Design and Implementation of an Embedded Software Online Testing System

Sammanfattning

With the deep integration of embedded products in consumer electronics and automotive fields, the quality of embedded software directly affects product safety and user experience. Aiming at the core defects of traditional embedded software testing methods, such as environment simulation distortion, single verification dimension, and insufficient real-time performance, this study proposes an online testing system based on the Assertion-Execution- Path-Timing (AEPT) model. The system achieves functional correctness verification through a configurable macro assertion library, traces the program execution path using a checkpoint chain list algorithm, and constructs a timestamp difference model to accurately measure the nonlinear timing behavior. The probe function library (developed in C language) and the client (developed in C) work together through a multi-channel communication protocol, supporting dynamic result verification, execution flow tracking and performance monitoring functions, and realizing real-time analysis and database management of test data. Experimental results show that the system achieves 98% assertion verification accuracy in execution result judgment, 96.2% critical path branch coverage, and timing measurement error is less than 0.8% of the target board hardware clock accuracy. This study provides a multi-dimensional solution for dynamic behavioral verification of embedded software. In the future, we will further explore the direction of machine- learning-driven intelligent test case generation and distributed test framework optimization to enhance the testing efficiency and reliability of complex embedded systems.