augmented system (AAS) within a ...
详细信息
Industrial robots perform tasks through tools installed on the end flange. The position and orientation of the tools are essential factors that affect the motion accuracy of industrial robots. However, existing calibr...
详细信息
ISBN:
(纸本)9798350321050
Industrial robots perform tasks through tools installed on the end flange. The position and orientation of the tools are essential factors that affect the motion accuracy of industrial robots. However, existing calibration methods for the tool frame mainly depend on manual observation. To solve this problem, this paper proposes an automatic calibration method of the tool frame based on the fact that the accurate position and orientation of the tools relative to the flange can be obtained through the calibration of the tool frame. First, the tool carried by the robot moves in a uniform circle at different heights. The origin and orientation calibration models of the tool frame are established respectively based on the similarity of the motion track of each point on a rigid body. Through two pairs of vertically mounted laser beam sensors, the time when the tool passes through the laser beam and the position of the corresponding robot flange are obtained. Second, the simulation platform with the robot and sensors is built in a 3-dimensional software to simulate the motion and measurement of the tool. The data required for calibration are acquired, by which the parameters of the origin and orientation of the tool frame are identified and compensated in the motion controller of the robot. Finally, the accuracy of the tool frame before and after calibration is tested in the simulation platform, and the simulation results verify the effectiveness of the proposed model and method.
In this paper, we infer remote measurements such as remote voltages and currents online with change in AC power flows using spiking neural network (SNN) as grid-edge technology for efficient coordination of power elec...
详细信息
ISBN:
(纸本)9798350318562;9798350318555
In this paper, we infer remote measurements such as remote voltages and currents online with change in AC power flows using spiking neural network (SNN) as grid-edge technology for efficient coordination of power electronic converters. This work unifies power and information as a means of data normalization using a multi-modal regime in the form of spikes using energy-efficient neuromorphic learning and event-driven asynchronous data collection. Firstly, we organize the synchronous real-valued measurements at each edge and translate them into asynchronous spike-based events to collect sparse data for training of SNN at each edge. Instead of relying on error-dependent supervised data-drivenlearning theory, we exploit the latency-driven unsupervised Hebbian learning rule to obtain modulation pulses for switching of power electronic converters that can now comprehend grid disturbances locally and adapt their operation without requiring explicit infrastructure for global coordination. Not only does this philosophy block exogenous path arrival for cyber attackers by dismissing the cyber layer, it also entails converter adaptation to system reconfiguration and parameter mismatch issues. We conclude this work by validating its energy-efficient and effective online learning performance under various scenarios in different system sizes, including modified ieee 14-bus system and under experimental conditions.
A hybrid (i.e., physics-guided data-driven) feedforward tracking controller is proposed for systems with unmodeled linear or nonlinear dynamics. The proposed controller is based on the filtered basis functions (FBF) a...
详细信息
A hybrid (i.e., physics-guided data-driven) feedforward tracking controller is proposed for systems with unmodeled linear or nonlinear dynamics. The proposed controller is based on the filtered basis functions (FBF) approach, and hence called a hybrid FBF controller. It formulates the feedforward control input to a system as a linear combination of a set of basis functions whose coefficients are selected to minimize tracking errors. To predict the system response and thereby the tracking errors, the basis functions are filtered using a combination of two linear models. The first model is physics-based and remains unaltered during the execution of the controller, while the second is data-driven and is continuously updated during the execution of the controller. To ensure its practicality and safe learning, the proposed hybrid FBF controller is equipped with the abilities to handle delays in data acquisition and to detect impending instability due to its inherent data-driven feedback loop. The effectiveness of the hybrid FBF controller is demonstrated via application to vibration compensation of a 3D printer with unmodeled linear and nonlinear dynamics. Thanks to the proposed hybrid FBF controller, the tracking accuracy of the 3D printer and the print quality are both significantly improved in experiments involving high-speed printing, compared to standard FBF controller that does not incorporate a data-driven model. Furthermore, the ability of the hybrid FBF controller to detect, and hence to potentially avoid, impending instability is demonstrated offline using data collected online from experiments.
Since the late 20th century, the demand for petrochemical materials and products in industry and daily life has increased, leading to the widespread use of pipelines for transporting these products. Recognizing the im...
详细信息
Rolling bearings, as a rotating component, are of great importance to ensure the normal operation and smooth running of important equipment. Remaining useful life (RUL) prediction is a hot research topic in the engine...
详细信息
ISBN:
(纸本)9798350321050
Rolling bearings, as a rotating component, are of great importance to ensure the normal operation and smooth running of important equipment. Remaining useful life (RUL) prediction is a hot research topic in the engineering field, which is helpful to ensure the operational safety of system equipment and reduce maintenance cost. The topic of how to utilize the important feature information in the time-series data and the reasonable use of attention mechanism are addressed in this study with a CBAM-CNN-BiLSTM-based technique for estimating the remaining service life of rolling bearings. Firstly, multi-domain features of vibration signals are extracted from time domain, frequency domain and time-frequency domain, and the features are normalized to the maximum-minimum value. Then, a convolutional neural network incorporating a hybrid convolutional attention module is used to extract the important features;a bidirectional long- and short-term memory network is employed to obtain the before-and-after dependencies in the features. Next, the self-attention mechanism is introduced into the bidirectional long and short-term network to focus on more important deep features. Finally, the effectiveness of the method is verified by the XJTU-SY dataset. The comparative study shows that the proposed CBAM-CNN-BiLSTM model outperforms other state-of-the-art methods in RUL prediction and system prediction, with higher prediction accuracy and generalization performance.
Model Predictive Path Integral (MPPI) is a recognized sampling-based approach for finite horizon optimal control problems. However, the efficacy and computational efficiency of prevailing MPPI methods are heavily reli...
详细信息
Model Predictive Path Integral (MPPI) is a recognized sampling-based approach for finite horizon optimal control problems. However, the efficacy and computational efficiency of prevailing MPPI methods are heavily reliant on the quality of rollouts. This is problematic because it is hard to sample a low-cost trajectory using random control sequences, thereby leading to inferior performance and computational efficiency, especially under constrained resources. To address this issue, we propose a data-efficient MPPI method called reinforcement learning-driven MPPI (RL-driven MPPI), which significantly reduces the dependency on the quantity and quality of samples. RL-driven MPPI employs an offline-online policy learning scheme, where the offline policy learned by RL serves as the initial solution and the initial rollout generator of MPPI, effectively combining the strengths of both RL and MPPI. The rollouts generated by RL typically correspond to a lower cost-to-go compared to random sampling, which significantly boosts the sample efficiency and convergence speed of MPPI. Moreover, the value function learned by RL offers an accurate estimation for infinite-horizon cost-to-go, enabling it to serve as a terminal term for the cost criteria of MPPI. This approach empowers MPPI to approximate an infinite-horizon cost with a shorter prediction horizon, thus enhancing real-time performance at each time step. An unmanned aerial vehicle control task is conducted to evaluate the proposed method. Results indicate that the proposed RL-driven MPPI method exhibits superior control performance and sample efficiency.
暂无评论