Scalable Component-based Model Revision of Cyber-Physical Systems with Separation of Concerns
pdf
Abstract:
As part of our CPS project, we have focused on the problem of model repair for cyber-physical systems. This work involves identifying constraints caused due to physical components during revision. We consider four types of constraints cyber-cyber, cyber-physical, physical-cyber and physical-physical. Based on the complexity limitations caused by these constraints we are developing efficient heuristics to mitigate the cost of model repair. We have also focused on extending revision to code level. In particular, we have focused on the problem of repairing a given C program to remove deadlocks and livelocks. We are also developing algorithms for repairing programs in SystemC to add fault-tolerance. We are also developing methods that allow addition of fault-tolerance in the presence of a potentially collaborative and potentially disruptive environment. This approach would help in the context of cyber-physical systems since it can guarantee that physical components remain unchanged during revision. As part of the current activities, we have developed algorithms and tools for analysis of faults in these programs. The results from this work show that the cost of fault-impact analysis is comparable with that of fault-free analysis.
Submitted by Anonymous
on
Abstract:
As part of our CPS project, we have focused on the problem of model repair for cyber-physical systems. This work involves identifying constraints caused due to physical components during revision. We consider four types of constraints cyber-cyber, cyber-physical, physical-cyber and physical-physical. Based on the complexity limitations caused by these constraints we are developing efficient heuristics to mitigate the cost of model repair. We have also focused on extending revision to code level. In particular, we have focused on the problem of repairing a given C program to remove deadlocks and livelocks. We are also developing algorithms for repairing programs in SystemC to add fault-tolerance. We are also developing methods that allow addition of fault-tolerance in the presence of a potentially collaborative and potentially disruptive environment. This approach would help in the context of cyber-physical systems since it can guarantee that physical components remain unchanged during revision. As part of the current activities, we have developed algorithms and tools for analysis of faults in these programs. The results from this work show that the cost of fault-impact analysis is comparable with that of fault-free analysis.