Performance analysis of task offloading based on compressed sensing in cloud-edge-device architecture

To address the challenges encountered by wireless sensor networks (WSN) in processing complex computational tasks due to limitations in energy and computational resources, this paper has devised a three-tier architecture that integrates device, edge, and cloud layers. By incorporating Mobile Edge Computing (MEC) technology, the architecture aims to reduce the overall network energy consumption and enhance data processing efficiency. Within this framework, sensor nodes have the flexibility to offload computational tasks to MEC or cloud servers, or execute them locally, adapting to various scenarios and demands. Considering the increased energy consumption associated with transmitting large volumes of task data to the cloud when MEC opts for cloud offloading, this paper employs compressed sensing technology to compress the task data, thereby reducing the data transmission volume and energy consumption. Finally, a comprehensive model based on compressed sensing is constructed and analyzed, taking into account factors such as task volume, compression rate, and transmission distance. Through a variety of experimental scenarios and parameter settings, the paper comprehensively evaluates two critical performance metrics: task accomplishment time and energy consumption. According to the experimental results, specific suggestions are proposed to guide the arrangement of the timing and location of actual task offloading, with the aim of maximizing network performance and energy efficiency.