To reduce the power consumption of autonomous robots with Artificial Intelligence (AI), an edge computing method called multi-access edge computing (MEC) is expected to offload processing to high-performance computers in close proximity via high-speed 5G wireless communications. Furthermore, field programmable gate arrays (FPGAs) are anticipated to play a crucial role as computing resources within MEC due to their low power consumption and high-speed parallel processing capabilities. In this research, we introduce an offloading method employing MEC-RM, a resource management middleware designed for MEC, aimed at reducing response times for image recognition processing on robots. MEC-RM serves as middleware enabling the offloading of processing tasks to compute resources like FPGAs and GPGPUs through the transmission of JSON-RPC requests from edge devices to a server responsible for resource management. This paper presents the evaluation results of response times when employing the proposed method to offload image recognition processing to MEC's FPGA and communication performance in a local 5G environment.

To reduce the power consumption of autonomous robots with Artificial Intelligence (AI), an edge computing method called multi-access edge computing (MEC) is expected to offload processing to high-performance computers in close proximity via high-speed 5G wireless communications. Furthermore, field programmable gate arrays (FPGAs) are anticipated to play a crucial role as computing resources within MEC due to their low power consumption and high-speed parallel processing capabilities. In this research, we introduce an offloading method employing MEC-RM, a resource management middleware designed for MEC, aimed at reducing response times for image recognition processing on robots. MEC-RM serves as middleware enabling the offloading of processing tasks to compute resources like FPGAs and GPGPUs through the transmission of JSON-RPC requests from edge devices to a server responsible for resource management. This paper presents the evaluation results of response times when employing the proposed method to offload image recognition processing to MEC's FPGA and communication performance in a local 5G environment.