Automatic Test Suite Generation for PLC Software in the Internet of Production

Automatic test suite generation is an established technique used to generate test suites adhering to structural coverage metrics of PLC software. In order to reduce redundancy in the test suite generation after a structural reconfiguration to the PLC software has occurred, reusable summaries of program parts should be employed. This paper presents a combination of state-of-the-art symbolic execution and static analysis algorithms for test suite generation and summary reuse. The general rationale is to improve efficiency by not doing redundant work. For this purpose, summaries of function blocks are cached and reused to benefit from the previous analysis. As code untouched from reconfigurations will result in equivalent path conditions summaries can aid in speeding up regression testing. The proto-typical implementations of several techniques are evaluated and compared using selected domain-specific benchmarks showing the ineffectiveness of using summarization during test suite generation for reconfigurable logic control software.

I. Introduction

Manufacturing experienced a paradigm shift through the advent of the Industrial Internet of Things (IIoT). The life cycles and value chains of conventional production systems are too rigid with regards to the increasing agility in the development process and changing customer and legislator requirements [1]. The main enabler for the increase in cyber-physical production systems (CPPSs) complexity is transformability, a property composed of flexibility and reconfigurability [2]. The goal of the Internet of Production (IoP) is to enable transformable CPPSs that are capable of coping with the highly iterative development and agile manufacturing practices. However, the ability to transform the CPPS during its operational phase takes a toll on the functional safety and reliability of software components [3]. While reconfigurability of the industrial control software helps in quickly adapting, each reconfiguration might also introduce regressions. In order to reduce long downtimes during static reconfigurations in which the entire system is stopped for maintenance, efficient regression verification and testing techniques are mandatory to reduce the time it takes to ensure the correctness of the reconfigured programmable logic controller (PLC) software. The subsequent section gives a brief survey of related approaches in the field of validating PLC software after a reconfiguration.