Search

Analysis of human interaction is one important research topic of human motion analysis. It has been studied either using first person vision (FPV) or third person vision (TPV). However, the joint learning of both types of vision has so far attracted little attention. One of the reasons is the lack of suitable datasets that cover both FPV and TPV. In addition, existing benchmark datasets of either FPV or TPV have several limitations, including the limited number of samples, participant subjects, interaction categories, and modalities. In this work, we contribute a large-scale human interaction dataset, namely, FT-HID dataset. FT-HID contains pair-aligned samples of first person and third person visions. The dataset was collected from 109 distinct subjects and has more than 90K samples for three modalities. The dataset has been validated by using several existing action recognition methods. In addition, we introduce a novel multi-view interaction mechanism for skeleton sequences, and a joint learning multi-stream framework for first person and third person visions. Both methods yield promising results on the FT-HID dataset. It is expected that the introduction of this vision-aligned large-scale dataset will promote the development of both FPV and TPV, and their joint learning techniques for human action analysis. The dataset and code are available at \href{https://github.com/ENDLICHERE/FT-HID}{here}.

To realize low-carbon energy systems, distributed energy storage systems and flexible loads have been integrated into power grids. System reliability, economy, and resilience, therefore, face significant challenges. This article presents modeling of a distributed energy micro-grid including wind turbines, micro gas turbines, waste heat recovery devices, electric boilers, direct-fired boilers, battery energy storage, interruptible loads, and transferable loads. At the same time, the optimal configuration of energy storage and the demand-side response modeling are studied, and the combined optimization control strategy of the two is demonstrated. The simulation results indicate that the proposed control strategy has better performance than the traditional operation. In addition, this article also clarifies the impact of control strategy on distribution system resilience. The results show that the control strategy proposed in this article can achieve the resource complementarity of demand-side response and energy storage, and realize the integrated coordination of source, network, load, and storage. The distributed energy micro-grid under this control strategy has the best overall economic benefit and the best capacity to accommodate load growth.

Zero-Shot Action Recognition (ZSAR) aims to recognize video actions that have never been seen during training. Most existing methods assume a shared semantic space between seen and unseen actions and intend to directly learn a mapping from a visual space to the semantic space. This approach has been challenged by the semantic gap between the visual space and semantic space. This paper presents a novel method that uses object semantics as privileged information to narrow the semantic gap and, hence, effectively, assist the learning. In particular, a simple hallucination network is proposed to implicitly extract object semantics during testing without explicitly extracting objects and a cross-attention module is developed to augment visual feature with the object semantics. Experiments on the Olympic Sports, HMDB51 and UCF101 datasets have shown that the proposed method outperforms the state-of-the-art methods by a large margin.

The traditional OV model and Gipps model have problems such as limited applicability frequent fluctuations of acceleration and so on when applied to autonomous driving vehicles. The driving experiment was carried out by using the driving simulator, and the driving state of vehicle was divided according to the experimental analysis results. Based on the driving targets in different states of the vehicle, a multi-mode autonomous car-following model based on driving simulator was proposed to realize the safe and stable operation of autonomous driving vehicle. The results show that the model is more stable in the driving simulator simulation environment, and it provides a basis for the research of autonomous driving behavior.

In this paper, a new color image encryption scheme based on DNA operations and spatiotemporal chaotic system is presented. Firstly, to hide the distribution information of the plain image, we convert the plain image into three DNA matrices based on the DNA random encoding rules. Then, the DNA matrices are combined into a new matrix and is permutated by a scramble matrix generated by mixed linear-nonlinear coupled map lattices (MLNCML) system. In which, the key streams are associated with the secret keys and plain image, which can ensure our cryptosystem plain-image-dependent and improve the ability to resist known-plaintext or chosen-plaintext attacks. Thereafter, to resist statistical attacks, the scrambled matrix is decomposed into three matrices and diffused by DNA deletion-insertion operations. Finally, the three matrices are decoded based on DNA random decoding rules and recombined to three channels of the cipher image. Simulation results demonstrate that the proposed image cryptosystem has good security and can resist various potential attacks.

Rehabilitation robots have been widely used in clinical rehabilitation in stroke subjects. The safety and comfort in rehabilitation training still are affected by many problems, however, such as the single rehabilitation training mode, the poor human-robot interaction and adaptability. In this paper, an adaptive trajectory planning method of lower limb rehabilitation robot based on surface Electromyography (sEMG) signal and interactive force (IF) is proposed to improve the human-robot interaction and adaptability. Firstly, the envelope difference method is used to identify flexion or extension of the lower limb based on the surface EMG signals from biceps femoris and vastus lateralis muscles; Secondly, the sEMG-IF fusion model is established on the plantar pressure and recognition result of flexion or extension with logical reasoning, and then the velocity trajectory of the robot can be planned with the model; Meanwhile, the evaluation of muscle activity level is introduced to adjust the robot's auxiliary coefficients and modify the velocity trajectory adaptively. In this paper, the corresponding three velocity trajectories of the robot were planned based on EMG signals during the three periodic movements. The result showed that the proposed method is feasibility and effectiveness.