Unambiguous and Reliable Positioning in the vehicle in terms of Functional Safety and Cyber Security

University of Lörrach, DHBW Lörrach, Germany Department of Electrical Engineering, Functional Safety and Cyber Securtiy

Abstract

Functional security and agile software development are two modern areas in product development, which initially have very opposite approaches. For example, formal tests are required by the relevant standards for the former, which must be documented very extensively. The agile software development, on the other hand, tries to come to its conclusion with as few documentation and flexible tests as possible. Also, the proof that testing and development are independent of each other for safety-critical projects is difficult in the context of the use of agile methods. However, taking the constraints of functional safety as given and taking advantage of the enormous flexibility of agile software development, e.g. With the use of Scrum, the Daily Team Meetings create new opportunities in product development. In contrast to previous positioning methods for linearly movable axles, a new developed approach for rear axle steering has not been used as an absolute value encoder, but a novel positioning concept has been researched and developed. Functional Safety first! A new safety concept must therefore be developed. The absolute value encoder, usually realized as an optical or magnetic bar-coded sensor, is used reliably but cost-effectively in a large number of systems. In order to save costs as well as space, the development of the new approach to the sensor will be dispensed with and the positioning will be realized via a new concept. The conventional concepts for position determination of axes is an absolute value encoder. However, this is not highly reliable and has no redundancy. With the new safety concept, the exact position of an axis can be determined and output with high accuracy by means of the various safety devices directly after switching on the system. As a result, the sensor system is hardly susceptible to errors. Here, a detailed error analysis has been carried out. Even after system crashes, there are enough detection points, which are constantly detected during normal operation and thus the plausibility check can be restored. The new explored approach allows the steering to work normally even in safe modes. However, the algorithms for protection have to take effect immediately if, for example, an expected index signal does not occur.