Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.561668
Title: Safety assurance in interlocking design
Author: Morley, Matthew J.
Awarding Body: University of Edinburgh
Current Institution: University of Edinburgh
Date of Award: 1996
Availability of Full Text:
Access through EThOS:
Full text unavailable from EThOS. Please try the link below.
Access through Institution:
Abstract:
This thesis takes a pedagogical stance in demonstrating how results from theoretical computer science may be applied to yield significant insight into the behaviour of the devices computer systems engineering practice seeks to put in place, and that this is immediately attainable with the present state of the art. The focus for this detailed study is provided by the type of solid state signalling systems currently being deployed throughout mainline British railways. Safety and system reliability concerns dominate in this domain. With such motivation, two issues are tackled: the special problem of software quality assurance in these data-driven control systems, and the broader problem of design dependability. In the former case, the analysis is directed towards proving safety properties of the geographic data which encode the control logic for the railway interlocking; the latter examines the fidelity of the communication protocols upon which the distributed control system depends. The starting point for both avenues of attack is a mathematical model of the interlocking logic that is derived by interpreting the geographic data in process algebra. Thus, the emphasis is on the semantics of the programming language in question, and the kinds of safety properties which can be expressed as invariants of the system's ongoing behaviour. Although the model so derived turns out to be too concrete to be effectual in program verification in general, a careful analysis of the safety proof reveals a simple co-induction argument that leads to a highly efficient proof methodology. From this understanding it is straightforward to mechanise the safety arguments, and a prototype verification system is realised in higher-order logic which uses the proof tactics of the theorem prover to achieve full automation. The other line of inquiry considers whether the integrity of the overall design that coordinates the activities of many concurrent control elements can be compromised. Therefore, the formal model is developed to specifically answer safety-related concerns about the protocol employed to achieve distributed control in the management of larger railway networks. The exercise reveals that moderately serious design flaws do exist, but the real value of the mathematical model is twofold: it makes explicit one's assumptions about the conditions under which the faults can and cannot be activated, and it provides a framework in which to prove a simple modification to the design recovers complete security at negligible cost to performance.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.561668  DOI: Not available
Share: