Search Results

On the way to automated driving, the installation rate of surround sensing systems will rapidly increase in the upcoming years. The respective technical progress in the areas of driver assistance and active safety leads to a numerous and valuable information and signals to be used prior to, during and even after an accident. Car makers and suppliers can make use of this new situation and develop integrated safety functions to further reduce the number of injured and even deaths in car accidents. Nevertheless, the base occupant safety remains the core of this integrated safety system in order to ensure at least a state-of-the-art protection even in vehicles including partial, high or full automation. Current networked safety systems comprehend a point-to-point connection between single components of active and safety systems. The optimal integration requires a much deeper and holistic approach.

With the publication of ISO26262 [1] and the concept of Functional Safety, being able to identify the required safety integrity level for software components and defining the respective development steps has become increasingly important. A number of Tier 1 automotive suppliers, including Robert Bosch LLC, have been developing software for safety relevant systems, and have experience with a number of methods and tools for software analysis. This paper will focus on the pros and cons of the Criticality Analysis method. Criticality Analysis (CA) is a method that rates outputs, sub-components and inputs to a function based on the ASIL rating of the function. Faller [2] proposed the use of CA in conjunction with IEC 61508 safety standard, and this author proposes that the CA can also be used in conjunction with ISO 26262. CA allows taking a function with any ASIL rating and breaking down the signal chain to develop safety requirements at each stage (see [2, 3]).

This paper describes an investigation about the efficiency of safety-related non-functional software unit tests (NFSWUT). Well defined design, implementation and test processes are widely used in the respective industry. In order to fulfill the ISO 26262[1] requirement, additional effort is necessary to execute the NFSWUT. However, the efficiency of these tests is still not confirmed. This paper will provide an overview about an investigation of the effort-benefit ratio of the NFSWUT.