Anthony M. Federico

Control fault detection for machine recovery and diagnostics prior to malfunction

The present invention is monitoring the control, in particular, software crashes in a multiprocessor machine control system to prevent machine malfunctions. Benchmarks or key operations tracked include, for example, monitoring that the number of tasks to be completed by the control is not beyond the capacity of the control to respond. These recorded benchmarks enable one or both of two paths of progressively more complex analysis. In particular, there is a diagnostic path that leads first to the isolation of a particular control element in a multi-element control that contains a fault and then to a more comprehensive recording of the fault history for each individual control element. There is also a path for maintaining machine operation. In this path, a first level of response is to merely reset the various control elements. Progressively more comprehensive responses include isolating a fault in an individual control element to maintain machine operation and the full recovery of the job in progress after a software crash.