Method and device for monitoring a process execution

Abstract

A method for monitoring a process execution of a plurality of sequentially executed processes starts one of a plurality of timers in cyclic permutation when one of the processes is started, and outputs a first error signal when a period of time recorded by one of the timers exceeds a predefined maximum period of time.

Description

The present invention relates to a method for monitoring a process execution and to a device for implementing the method according to the present invention.

BACKGROUND INFORMATION

Numerous control methods and their corresponding devices provide for regularly recording signals from sensors, for analyzing the recorded signals, and for generating appropriate control signals in response to the analysis. An example is a control system for triggering airbags which regularly executes a process that includes the following incremental steps: recording the signals from sensors, processing the signals, and generating a control signal for the airbag.

A malfunction of the real-time system or of the executed processes can lead to a triggering of the airbag. This is prevented in that the real-time system monitors the execution of the processes and, in response to detection of a malfunction, an outputting of a control signal to the airbag is blocked.

A method for clarifying a problem definition of the exemplary embodiment and/or exemplary method of the present invention is shown in FIG. 8. Processes P₆₀, P₆₁, P₆₂ are started by a real-time system S′ in accordance with the time-slot pattern at points in time t₆₁, t₆₂ at time interval dT. Two processes P₆₀, P₆₁ are combined into pairs, respectively, and, in response to the starting of first process P₆₀ of the two processes, a counter Z′ is started, and, in response to the ending of second process P₆₁, counter Z′ is stopped. Thus, counter count C₆₀ is a measure of the entire execution duration of the two processes P₆₀, P₆₁. If counter count C₆₂ exceeds a value predefined by a maximum period of time T_Max, an overflow error message F₁ is output. If a malfunction occurs during a first process P₆₂ of a pair, then, starting with the starting point in time t₆₂ of this process P₆₂, the maximum period of time T′_Max lapses until error signal F′₁ is output. This is undesirably long, in particular since the maximum period of time T′_Max must be longer than two time intervals dT of the execution grid. In addition, process execution times that are too short, likewise pointing to a malfunction of the real-time system or of the processes, are not always able to be detected. In FIG. 9, both process P₆₆, as well as process P₆₉ are atypically short. If the too short process is the second process of a pair P₆₈, P₆₉, the counter Z′ records a count C₆₈, which is shorter than predefined by the minimum period of time T′_min, and, as a consequence thereof, an underflow error signal F′₂ is output. However, if the too short process P₆₆ is the first of a pair P₆₆, P₆₇, then, due to the start of the second process in accordance with the time-slot pattern at point in time t₆₇, no count C₆₆ is recorded that is shorter than the corresponding minimal period of time T′_min.

SUMMARY OF THE INVENTION

A method is devised by the exemplary embodiment and/or exemplary method of the present invention which has the features described herein and is able to be implemented by the device having the features described herein. The method according to the present invention provides for every underflow and overflow to be recorded, a reaction time, measured from a start of a faulty process to the point of outputting of an error signal, being short.

The method according to the present invention for monitoring a process execution of a plurality of sequentially executed processes starts one of a plurality of timers in cyclic permutation when one of the processes is started, and outputs a first error signal when a period of time recorded by one of the timers exceeds a predefined maximum period of time.

One advantage of the exemplary embodiment and/or exemplary method of the present invention is that a malfunction of a process, i.e., an atypically long execution duration, causes a first timer, which had been started prior to the process by a preceding process, to overflow. For that reason, measured from the start of a process, the reaction time until the malfunction of the process is recognized, is shorter than the maximum period of time.

The option of starting a timer by a process and stopping it by the same process requires selecting a longer average maximum period of time than in the method according to the present invention.

This has to do with the typical outliers of the execution duration which occur in isolated instances and should not result in an outputting of the error message. For that reason, in the case that the timer is started and stopped by the same process or is restarted by the immediately following process, the maximum period of time must be greater than the execution duration of the typical outliers. In the method according to the present invention, it is assumed that N−1 processes are executed in the typical manner and, at most, that one single process is an outlier, N being the number of timers. Thus, the excess length of the outlier is distributed over N processes, and the average maximum period of time, in terms of one single process, is shorter than in the aforementioned method that is not in accordance with the exemplary embodiment and/or exemplary method of the present invention.

Advantageous further refinements of the method set forth in claim 1 and improvements thereto are indicated in the dependent claims.

One exemplary embodiment provides that, in response to the ending of a process, a recorded period of time of the timer be read out which is restarted by the next process to be executed, and that a second error signal be output when the period of time falls below a predefined minimal period of time. This method advantageously makes it possible for a malfunction of the process execution having an atypically short execution duration of a process to be recognized.

Another embodiment provides for a test datum to be transmitted to a started process for processing purposes and, upon termination of the process, for a third error signal to be output in response to a deviation of the processed test datum from a setpoint value. In one embodiment, a process is composed of subprocesses, and at least one subprocess executes an operation on the test datum which confirms the execution of this subprocess reproducibly by the processed test datum. This makes it possible to recognize whether all relevant subprocesses have been executed in which the processed test datum is compared to the setpoint value. Another embodiment provides for at least two subprocesses to execute mutually non-commutative operations on the datum, with the result that the execution of these subprocesses is able to be confirmed in the correct sequence by the processed test datum. It is a feature of mutually non-commutative operations that an operation in a different sequence on the test datum leads, for most test data, to different processed test data.

In accordance with one exemplary embodiment, two immediately successive processes are started at a minimal time interval. This makes it possible for processes to be regularly executed.

Another embodiment provides for the maximum period of time to be greater than a first product of the minimal time interval and for the number of timers to be smaller than the sum of the first product and of the minimal time interval. Another embodiment provides for the minimal period of time to be greater than or equal to a second product of the minimal time interval and of the number of timers reduced by one.

In accordance with one specific embodiment of the present invention, a device for implementing the method includes a monitoring device which is connected to a processor device that executes the processes in order to receive a start signal from the control device when a process is started, and is configured to compare a transmitted period of time to the maximal period of time and, in response to exceeding of the same, to output an error signal via an interface; and includes at least two timers which are connected to a multiplexer of the monitoring device in order to be able to be cyclically restarted by a reset signal and are connected to an input of the monitoring device in order to transmit a recorded period of time to the monitoring device.

Another embodiment provides for the timers to have regressively operated counters which, in response to a restart, are set to a value that corresponds to the predefined maximal period of time. Another embodiment provides for the timer to be configured to enable the recorded period of time to be read out by the monitoring device and for the timer to be configured to be stoppable by a control signal.

Yet another embodiment provides for the monitoring device to have a comparator which compares the setpoint value to a test datum processed by the process and, in the case of a deviation, outputs a third error signal.

Exemplary embodiments of the present invention, as well as advantageous further refinements, are shown in the figures of the drawing and are explained in greater detail in the following description.

Although the exemplary embodiment and/or exemplary method of the present invention is described with reference to a monitoring of a process control by a real-time system, it is not limited thereto, but rather relates generally to methods for monitoring processes.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a schematic representation of a specific embodiment, as a flow diagram.

FIG. 2 shows a schematic representation of a specific embodiment, as a flow diagram.

FIG. 3 shows a schematic representation of the specific embodiment for clarifying an overflow.

FIG. 4 shows a schematic representation of the specific embodiment in the case of an underflow.

FIG. 5 shows a flow diagram of a specific embodiment having a multi-step process.

FIG. 6 shows a schematic block diagram of a specific embodiment of the device according to the exemplary embodiment and/or exemplary method of the present invention.

FIG. 7 shows a flow diagram for clarifying a problem definition underlying a specific embodiment of the exemplary embodiment and/or exemplary method of the present invention.

FIG. 8 shows a flow diagram for clarifying the problem definition underlying the method according to the present invention.

FIG. 9 shows another flow diagram for clarifying the problem definition underlying the method according to the present invention.

DETAILED DESCRIPTION

In the figures, like or functionally equivalent components are denoted by like reference numerals, provided that nothing is indicated to the contrary.

FIG. 1 shows the execution in sequential order of a plurality of individual processes P₁, P₂, . . . by a processor system S. To the greatest degree possible, processor system S attempts to start each of the processes at predefined points in time t₁, t₂, . . . , provided that a preceding process P₃, P₄ is not still active at predefined points in time t₄, t₅. In such a case, processor system S starts process P₄, P₅ instantaneously at point in time T₃, T₄ at which process P₃, P₄ is ended. The described processor system and the execution of the processes are typical for a real-time operating system and a corresponding device, as is frequently used in control systems. In most cases, time interval dT₁, dT₂ between two successive points in time t₁, t₂, . . . is equidistant. In another specific embodiment, time intervals dT₁, dT₂, as illustrated in FIG. 1, may differ and form pairs or longer sequences that are periodically repeated. With regard to the following descriptions, it is assumed that time intervals dT₁, dT₂ are equidistant, unless indicated otherwise.

Various malfunctions may occur during the process execution by processor system S. For that reason, it is desirable to obtain indicators of a malfunction of processor system S. A very simple and nevertheless effective method provides for monitoring the execution duration of processes P₁, P₂, . . . . To that end, FIG. 1 shows a specific embodiment of the present invention that is used for monitoring two timers Z_A, Z_B. The two timers Z_A, Z_B are alternatingly restarted when a new process P₁, P₂, . . . is started nearly simultaneously by processor system S. Thus, each timer Z_A, Z_B is only restarted by every second process. In this connection, restarting is understood to mean that the timer is reset to a start value and begins to measure again. First timer Z_A is started, together with first process P₁, at predefined point in time t₁. First timer Z_A is restarted when third process P₃ is started at predefined point in time t₃. Second timer Z_B is not influenced by the start of first process P₁ and of third process P₃, but rather is restarted together with second process P₂ started in between the aforementioned processes.

The period of time between two restarts of one of timers Z_A, Z_B is read out prior to a restarting of corresponding timer Z_A. In the specific embodiment shown in FIG. 1, a timer Z_A is stopped when second process P₂ is ended at point in time T₂. Thus, period of time A₁ recorded in this manner by timer Z_A corresponds to the period of time between the start of first process P₁ and the end of second process P₂. In the illustrated example, second process P₂ is ended prior to third point in time t₃ at which third process P₃ is started, so that the reading out of first timer Z_A at point in time T₂ does not coincide with the restarting of first timer Z_A at predefined point in time t₃.

Another specific embodiment provides for only stopping timer Z_A upon termination of a process and for reading out this timer before it is restarted. In the case of fourth process P₄, the point in time of reading out or of stopping of first timer Z_A must necessarily coincide with the restarting of first timer Z_A, since fifth process P₅ is immediately started upon termination of fourth process P₄. The performance characteristics of second timer Z_B correspond to that of first timer Z_A, with the exception that the process of reading out and restarting of second timer Z_B is delayed by the execution time of a process.

This pattern may be applied universally to N timers Z_A, Z_B, . . . Z_N, these timers then being started one after another in a cyclically permuted sequence, thus, following N-th timer P_N, first timer P₁ is restarted again. Thus, first timer Z_A is restarted in response to processes P₁, P_N+1, P_2N+1, . . . ; second timer Z_B in response to processes P₂, P_N+2, P_2N+2, . . . ; and k-th timer Z_N in response to processes P_k, P_N+k, P_2N+k. The reading out or stopping of a k-th timer Z_k is carried out in accordance with the specific embodiment of FIG. 1 upon termination of process P_k−1, P_N+k−1, P_2N+k−1, which had been started prior to process P_k, P_N+k, P_2N+k, in response to which k-th timer Z_k is restarted.

In FIG. 1, processes P₃ are shown whose execution duration is longer than time interval dT₁, that is made available for the processes. Time interval dT₁, dT₂ is exceeded in isolated instances when a process P₃ is required to carry out a greater number of calculations or individual steps than is otherwise customary. This may occur in an airbag control system, for example, which normally queries all sensors and, on the basis of a few calculations, is able to comprehend that a situation does not exist that would require airbag activation. In isolated instances, however, the sensors signal increased acceleration values, indicating the occurrence of an accident, and, accordingly, processor system S must perform a greater number of calculations in order to verify these signals and/or to activate the airbag. Since these situations and thus the long execution durations of the processes are very infrequent, time intervals dT₁, dT₂ are selected to enable processes P₁, P₂, . . . having a normal execution duration to be executed within these time intervals dT₁, dT₂, and, at the same time, for individual outliers to be compensated over a longer period of time or a plurality of processes, so that subsequent processes are again executed in accordance with the time schedule, as illustrated by the example of third process P₃ having excess length and by the subsequent sixth process P₆, which is executed in conformance with the time schedule at point in time t₆.

It follows from the preceding description that not every instance of a process execution exceeding a time interval dT₁, dT₂ is based on a malfunction of processor system S or of executed process P₃, but rather may also correspond to infrequently occurring outliers of the execution duration. For this purpose, a time buffer ΔT is introduced which corresponds to the expected excess length of the outliers. Since it is to be assumed that a typical outlier is very rare and thus only occurs once in 2 or N successive processes P₁, P₂, . . . , a maximum period of time T_Max is defined that is longer by time buffer ΔT than 2−times or N−times time interval dT. Generally, maximum period of time T_Max is longer by the time buffer than the sum calculated over all time intervals dT₁, dT₂ between two scheduled starts of a timer.

It is explained exemplarily with reference to FIG. 2 how, on the basis of a predefined maximum period of time T_Max and the periods of time recorded by timers Z_A, Z_B, a malfunction is recognized in the execution of processes. Eighth and ninth processes P₈, P₉ are executed without any malfunction being experienced by processor device S. Analogously to the explanations regarding FIG. 1, in response to the start of ninth process P₉, second timer Z_B is started, and period of time B₉ is recorded by second timer Z₈. In conformance with the time schedule, tenth process P₁₀ is started at point in time t₁₀, a malfunction occurring during execution thereof, for example an endless loop within tenth process P₁₀ or a hardware defect of processor system S, which leads to an atypically long execution time of tenth processor P₁₀. In response to tenth process P₁₀ first timer Z_A was also started. Both the first, as well as second timer Z_A, Z_B monitor periods of time A₁₀ or B₉ recorded by the same. The monitoring of timers Z_A, Z_B may also be undertaken by an external unit. If one of recorded periods of time A₁₀ B₉ exceeds predefined maximum period of time T_Max, an error signal is generated. The signal may be generated as closely in time as possible to point in time F₁ of the beginning of exceedance in time of maximum period of time T_Max. If maximum periods of time T_Max for the two timers Z_A, Z_B are selected to not differ too greatly (normally they are the same), recorded period of time B₉ of second timer Z_B first exceeds maximum period of time T_Max, since this second timer Z_B was started already prior to tenth process P₁₀. From this, it follows that reaction time R₁, measured from the beginning of the start of tenth process P₁₀ at which a malfunction occurs, until the point when the malfunction is recognized at point in time F₁, is shorter by one or N−1 times the time interval dT than maximum period of time T_Max. Thus, resultant reaction time R₁ is the sum of a time interval dT and of time buffer ΔT. This makes it possible for a malfunction of processor system S or of the executed process to be quickly reacted to. In this case, in the context of an airbag control system, one specific embodiment of the present invention provides for an output stage to be blocked, so that a potentially erroneously generated control signal of the airbag control device is not able to activate an airbag.

In a comparison with FIG. 7, it becomes apparent that the method according to the present invention reacts just as quickly to a malfunction of a process as when a monitoring system having one single timer Z″ is used that is restarted in response to each process P₅₀, P₅₁, . . . . During process P₅₂, a malfunction occurs in processor system S″, which, in response to the exceedance of a maximum period of time T″_Max, is ascertained at point in time F″₁ due to the maximum period of time T″_Max corresponding to the sum of time interval dT and of time buffer ΔT. Thus, reaction time R₁ of the specific embodiment described in the preceding and of this system are equal.

Another indication of a malfunction in a processor system S is when processes are ended much more quickly than usual. It is described with reference to FIG. 3 how such a malfunction is ascertained with the aid of a minimal period of time T_min. Thirteenth and fourteenth processes P₁₃, P₁₄ are executed and terminated in accordance with the time schedule. Second timer Z_B is started together with fourteenth process P₁₄. Fifteenth process P₁₅ is started together with first timer Z_A. During fifteenth process P₁₅, a malfunction arises, which, for example, prematurely ends the execution of fifteenth process P₁₅. At the end of this fifteenth process, second timer Z_B is stopped in the usual manner, and period of time B₁₄ recorded by the same is read out. This period of time B₁₄ is compared to minimal period of time T_min. If recorded period of time B₁₄ falls below minimum period of time T_min, a second error signal is output. The error signal may be output at point in time F₂ as closely in time as possible to the reading out of second timer Z_B. Minimum period of time T_min is shorter by a time difference δT than twice time interval dT. All previous considerations regarding time buffer ΔT for maximum period of time T_Max are applicable to minimum period of time T_min, with the difference that minimum period of time T_min is shorter by time difference δT than twice or N−times time interval dT.

Besides monitoring the processes on the basis of their execution times, another possible diagnostic method is provided by the specific embodiment illustrated with reference to FIG. 4.

If, in a processor system S, a first process P₁ is started at 200, processor control S transmits a start signal 100 to a monitoring device 2. In response thereto, monitoring device 2 starts a first timer Z_A at 250, as described in the preceding exemplary embodiments. In addition, a test datum X₁ is selected at 251 and transmitted via a signal 101 to processor system S. In first process P₁, at least one operation f is applied to test datum X₁ at 201, so that a processed test datum Y₁ is thereby obtained. At termination 202 of first process P₁, a stop signal 102 is transmitted from processor system S to monitoring device 2. Subsequently, in monitoring device 2, second timer Z_B is stopped at 252, in order to check whether period of time B recorded by second timer Z_B falls below minimal period of time T_min at 254; and, if it has, an error message is output at 259, as already described in the preceding specific embodiments. In addition, processed test datum Y₁, which was transmitted via stop signal 102, is compared to a setpoint value Y_setpoint1. In the case of a deviation, an error message is likewise output. In addition, it is checked whether a recorded period of time has not already exceeded maximum period of time T_Max at 258, which likewise leads to the outputting of an error message, as already explained.

At the start of another process, the previously described steps are repeated, as can be inferred from FIG. 4, the other timer being started or stopped accordingly. By using various test data X₁, X₂, . . . , it may be ensured that the operations are executed on test data X₁, X₂, . . . and that the process does not only read out one single previously processed test datum Y₁, Y₂, . . . from one memory cell of processor system S. Test data X₁, X₂, . . . may be cyclically used again. For every test datum X₁, X₂, . . . , corresponding setpoint values Y_setpoint1, Y_setpoint2, . . . are to be reserved (held ready) in monitoring device 2.

With reference to FIG. 5, two applications of a test datum are described. Typically, one single process may be subdivided into a plurality of subprocesses. In the context of an airbag control, these would include readout sensors I₁, analyzing sensor values 12, and generating a control signal for airbag system 13, for example. In this connection, it is absolutely necessary for the correct execution of process P₁ that all subprocesses I₁, I₂, I₃ be executed. To this end, every subprocess I₁, I₂, I₃ subsequently executes an operation F on test datum X₁. Three very simple operations F₁, F₂, F₃ would entail negating the first, the second and, respectively, the third bit of test datum X₁. Thus, by making a comparison to the setpoint value, it is possible to directly ascertain whether all three operations were executed or if one, and in this case even which one, of the operations was not executed. Important for the correct execution of process P₁ is also the sequence of executions of subprocesses I₁, I₂, I₃. This may be accomplished in that operations F₁, F₂, F₃ are mutually or at least partially non-commutative. As a result, an incorrect sequence of the execution yields a processed test datum Y that differs from the setpoint value. Assuming, for example, test datum X₁ is five and operation F₁ is multiplication by two and operation F₂ is addition by three, a correct sequence yields the result thirteen, which in this case is the setpoint value, and a permuted sequence yields the result sixteen, which does not correspond to the setpoint value.

With reference to FIG. 6, a monitoring device 2 is described which has an interface and a processing module 1 that is linked to a processor system S via a data and control line 20 and to an output stage 3 of an airbag system 4 in order to transmit a suppression signal 201. Suppression signal 201 is generated by processing device 1 when a malfunction of the process execution is ascertained by processor system S using one of the methods described above. Processing device 1 is capable of restarting and, if indicated, also of stopping a multiplicity of timers Z_A, Z_B, Z_N directly or via a multiplexer device 10. In one exemplary embodiment, timers Z_A, Z_B, Z_N are down counters which count down in time with a timer. If a count of one of the counters reaches value 0, then this triggers an interrupt in a cell readout device 11. When they are restarted, counters Z_A, Z_B, Z_N, are set to a value that corresponds to the maximum period of time. As a result, the triggering of the previously mentioned interrupt is interpreted as an error in processor system S during the process execution. In another variant, counter readout device 11 may read out the count of the counters or of timer Z_A, Z_B, Z_N and, when prompted, transfer this value to processing device 1. Readout device 11 may likewise be linked to a multiplexer which is configured to only allow that timer to be read out which has been running for the longest period of time. This may be implemented by employing a simple cyclical multiplexer.

Control device S outputs a control signal to output stage 3 when airbag system 4 is to be triggered. When an error of the process execution is recognized by processing device 1, output stage 3 is blocked, thereby preventing a triggering of airbag system 4.

Another advantage of the described specific embodiments is achieved in accordance with the following. As described, the outliers require introduction of time buffer ΔT. In the context of a system having N timers Z₁, Z₂, . . . , time buffer ΔT is effectively distributed over N processes P₁, P₂ . . . . Thus, the average execution duration for each individual process is specified more precisely than when working with only one single timer Z″ and, nevertheless, isolated outliers do not lead to the outputting of an error signal.

Claims

1. A method for monitoring a processor system, the method comprising: executing a plurality of processes, one of the processes being started after a preceding one of the processes is ended, wherein one of at least two timers is restarted in cyclic permutation when one of the processes is started; andoutputting a first error signal when a period of time recorded by one of the at least two timers exceeds a predefined maximum period of time;wherein, in response to an ending of a process, a recorded period of time of a corresponding timer, which is restarted by a next process to be executed, is read out, and wherein a second error signal is output when a recorded period of time falls below a predefined minimal period of time.

2. The method of claim 1, wherein one of the processes is started, at the earliest, following a predefined time interval that has elapsed since the start of another of the processes that was started immediately prior to the one of the processes.

3. The method of claim 2, wherein: a length of the maximum period of time is defined by N dT<TMax<(N+1) dT;TMax is the maximum period of time;dT is the predefined time interval;N is a number of the timers;N dT is a first product of the predefined time interval and of the number of timers; and(N+1) dT is a sum of the first product and of the predefined time interval.

4. A method for monitoring a processor system, the method comprising: executing a plurality of processes, one of the processes being started after a preceding one of the processes is ended, wherein one of at least two timers is restarted in cyclic permutation when one of the processes is started; andoutputting a first error signal when a period of time recorded by one of the at least two timers exceeds a predefined maximum period of time;wherein a test datum is transmitted for processing to a started process, and wherein upon termination of the process, a second error signal is output in response to a deviation of the processed test datum from a setpoint value.

5. The method of claim 4, wherein a process includes subprocesses, and at least one of the subprocesses executes an operation on the test datum which confirms an execution of the at least one of the subprocesses reproducibly by the processed test datum.

6. The method of claim 5, wherein at least two subprocesses execute mutually non-commutative operations on the test datum, so that the execution of the at least two subprocesses in a correct sequence is able to be confirmed by the processed test datum.

7. The method of claim 4, wherein one of the processes is started, at the earliest, following a predefined time interval that has elapsed since the start of another of the processes that was started immediately prior to the one of the processes.

8. The method of claim 7, wherein: a length of the maximum period of time is defined by N dT<TMax<(N+1) dT;TMax is the maximum period of time;dT is the predefined time interval;N is a number of the timers;N dT is a first product of the predefined time interval and of the number of timers; and(N+1) dT is a sum of the first product and of the predefined time interval.

9. A method for monitoring a processor system, the method comprising: executing a plurality of processes, one of the processes being started after a preceding one of the processes is ended, wherein one of at least two timers is restarted in cyclic permutation when one of the processes is started; andoutputting a first error signal when a period of time recorded by one of the at least two timers exceeds a predefined maximum period of time;wherein: one of the processes is started, at the earliest, following a predefined time interval that has elapsed since the start of another of the processes that was started immediately prior to the one of the processes;a second error signal is output when a recorded period of time falls below a predefined minimal period of time;the minimal period of time is greater than or equal to (N−1) dT;dT is the predefined time interval; andN is a of the number of the timers.

10. The method of claim 9, wherein: a length of the maximum period of time is defined by N dT<TMax<(N+1) dT;TMax is the maximum period of time;N dT is a first product of the predefined time interval and of the number of timers; and(N+1) dT is a sum of the first product and of the predefined time interval.

11. A device for monitoring a processor system, comprising: a monitoring device, which is linked to the processor system to receive a start signal from the processor system when a process is started;at least two timers connected to a multiplexer of the monitoring device to cyclically restart the timers using a reset signal, and which are connected to an input of the monitoring device to transmit a recorded period of time to the monitoring device;wherein: a plurality of the processes are executed, one of the processes being started after a preceding one of the processes is ended;one of at least two timers is restarted in cyclic permutation when one of the processes is started;a first error signal is outputted when a period of time recorded by one of the at least two timers exceeds a predefined maximum period of time; andthe monitoring device has a comparator to compare a setpoint value to a test datum processed by the process, and to output a second error signal for a deviation.

12. The device of claim 11, wherein the timers are regressively operated counters which, in a restart, are set to a value that corresponds to the predefined maximal period of time.

13. The device of claim 12, wherein the input is configured to trigger an output of the first error message when a zero of one of the counters is present at the input.

14. The device of claim 11, wherein the count of the counters is able to be read out by the monitoring device, and the timer is configured to be stoppable by a control signal.

15. The device of claim 11, wherein a process includes subprocesses, and at least one of the subprocesses executes an operation on the test datum which confirms an execution of the at least one of the subprocesses reproducibly by the processed test datum.

16. The device of claim 15, wherein at least two subprocesses execute mutually non-commutative operations on the test datum, so that the execution of the at least two subprocesses in a correct sequence is able to be confirmed by the processed test datum.

17. The device of claim 11, wherein: one of the processes is started, at the earliest, following a predefined time interval that has elapsed since the start of another of the processes that was started immediately prior to the one of the processes;a length of the maximum period of time is defined by N dT<TMax<(N+1) dT;TMax is the maximum period of time;dT is the predefined time interval;N is a number of the timers;N dT is a first product of the predefined time interval and of the number of timers; and(N+1) dT is a sum of the first product and of the predefined time interval.

18. A device for monitoring a processor system, comprising: a monitoring device, which is linked to the processor system to receive a start signal from the processor system when a process is started;at least two timers connected to a multiplexer of the monitoring device to cyclically restart the timers using a reset signal, and which are connected to an input of the monitoring device to transmit a recorded period of time to the monitoring device;wherein: a plurality of the processes are executed, one of the processes being started after a preceding one of the processes is ended;one of at least two timers is restarted in cyclic permutation when one of the processes is started;a first error signal is outputted when a period of time recorded by one of the at least two timers exceeds a predefined maximum period of time;in response to an ending of a process, a recorded period of time of a corresponding timer, which is restarted by a next process to be executed, is read out; anda second error signal is output when a recorded period of time falls below a predefined minimal period of time.

19. A device for monitoring a processor system, comprising: a monitoring device, which is linked to the processor system to receive a start signal from the processor system when a process is started;at least two timers connected to a multiplexer of the monitoring device to cyclically restart the timers using a reset signal, and which are connected to an input of the monitoring device to transmit a recorded period of time to the monitoring device;wherein: a plurality of the processes are executed, one of the processes being started after a preceding one of the processes is ended;one of at least two timers is restarted in cyclic permutation when one of the processes is started;a first error signal is outputted when a period of time recorded by one of the at least two timers exceeds a predefined maximum period of time;one of the processes is started, at the earliest, following a predefined time interval that has elapsed since the start of another of the processes that was started immediately prior to the one of the processes;a second error signal is output when a recorded period of time falls below a predefined minimal period of time;the minimal period of time is greater than or equal to (N−1) dT;dT is the predefined time interval; andN is a of the number of the timers.

20. A method for monitoring a processor system, the method comprising: sequentially executing a plurality of processes, a predefined time interval dT being assigned to each of the plurality of processes such that, for each respective one of the processes, a process to be executed following the respective process is not started at least until lapse of the predefined time interval measured from a start of execution of the respective process;restarting a first of a plurality of timers upon a respective start of each of a first subset of the sequentially executed plurality of processes and not upon a respective start of any of a second subset of the sequentially executed plurality of processes;restarting a second of the plurality of timers upon the respective start of each of the second subset of the sequentially executed processes and not upon the respective start of any of the first of the sequentially executed plurality of processes; andoutputting an error signal when a period of time recorded by one of the first and second timers exceeds a predefined maximum period of time;wherein: for each of the plurality of timers, after the respective timer is restarted upon the respective start of one of the respective subset of the sequentially executed plurality of processes upon start of which the respective timer is configured to be restarted, where an error does not occur, the respective timer is restarted only after successive execution of a number of the sequentially executed plurality of processes at least equal to a number N of the plurality of timers configured to be restarted upon each respective start of a respective subset of the plurality of processes; anda length of the maximum period of time TMax is defined by (N*dT)<TMax<((N+1)*dT).

21. A method for monitoring a processor system, the method comprising: sequentially executing a plurality of processes, a predefined time interval dT being assigned to each of the plurality of processes such that, for each respective one of the processes, a process to be executed following the respective process is not started at least until lapse of the predefined time interval measured from a start of execution of the respective process;restarting a first of a plurality of timers upon a respective start of each of a first subset of the sequentially executed plurality of processes and not upon a respective start of any of a second subset of the sequentially executed plurality of processes;restarting a second of the plurality of timers upon the respective start of each of the second subset of the sequentially executed processes and not upon the respective start of any of the first of the sequentially executed plurality of processes; andoutputting an error signal when a period of time recorded by one of the first and second timers is less than a predefined minimal period of time;wherein: for each of the plurality of timers, after the respective timer is restarted upon the respective start of one of the respective subset of the sequentially executed plurality of processes upon start of which the respective timer is configured to be restarted, where an error does not occur, the respective timer is resstarted only after successive execution of a number of the sequentially executed plurality of processes at least equal to a number N of the plurality of timers configured to be restarted upon each respective start of a respective subset of the plurality of processes; anda length of the minimal period of time TMin is defined by ((N−1)*dT)≧TMin .