Self-repair Measurement in FPGA-Based Partial Reconfigurable Systems

Intelligent Embedded Systems (IES) represent a new discipline in which Artificial Intelligence (AI) is coupled with Embedded Systems (ESs) in order to create a new self-X-based Embedded Systems generation. Self-X capabilities, like: self-repair, self-awareness, self-adaptation, etc. brought, for an ES, the ability to reason about their external environments and, as a result, adapt their behavior appropriately. Also, self-repair capability defines the fact of identifying and repairing failures in order to avoid the high cost of the ordinary repair process. FPGA (Field Programmable Gate Arrays) provide support for implementing the self-repair capability based on FPGA Partial Reconfiguration (PR) feature. Despite that there are many approaches which proposed self-repair processes, the measurement aspects are often omitted. Indeed, the measurement allows controlling the repair process and comparing the different approaches. In this paper, we discuss the self-repair capability measurement by proposing a common FPGA PR self-repair process followed by a set of hardware (HW) and software (SW) metrics dedicated particularly for FPGA PR-based IES.