Scope
The paper’s scope extends beyond theoretical analysis to practical implications for real-world applications, aiming to contribute to the development of optimal solutions for energy-efficient, real-time ODV transmission via UAVs.
Summary
The paper presents a comprehensive analysis of energy consumption and coding efficiency in real-time omnidirectional video (ODV) transmission via Unmanned Aerial Vehicles (UAVs). It explores the challenges and requirements inherent in such applications, emphasizing the need for low-latency, energy-efficient solutions to ensure a superior Quality-of-Experience (QoE) while operating within the constraints of UAV energy budgets.
Critical areas of the paper include:
Energy Cost Analysis: Investigates the energy costs associated with encoding and decoding UAV ODV sequences, focusing on hardware versus software implementations across different platforms (ARM and x86). It encompasses various video coding standards such as AVC, HEVC, AV1, and VP9.
Encoding Efficiency: Examining the trade-offs between encoding efficiency, bitrate, and quality metrics to optimise ODV transmission pipelines. This involves considering factors like resolution, latency, and computational resources to achieve high-quality transmission while minimising energy expenditure.
Hardware and Software Encoders: Assessing the performance of both software and hardware video encoders across different architectures and devices. This includes evaluating their energy consumption, coding efficiency, and real-time capabilities, with insights into the advantages and limitations of each approach.
Testing Environment and Methodology: Utilising diverse datasets and testing environments to evaluate video encoders comprehensively. This involves describing the dataset used, the testing hardware (including CPUs and GPUs), and the energy measurement methodology (e.g., RAPL interface, Jetstats tool).
Optimisation Strategies: Providing recommendations and insights for designing energy-efficient video streaming pipelines tailored to UAV-based ODV transmission. This includes optimising encoding parameters, selecting appropriate codecs, and leveraging hardware acceleration where feasible.
The investigation illuminates the energy costs of encoding and decoding UAV ODV sequences, highlighting the advantages of hardware-based implementations over software-based ones across platforms like ARM and x86. Through an in-depth examination of various video coding standards, including AVC, HEVC, AV1, and VP9, the study elucidates the intricate trade-offs between encoding efficiency, bitrate, and quality metrics essential for optimising ODV transmission pipelines.
Relevance for EXIGENCE
Relevant for energy measurements, resource interconnectivity and orchestration, focusing on the compute nodes side, the metrics provided by the interfaces and tools to measure the consumption of energy could be used for providing an initial direction to focus on differences for hardware and software side (time to complete task and energy consumption while a compute task is running). On the other hand, the focus on encoders for video decoding only highlights a specific area for the energy cost measurements.
