Abstract: This paper reports on an industrial pilot project that introduces systematic, automated module testing for embedded software in distributed, real-time, control systems. The systems are used in safety-related applications, are complex in nature, and hence have strong requirements for test coverage, auditability and repeatability. This paper explores issues of isolating modules from the run-time environment, improving integration of testing into the development environment, automating testing, and improving test planning and documentation. Metrics were gathered throughout the project that allow a coarse cost-benefit evaluation. Code coverage metrics for statement and branch coverage were also gathered using a commercial code coverage analysis tool. The testing exposed a number of latent faults within the software, and the overall results of the project show that module testing is feasible for this complex, embedded software.