No abstract

Edge-cloud collaborative video analytics is transforming the way data is being handled, processed, and transmitted from the ever-growing number of surveillance cameras around the world. To avoid wasting limited bandwidth on unrelated content transmission, existing video analytics solutions usually perform temporal or spatial filtering to realize aggressive compression of irrelevant pixels. However, most of them work in a context-agnostic way while being oblivious to the circumstances where the video content is happening and the context-dependent characteristics under the hood. In this work, we propose VaBUS, a real-time video analytics system that leverages the rich contextual information of surveillance cameras to reduce bandwidth consumption for semantics compression. As a task-oriented communication system, VaBUS dynamically maintains the background image of the video on the edge with minimal system overhead and sends only highly confident Region of Interests (RoIs) to the cloud through adaptive weighting and encoding. With a lightweight experience-driven learning module, VaBUS is able to achieve high offline inference accuracy even when network congestion occurs. Experimental results show that VaBUS reduces bandwidth consumption by 25.0%-76.9% while achieving 90.7% accuracy for both the object detection and human keypoint detection tasks.Index Terms-edge-cloud collaborative computing, semantic compression, video analytics, task-oriented communication system I. INTRODUCTIONT HE recent advances of deep learning techniques boost the performance of many computer vision applications, such as object detection, semantic segmentation and keypoint detection. These advances facilitate the commercialization of video analytics applications including traffic control [1], video surveillance [2], and safety anomaly detection [3]. In order to make video analytics technologies ready to use, three main challenges need to be addressed, i.e., latency, bandwidth and accuracy. First, analytics applications require very low latency, Manuscript

Dockless sharing bikes play an increasingly significant role in transit transfer, especially for the first/last mile. However, it is not always accessible for users to find sharing bicycles. The objective of this paper is to assess the accessibility of dockless sharing bikes from a network perspective, which would provide a decision-making basis not only for potential bike users but also for urban planners, policymakers, and bicycle suppliers to optimize sharing-bike systems. Considering bicycle travel characteristics, a hierarchical clustering algorithm was applied to construct the dockless sharing-bike network. The social network analysis (SNA) method was adopted to assess the accessibility of the bike network. Then, a spatial interaction model was chosen to conduct a correlation analysis to compare the accessibility obtained from the SNA approach. The case study of Shanghai indicates a strong connection between the accessibility and the SNA indicators with the correlation coefficient of 0.779, which demonstrates the feasibility of the proposed method. This paper contributes to a deep understanding of dockless sharing-bike network accessibility since the SNA approach considers both the interaction barriers and the network structure of a bicycle network. The developed methodology requires fewer data and is easy to operate. Thus, it can serve as a tool to facilitate the smart management of sharing bikes for improving a sustainable transportation system.

No abstract