WCSE

Prototype Framework of Low-Latency Collection of Multi-Point Video Stream over Multi-Hop Multi-Channel Wireless Network

Abstract— Big cit ies are facing heavy road traffic congestion. The main reason is the increasing number of vehicles beyond the road network capacity, which forces traffic police at t imes to be incapable of controlling the inevitable traffic jams. To resolve those traffic jams, traffic police need the near real-time informat ion of the road network e.g. from traffic cameras. In this paper, a prototype framework of low-latency collection of multi-point video streams over a multi-hop multi-channel wireless network is proposed. Video frames are captured by Raspberry Pi cameras and sent by the wireless link to the herein-called IoT Concentrators. IoTCloud Hub then pulls and stores those video frames from the IoT Concentrators. Users (traffic police) can watch traffic videos from the display screen of the IoT-Cloud Hub. As a proof-of-concept implementation, two Pi Came ras have been set up with Apache Kafka Producers and two additional Raspberry Pi’s have been set up as the IoT Concentrators, which run the leader and follower Apache Kafka Brokers. To complete our system, an Intel® NUC as IoT-Cloud Hub and another Intel® NUC as Orchestration-Hub have been also set up. We assume the available wired communicat ion linkage between the Orchestration -Hub and the other devices. However, the links between the Pi Cameras, IoT Concentrators and IoT-Cloud Hub are wireless with Wi-Fi. To achieve a high performance, mult iple Wi-Fi channels can be dynamically configured by the Orchestration-Hub to steer the VDO t raffic (based on TCP) away from the congested channels. Based on the implementation prototype, the average TCP throughputs over all the wireless links have been improved by 56.046% on average, compared with the case without the dynamic Wi-Fi channel selection.

Index Terms— Apache Kafka, Raspberry Pi, video streaming, wireless

Aung Myo Htut, Chaodit Aswakul
Wireless Network and Future Internet Research Unit, Department of Electrical Engineering, Faculty of Engineering, Chulalongkorn University,THAILAND
Juseong Kim, JongWon Kim
Networked Computing Systems (NetCS) Laboratory, School of Electrical Engineering and Computer Science, Gwangju Institute of Science and Technology, KOREA

[Download]

Cite: Aung Myo Htut, Juseong Kim, JongWon Kim, Chaodit Aswakul, "Prototype Framework of Low-Latency Collection of Multi-Point Video Stream over Multi-Hop Multi-Channel Wireless Network," Proceedings of 2019 the 9th International Workshop on Computer Science and Engineering WCSE_2019_SPRING, pp. 11-16, Yangon, Myanmar, February 27-March 1, 2019.