Development of innovative features such as advanced driver assistance systems in modern day automobiles have led to an increased complexity in product development and maintenance. This imposes an increased risk in terms of system failure that could lead to unacceptable hazards. Thus it becomes crucial to ensure functional safety. The ISO 26262 standard [15] defines functional safety for automotive Electric/Electronic (E/E) safety-related systems. Its objective is to address possible hazards caused by the malfunctioning behavior of E/E systems throughout the product development cycle.
Most of the automotive companies have already started using safety analysis, verification and validation techniques to ensure vehicle safety [22]. One of the main objectives of the ISO 26262 is that these techniques should be applied as a standardized methodology for all automobile manufacturers. These techniques focus mainly on traceability which is the ability to track the safety requirements from initial concept design till the production and operation phase. Upon trying to improve the traceability, the researchers seek more techniques for effective product development process.
The introduction of the ISO 26262 functional safety standard provides more specific development processes that help to avoid the hazards and threats in the development phases. Following steps should be taken to ensure compliance with the standard: a) The manufacturers should adopt the development processes; b) The manufacturers should determine the Automotive Safety Integrity Level (ASIL) for safety-critical systems; c) The manufacturers should satisfy the additional requirements.
The standardization process requires the consistency of methods, languages and tools across all the sub-phases of the software lifecycle as well as system and hardware development phases as stated in the section 5.4.4 of the ISO 26262 Part 6 [15, p. 4]. In recent years, safety related platforms such as OPENCOSS [6] and AutoFOCUS3 [2] have been developed. OPENCOSS provides a common safety certification platform for the railway, avionics and automotive markets. AutoFO-CUS3 provides a model-based tool for distributed, reactive, embedded software systems. The consistency can be assured through the availability of a tool that ensures the compatibility within the ISO 26262 (sub-) phases. The automobile manufacturers are challenged in the selection of the optimal techniques to ensure this compatibility which helps to prove the functional safety. This paper focuses on examining the gap between the ISO 26262 standard objectives and state-of-the-art safety related techniques.
The remainder of the paper is organized as follows: In Section 2 we provide background information on the V-model of the ISO 26262 standard. In Section 3, we describe the systematic literature review process and the summary of the papers selected for the analysis. Section 4 presents the gap analysis results and Section 5 discusses the findings. Finally, we present the concluding remarks and some related future works.
The safety standard ISO 26262 [15] is an adaptation of the functional safety standard IEC 61508 [14] for automotive E/E systems. Similar to IEC 61508, ISO 26262 is also a risk-based safety standard. It provides a risk-driven safety life-cycle for developing safety-critical systems in the automotive domain.
The ISO 26262 consists of ten parts as shown in Figure 1. Part 1, 2, and Part 8 to 10 are out of the scope of this paper, because Part 3 to Part 7 correspond to the safety life-cycle. The main part of ISO 26262 is structured based upon the V-model, as well as Part 5 and Part 6. Part 3 and Part 7 focus on the vehicle level. The main goal of Part 3 is to identify system hazards and risks through Hazard Analysis and Risk Assessment (HARA), then derive safety goals and Functional Safety Concepts (FSC) from them. Part 4 focuses on the system level. In this part, Technical Safety Requirements (TSR) are derived from FSC. Then system design can be carried out based on TSR. Part 5 and Part 6 focus on the subsystem/component level. In these two parts more detailed safety requirements are derived from TSR. Those safety requirements are assigned to the concrete subsystems or components for implementation.
In the following section, we present state-of-the-art techniques complying to the ISO 26262 standard. Peer-reviewed articles on the topics "ISO 26262" and "vehicle safety", published between 2008 and 2015, are included. We exclude duplicate reports of the same or similar studies as well as white papers are excluded. After the search and inclusion/exclusion processes, we identify 120 unique papers. In our findings, we discover that higher number of papers are published in the concept phase (63 papers) than the development phases (51 papers) i.e., product development, software development, and hardware development phases of the ISO 26262 Vmodel. The remaining six papers are considered as general publications, since they cover all the phases of the V-model. To further narrow down the search results, citations are used as a key tool to assess the quality of the identified papers. Publications between 2013 and 2015 are included.
In the case of concept and product development phases, more than half of the papers have been cited at least once and number of papers cited more than five are 18. Figure 2 shows the trend of papers published in each sub-phases from the selected sources. It can be inferred that the focus of the papers are more on the improvement of FSC (Functional Safety Concepts) in the conceptual phase and IVTA (Integration,Validation,Testing and Assessment) in the development phase. This shows the following observations: -More additional standardized procedures have been implemented from the IEC 61508 standard on the conceptual and development phases where automobile manufacturers required clear process for implementation. -Engineers and researchers were involved in the development of methodologies to ensure safety compliance of the system at these phases.
The summary of the selected papers mapped to the standard phases is presented in Table 1. Following section presents the gap analysis results between the ISO 26262 standard and the techniques identified from the selected papers.
A gap analysis helps to understand the shortcoming of existing approaches suggested by literatures. The gap analysis is carried out between the ISO 26262 objectives of the Part 3 till Part 7 sub-phases. Key objective is to support an adequate understanding of the item so that the activities in subsequent phases can be performed.
Need further analysis to understand the method used for item definition.
Need further analysis to understand the gap.
To derive the functional safety requirements and allocate them to the architectural elements of the item.
Enhanced architecture description language techniques are developed that helps for allocation and reduce ambiguity.
Need further analysis to understand the gap.
Table 2 summarizes the finding of a gap analysis for the concept phase. In the area of Hazard Analysis and Risk Assessment (HARA), various techniques are available to identify and categorize the hazards. Techniques suggested by the literature elucidate the way of estimating the hazard parameters (i.e., severity, exposure and controllability) and help to formulate the safety goals. After identifying the safety goals, safety requirements can be derived for each goals. Literature provides more options for writing the requirement by different notations [10]. Once the requirements are elicited, they are allocated to the relevant architectural elements. This is performed using various architecture description languages such as EAST-ADL [17] and AADL [9]. Though existing techniques fulfill the objectives given for HARA in the standard, more techniques are required to achieve this effort. There is no standard common method or tool suggested by literature for meeting this objective. This is found to be one of the gaps by contrasting the standard objectives and literature approaches. A gap analysis for other sub-phases of the concept phase To ensure all the safety cases generated in the concept phase are validated.
As mentioned above, to verify the design and concepts compilation with the specification, fewer tools are developed that also ensures the traceability.
Need for enhanced tool that integrates both design and verification process together.
To verify whether TSR comply with the FSR. To manage the system requirements with complete traceability across the product life cycle.
Fewer tools like IBM Rational Team Concert, PTC Integrity, Papyrus are developed for requirement specification to improve the traceability. But the detailed semantic traceability for each sub phase has not been explored.
Integration, testing and validation.
To test compliance with each safety requirement and to verify the system design covering those requirements.
Methods are available that ensures the requirement traceability and verify the system design compliance.
Need for tool that combines all the sub phases of the product development.
i.e., Item Definition, Functional Safety Concept, and ASIL [13,19,20,12,18] is presented in the Table 2.
From the gap analysis of the product development phase, it is observed that there are few tools [23,21] suggested by literature and industrial technical report for requirement specification. These tools support only for specific sub-phases and there are more opportunities to integrate these tools with testing and validation tools [4,17]. By this integration, it becomes more sophisticated to perform all the activities of a phase using single technique. This also gives clear way of understanding the standard norms to the developers and verifying it by testers using same platform. The finding of this gap analysis can be found in the Table 3 on the previous page.
Need further analysis to understand the gap.
To specify and implement the software units identifies as specified in accordance with software design and the associated software safety requirements.
Need further analysis to understand the methods used for unit testing.
To demonstrate the software units fulfil the software unit design specification and do not contain undesired functionality.
Need further analysis to understand the methods used for implementation process.
Need further analysis to understand the gap.
To demonstrate that the embedded software fulfils the software safety requirements
As mentioned in the previous phase, to verify the safety requirements with the software, fewer tools are developed that also ensures the traceability.
Need for enhanced tool that integrates both design and verification process together.
To develop and verify the architectural design that realizes the software safety requirements.
Several methods such as GSN (Goal Structuring Notation) are used to reduce the developing Cost and time. This also helps for verification with the safety requirements.
Need tools for integrating the architectural design and verification the safety Requirements with the elements.
Similar to the system architecture level, more techniques are used for the software level [11]. Some of the common architecture description languages are EAST-ADL [17] and AADL [9] which help to reduce the development cost and time.
In addition, such techniques provide a way to make the verification of safety requirements easier. But there is no tool available that integrates both architectural design and safety verification together. This is found to be one of the gap. Table 4 on the previous page shows the gap analysis performed for the software development phase.
In the case of hardware development phase, only few literatures are published about the development required for the evaluation of safety violation. These literatures provide techniques mainly to support two claims. One is hardware architectural metrics and second is evaluation of safety goal violations. Techniques like UML based meta-model [9] support for design process and help to To demonstrate that the hardware fulfils the hardware safety requirements As mentioned in the product development phase, to verify the safety requirements with the components, fewer tools are developed that also ensures the traceability.
Need for enhanced tool that integrates both design and verification process together.
Need further analysis to understand the gap.
Need more analysis to understand the evaluation techniques used along with the hardware architectural metrics.
To demonstrate the compliance of the design with the safety metrics.
To demonstrate the hardware components fulfil the hardware design specification and do not contain undesired functionality.
Need more analysis to understand the hardware testing procedures.
Need further analysis to understand the gap. perform safety evaluation in a unified model based environment. The findings of the gap analysis for the hardware development phase are shown in the Table 5. Following section discusses the main results of the gap analysis.
Based on the gap analysis, the shortcoming and challenges of the techniques suggested by literature while fulfilling the standard objectives are found. In the concept phase, gap analysis identified the lack of mature techniques that provide wider possible solutions for ASIL decomposition. It showcases the opportunity for integrating various techniques within the phase. For product development phase, gap analysis shows similar results. There are tools used for each sub phases of the product development but there is no common platform where all sub phase activities can be performed. This tool integration could facilitate the understanding and correct interpretation of the standard norms.
For the software and hardware development phase, same type of architecture description languages, such as EAST-ADL and AADL, are used. But there is a lack of common platform that supports both design and safety evaluations.
Since the ISO 26262 standard does not specify which techniques to be applied in fulfilling the safety requirements, variety of techniques are developed for each phase of the ISO 26262 standard. However, a general overview of existing and emerging ISO 26262 related techniques is lacking. Therefore, in this paper, we carried out a gap analysis to identify the challenges and future trends to fulfill the ISO 26262 (part 3 to Part 7) safety objectives. We identified that the focus of research techniques is for the concept and product development phases. However, more techniques are needed for fulfilling the objectives of the software and hardware phases.
As a future work, we plan to conduct similar study on the remaining phases of the ISO 26262 and develop a method for the software and hardware development phases. Furthermore, our analysis focused on the research results rather than the practical application of the standard. This requires further survey on the gap between research results and the practical applicability of the standard to reflect the actual situation in the automotive industry.