Modeling of Power Consumption of Video Transmission in Internet of Things Devices with Acceptable Quality of Experience

Abstract

Internet of Things applications such as environmental monitoring and healthcare may involve multimedia communications from IoT devices to humans for decision-making. Therefore, the delivered multimedia should be in good perceived quality. Higher video quality results into higher energy consumption due to encoding and decoding processes and as a result, will affect the performance of IoT devices due to their inherent energy constraints. This thesis presents the impact of video encoding parameters on the energy consumption of IoT devices. The experimental results from Cooja simulator show that the videos with high bitrates and low frame rates consume more power than videos with low bitrates and high frame rates. It was also found that videos with high movement consume more power than videos with medium and slow movement. Also this thesis proposes a power model that takes into account video parameters such as bit rate, frame rate and content types. The proposed model can play a vital role in video quality adaptation in multimedia communication over IoT devices. The experiment results from the subjective test show that the video with the same bitrate ,frame rate and resolution, slow and medium movement recorded better MOS values and hence better quality than fast movement video sequence. Finally, this thesis presents the impact of gender on the Quality of Experience for video services. Initial results from subjective tests suggested that male viewers request higher video quality compared to female viewers, while female viewers possibly concentrate more on the content of the video.