It is a widely accepted view that considering the memory effects of historical information(driving operations) is beneficial for vehicle trajectory prediction models to improve prediction accuracy. However,many common...
详细信息
It is a widely accepted view that considering the memory effects of historical information(driving operations) is beneficial for vehicle trajectory prediction models to improve prediction accuracy. However,many commonly used models(e.g., long short-term memory, LSTM) can only implicitly simulate memory effects, but lack effective mechanisms to capture memory effects from sequence data and estimate their effective time range(ETR). This shortage makes it hard to dynamically configure the most suitable length of used historical information according to the current driving behavior, which harms the good understanding of vehicle motion. To address this problem, we propose a modified trajectory prediction model based on ordered neuron LSTM(ON-LSTM). We demonstrate the feasibility of ETR estimation based on ON-LSTM and propose an ETR estimation method. We estimate the ETR of driving fluctuations and lane change operations on the NGSIM I-80 dataset. The experiment results prove that the proposed method can well capture the memory effects during trajectory prediction. Moreover, the estimated ETR values are in agreement with our intuitions.
This paper tackles the optimal tracking control problem for reconfigurable manipulators based on critic-only policy iteration(Co PI) algorithm. By system transformation, the optimal tracking control problem is trans...
详细信息
ISBN:
(纸本)9781538629185
This paper tackles the optimal tracking control problem for reconfigurable manipulators based on critic-only policy iteration(Co PI) algorithm. By system transformation, the optimal tracking control problem is transformed into an optimal regulation problem. The optimal tracking controller is composed of the desired controller and the approximate optimal feedback one. The desired controller is developed to maintain the desired tracking performance at the steady-state, while the approximate optimal feedback controller is designed to stabilize the tracking error dynamics in an optimal manner. Then, a critic neural network is used to estimate the optimal performance index function, and the optimal feedback control is obtained by the Co PI algorithm. The convergence of the proposed method is analyzed and it is shown that the closed-loop system based on Co PI is uniformly ultimately bounded by using the Lyapunov approach. Finally, simulation studies are given to show the effectiveness of the developed method.
This paper presents a decentralized optimal control method for modular and reconflgurable robots(MRRs) based on adaptive dynamic ***,the dynamic model of MRRs is formulated by using the Newton-Euler iterative algori...
详细信息
ISBN:
(纸本)9781538629185
This paper presents a decentralized optimal control method for modular and reconflgurable robots(MRRs) based on adaptive dynamic ***,the dynamic model of MRRs is formulated by using the Newton-Euler iterative algorithm,and then the state space description is ***,the optimal control policy of the MRRs system is obtained based on the policy iteration algorithm,which is used to solve the Hamilton-Jacobi-Bellman(HJB) equation via the critic neural ***,the stability of the closed-loop system is proved by using the Lyapunov ***,simulations are conducted to illustrate the effectiveness for the 2-DOF MRRs.
WE are in an exciting new intelligent era where various Web 3.0 systems emerge and flourish.[1]–[3].In this new epoch,the collaboration of data and knowledge,humans and machines,actual and virtual worlds is undergoin...
详细信息
WE are in an exciting new intelligent era where various Web 3.0 systems emerge and flourish.[1]–[3].In this new epoch,the collaboration of data and knowledge,humans and machines,actual and virtual worlds is undergoing an unprecedented diversification and community-driven transformation,unveiling an open future full of boundless ***,the value of dispersed data extends far beyond passive storage and application.
This paper presents a decentralized adaptive super-twisting control method for modular and reconflgurable robots(MRRs) with uncertain environment *** conventional methods that rely on robot-environment contact model...
详细信息
ISBN:
(纸本)9781538629185
This paper presents a decentralized adaptive super-twisting control method for modular and reconflgurable robots(MRRs) with uncertain environment *** conventional methods that rely on robot-environment contact model or force/torque sensing,this paper addresses the problem of controlling MRRs in contact with uncertain environment that using only local dynamic information of each joint *** dynamic model of MRR is formulated as a synthesis of interconnected *** on the integral sliding mode control(ISMC) technique and the adaptive super-twisting algorithm(ASTA),the decentralized controller is designed to compensate the model uncertainty in which the up-bound is *** stability of the MRR system is proved by using the Lyapunov *** last,simulations are conducted for 2-DOF MRRs with different configurations under the situations of dynamic contact and collision to investigate the advantage of the proposed approach.
Privacy preserving in distributed control is getting more attention, and differential privacy (DP) is the common tool to protect data privacy, in which additive noise is applied in the algorithm function. However, DP ...
Privacy preserving in distributed control is getting more attention, and differential privacy (DP) is the common tool to protect data privacy, in which additive noise is applied in the algorithm function. However, DP can be leveraged by false noise (FN) attacks because attack vectors can be disguised as artificial noise in DP. FN attacks are a concern as the stealth attacks are hard to detect. Moreover, DP in distributed control makes FN attack detection more difficult. Hence, detecting FN attacks in privacy-preserving distributed control is critical and challenging. In this paper, taking distributed energy management systems as the control object, we propose a novel peer-to-peer attack detection approach, named False Noise Attack Detection (FNAD). In FNAD, each device observes the power decisions of its neighbors based on the data from its two-hop neighbors, estimates the power decisions of its neighbors by a Kalman filter, and updates the detection index of each neighbor according to the residues of the Kalman filter at each iteration. The detection index is developed based on information entropy, without any prior knowledge of the FN attacks. If a device’s detection index is out of well-defined thresholds, its neighbors can perform a majority vote to decide whether it is malicious. We theoretically prove the detection effect of FNAD against three representative attacks in the literature and analyze the advantages of FNAD compared with the traditional methods. The effectiveness of FNAD is demonstrated by extensive simulations.
Human trajectory prediction is essential and promising in many related applications. This is challenging due to the uncertainty of human behaviors, which can be influenced not only by himself, but also by the surround...
详细信息
Human trajectory prediction is essential and promising in many related applications. This is challenging due to the uncertainty of human behaviors, which can be influenced not only by himself, but also by the surrounding environment. Recent works based on long-short term memory(LSTM) models have brought tremendous improvements on the task of trajectory prediction. However, most of them focus on the spatial influence of humans but ignore the temporal influence. In this paper, we propose a novel spatial-temporal attention(ST-Attention) model,which studies spatial and temporal affinities jointly. Specifically,we introduce an attention mechanism to extract temporal affinity,learning the importance for historical trajectory information at different time instants. To explore spatial affinity, a deep neural network is employed to measure different importance of the neighbors. Experimental results show that our method achieves competitive performance compared with state-of-the-art methods on publicly available datasets.
Accurate estimation for state-of-health (SOH) of lithium-ion batteries is critical to ensure the security and durability of battery operation. Accordingly, this paper proposes a SOH estimation method based on Attentio...
详细信息
In this article, we investigate the problem of fast finite-time target tracking control of the inertially stabilized platform (ISP) with a camera that mounted on a mobile robot. At first, the kinematics of the onboard...
详细信息
To make humanoid robots walking fast, it's important to improve driving force of their leg joints. Usually, each joint of humanoid robots is driven by a single motor. Dual-motor joint, on the other hand, is one of...
详细信息
ISBN:
(纸本)9781467355339
To make humanoid robots walking fast, it's important to improve driving force of their leg joints. Usually, each joint of humanoid robots is driven by a single motor. Dual-motor joint, on the other hand, is one of the candidate solutions to meet the power requirement needed for fast walking. This paper proposed a new dual-motor control model. In the model, two motors are treated as a single control plant instead of two parallel control plants. With the usage of current distributor, the control model can pump different current to each motor freely so as to eliminate the unbalance of the load imposed on each motor. Simulation and experiment show that the proposed model works well under high joint load and it can be used on a fast walking humanoid robot.
暂无评论