DE102019209484A1 - Room surveillance method and room surveillance system for monitoring a traffic area - Google Patents

Abstract

The invention relates to a space monitoring method for monitoring a traffic area (1), - in which an operating situation (st) of the traffic area (1) is recognized as a predetermined situation (sT) of the traffic area (1) if a based on first data (DS1t , DS2t, DS3t) the operating situation result (st *) determined for the operating situation (st) corresponds to a specified situation result (sT *) determined for the specified situation (sT), - where the specified situation result (sT *) or a part of the specified situation result ( sT *) is a combination of a given state result (partly *) and a given object result (oT *). In order to improve the reliability of the detection of objects in the traffic area and the disclosure of errors, it is provided that - with regard to the operating situation (st) of the traffic area (1) based on the first data (DS1t, DS2t, DS3t) a first operating status result (zt * .1) and based on the second Da th (DP1t, DP2t, DP3t, DP4t) a second operating status result (zt * .2) can be determined and the two operating status results (zt * .1, zt * .2) are checked for correspondence. The invention also relates to a room surveillance system Computer program and a provision device for the computer program.

Description

It is known that level crossings, in particular level crossings with full barriers, or platforms are monitored by staff or by means of expensive radar scanners, which require a SIL3 safety approval. In the case of simpler operating conditions for level crossings, other solutions, for example time-controlled closing after a warning (known as a so-called "call barrier"), are also permissible.

Traffic areas such as level crossings and platforms have in common that their situations are characterized in particular by existing objects and prevailing conditions (appearances).

The invention relates to a space monitoring method for monitoring a traffic area, in which an operating situation of the traffic area is recognized as corresponding to a given situation of the traffic area when an operating situation result determined on the basis of first data on the operating situation corresponds to a given situation result determined for the given situation Situation result or part of the predefined situation result is a combination of a predefined status result and a predefined object result.

Such a room monitoring method is from the document DE 103 41 426 A1 known. In this case, it is provided that the type of the current situation - that is, the operating situation - is detected via a classification. In particular, all things occurring in the traffic area, such as, for example, every train, every person, animal, pieces of luggage and also other objects, if they are moved or move for any reason, are considered objects. Appearances - i.e. states - should be able to be related to the processes taking place in the traffic area. They should include light and weather conditions, for example.

Furthermore, from the publication DE 10 2017 126 553 A1 a method for operating a detection system is known, the detection system serving to classify an object. In this known method, at least one value is transmitted from an external device to the detection system, the at least one value being, for example, a current day of the week or a current season or a current time or current weather data.

Based on a generic known method ( DE 103 41 426 A1 ) The invention is based on the object of improving the reliability of the detection of objects in the traffic area and the disclosure of errors which can occur, for example, as a result of failure of a room surveillance system used for carrying out the method.

This object is achieved with the method according to the invention according to claim 1, in which a first operating state result is determined based on the first data and a second operating state result based on second data and the two operating state results are checked for correspondence with regard to the operating situation of the traffic area.

Advantageous embodiments of the method according to the invention are the subject of the dependent claims 2 to 7th .

With regard to the operating situation of the traffic area, an object result is preferably determined using the first data, which in combination with the first operating state result forms the operating situation result or a part of the operating situation result.

It is considered to be particularly advantageous if an error message is output if the two operating state results do not match.

It is also considered advantageous if a further error message is output if, on the one hand, the first operating status result corresponds to a predetermined operating mode of the traffic area, if at least one selected object is to be expected in the traffic area, and if, on the other hand, in a predetermined time interval after the determination the first operating state result does not determine a predetermined object result which corresponds to the expected object.

It is also considered advantageous if a monitoring message is output when the two operating status results match, when the operating situation result determined for the operating situation also corresponds to the predefined situation result determined for the predefined situation and when the predefined situation result is predefined as a hazard.

It is advantageous if a computer system of a room surveillance system, in particular a first computer of the computer system, receives the first data from a sensor device and the first operating status result and the object result are determined using a machine learning algorithm.

It is also advantageous if the computer system, in particular a second computer of the computer system, receives the second data from a status parameter data provision device and determines the second operating status result by means of a status model algorithm or by means of a further machine learning algorithm.

The above object is achieved according to a second aspect by a room surveillance system according to claim 8.

Such a room surveillance system is set up to carry out the method according to claim 1. This is a room surveillance system for monitoring a traffic area, with a computer system,
which is set up to recognize an operating situation of the traffic area as corresponding to a predetermined situation of the traffic area when an operating situation result determined on the basis of first data on the operating situation corresponds to a predetermined situation result determined for the predetermined situation, the predetermined situation result or a part of the predetermined situation result being a A combination of a predefined status result and a predefined object result, and which is also set up to determine a first operating status result based on the first data and a second operating status result based on second data and to check the two operating status results for correspondence with regard to the operating situation of the traffic area.

The computer system is set up in particular to determine an object result with regard to the operating situation of the traffic area using the first data, which in combination with the first operating state result forms the operating situation result or part of the operating situation result.

The computer system is preferably set up to output an error message if the two operating status results do not match.

In addition, the computer system is preferably set up to output a further error message if, on the one hand, the first operating state result corresponds to a predetermined operating mode of the traffic area, if at least one selected object is to be expected in the traffic area, and if, on the other hand, in a predetermined time interval after the first one has been determined Operating state result no predetermined object result is determined which corresponds to the expected object.

In addition, the computer system is preferably set up to output a monitoring message if the two operating status results match and if the operating situation result determined for the operating situation also corresponds to the predetermined situation result determined for the predetermined situation.

It can advantageously be provided that the computer system, in particular a first computer of the computer system, is set up to receive the first data from a sensor device and to determine the first operating state result and the object result using a machine learning algorithm.

It can also advantageously be provided that the computer system, in particular a second computer of the computer system, is set up to receive the second data from a state parameter data provision device and the second operating state result by means of a state model algorithm or by means of a further machine learning algorithm to investigate.

The invention also relates to a computer program with program instructions which, when the computer program is executed by a computer system, cause the computer system to carry out the steps of the method according to one of claims 1 to 7, as well as a provision device for this computer program which stores and / or provides the computer program.

• 1 ein Ausführungsbeispiel der erfindungsgemäßen Raumüberwachungsanlage während einer Trainingsphase und die
• 2 das eine Ausführungsbeispiel der erfindungsgemäßen Raumüberwachungsanlage während einer Betriebsphase.

The invention is further illustrated by the 1 and 2 explained in more detail. The

• 1 an embodiment of the room surveillance system according to the invention during a training phase and the
• 2 one embodiment of the room surveillance system according to the invention during an operating phase.

The 1 shows a traffic area 1 as well as a room surveillance system according to the invention 2 showing the traffic area 1 supervised.

The traffic area 1 is a fixed area, the situation of which is characterized by a prevailing state and the presence of one or more objects.

The room surveillance system 2 has a sensor device 3 , a state parameter data providing device 4th and a computer system 5 on.

The sensor device 3 comprises at least one sensor unit. In the exemplary embodiment shown here, the sensor unit comprises, for example, three sensor units S1 , S2 and S3 on.

The state parameter data provision device 4th includes at least one unit. In the exemplary embodiment shown here, the state parameter data provision device comprises 4th for example four units P1 , P2 , P3 and P4 on.

The computer system 5 has a first computer here as an example 6th and a second computer 7th on that via a communication path K are connected.

Also, the computer system 5 a provision device, not shown here, which stores a computer program. The computer program is a computer program with program instructions that are used when the computer program is executed by the computer system 5 cause this to carry out the steps of the method according to the invention described in detail below.

The computer program has two program parts. A first of the program parts is a machine learning algorithm MLA which by the first computer 6th is performed. The second part of the program is a state model algorithm ZMA .

The sensor units S1 , S2 and S3 are established via suitable mechanisms of action W1 , W2 and W3 Measurements regarding the traffic area 1 and the results of their measurements in the form of initial data (sensor data) DS1 _T , DS2 _T , DS3 _T via suitable communication paths K11 , K12 , K13 to the first computer 6th of the computer system. Since the 1 is a training phase, is in the 1 the index "T" is used.

In the traffic area 1 different objects can be located, some of which are particularly relevant for the operation of the traffic area and whose presence in the traffic area must be monitored. The recognition of these objects is learned in a training phase, as will be described below.

Real can be in the traffic area 1 a variety of different objects are located.

As shown here, a first object o _{1 is,} for example, a train. A second object o ₂ is, for example, a vehicle. A third object o ₃ is, for example, a person and an m-th object o _m is, for example, a tree whose exterior can change due to growth and weather conditions.

The set of these objects is here with O _T = {o _T | o _T = 01 or O _T = O ₂ ... or O _T = O _m } denotes and at the same time forms a set of predetermined objects or a set of training objects.

The sensor units S1 , S2 and S3 are commercially available sensor units. So is the sensor units S1 a radar sensor unit. The sensor units S2 is a lidar sensor unit and sensor units S3 a camera sensor unit.

The units P1 , P2 , P3 and P4 are set up via suitable additional communication paths K21 , K22 , K23 , K24 to the second computer 7th of the computer system 5 output second data (status parameter data).

This one with P1 The designated first unit provides exemplary parameter data on the current weather in the traffic area 1 ready (weather, e.g. provided by a weather service). With its parameter data it indicates, for example, whether it is sunny or whether it is overcast, whether it is raining or whether it is snowing.

This one with P2 The designated second unit provides exemplary parameter data for the current operating mode of the traffic area 1 ready. Is the traffic area 1 For example - as shown here - a restricted level crossing, then the second unit with its parameter data indicates, for example, whether the level crossing is currently secured (cf. BA in 2 ), i.e. whether the barriers are closed, or whether it is currently not secured, i.e. whether the barriers are currently open.

This one with P3 The third unit designated provides exemplary parameter data for the current day segment. With its parameter data, it indicates, for example, whether it is day (daytime) or night (nighttime).

This one with P4 The fourth unit designated provides exemplary parameter data on the current lighting conditions (for example provided by a brightness sensor) in the traffic area 1 ready. With its parameter data, it indicates, for example, whether it is currently light or dark.

Regarding the parameters defined by the units P1 to P4 recorded and then provided as the parameter data, the traffic area can 1 so n different states are located. The set of these states is here with Z _T = {Z _T | Z _T = z ₁ or z _T = z ₂ ... or z _T = z _n denotes and at the same time forms a set of predetermined states or a set of training states.

A first state is for example: z _T = z ₁ = (sunny, secured, daytime, bright).

A second state is, for example: z _T = z ₂ = (covered, secured, daytime, bright). For example, an n-th state is: z _T = z _n = (it is snowing, unsecured, at night, dark).

The second computer 7th of the computer system 5 receives the second data DP1 _T , DP2 _T , DP3 _T from the state parameter data provision device 4th and determined to the amount Z _T = {Z _T | Z _T = z ₁ or z _T = z ₂ ... or z _T = z _n a lot of given state results Z _T * = {z _T * | z _T * = z ₁ * or z _T * = z ₂ * ... or z _T * = z _n *} using the state model algorithm ZMA .

The second computer (for example a state transition machine) thus generates with the set Z _T * = {z _T * | z _T * = z ₁ * or z _T * = z ₂ * ... or z _T * = z _n *} Assessments of the surveillance space, which is the finite number of states Z _T = {Z _T | Z _T = z ₁ or z _T = z ₂ ... or z _T = z _n can take.

The first computer 6th of the computer system 5 receives the first data DS1 _T , DS2 _T , DS3 _T from the sensor device 3 and the second data DP1 _T , DP2 _T , DP3 _T from the state parameter data provision device 4th .

From the crowd O _T of given objects (training objects) and the amount Z _T from given states (training states) there is a set of given situations (training situations), which are included here S _T = {S _T | S _T = S ₁ or S _T = S ₂ ... or S _T = z _x } is designated. Some of these given situations are dangerous. For example, a given situation in which a person or a vehicle is between the barriers when the barriers are closed is dangerous. The situational outcome on this situation is therefore appropriately considered a hazard G specified or categorized (cf. 2 ).

The machine learning algorithm MLA is trained in the training phase until it reliably recognizes each of the set of given situations (training situations), i.e. for each given situation (for each of the various combinations ( O _T , S _T ) from one of the training objects and one of the training states) the respectively appropriate, i.e. correct, predetermined situation results are determined.

The machine learning algorithm MLA is therefore for every given state Z _T trained separately until the required reliability is achieved on the training data, with the status associated with the respective training data being included in the assessment of the machine learning algorithm MLA is recorded; As a result of the training phase, the ML algorithm is therefore able to generate a given situation as a combination of a given object O _T and a predetermined state Z _T to recognize.

In addition, the second computer recognizes whether the recognized situation is a danger. So, for example, whether the recognized object in combination with the recognized state is an obstacle H represents.

The 2 shows the traffic area 1 with the room surveillance system 2 in an operating phase. The operating phase is indicated by the index "t".

In order to be able to prove the reliability of the recognition of the given situations in the operating phase, the room surveillance system 2 the method steps described below are carried out.

During the operational phase of the traffic area 1 is initially with regard to an operating situation in the operating phase s _t of the traffic area 1 based on the first data DS1 _t , DS2 _t , DS3 _t a first operating state result z _t *. 1 and based on the second data DP1 _t , DP2 _t , DP3 _t , DP4 _t a second operating status result z _t * .2 determined. In addition, the first data DS1 _t , DS2 _t , DS3 _t an object result Z _t * is determined, which in combination with the first operating state result z _t * .1 an operating situation result s _t * forms. That it is in the present operating situation s _t A preferably current operating situation is indicated by the index “t” behind the “s”.

As shown here is for example s _t = s ₁ and o _t = 0 ₁ . Accordingly, both z _t * .1 = z ₁ * as well as z _t * .2 = z ₁ * be and it should o _t * = o ₁ * provided that no errors occur. The current situation is therefore the current status: (sunny, secure, daytime, light) in combination with the current object: train.

The two operating status results z _t * .1 , z _t * .2 are now checked for consistency.

An FM1 error message is output if the two operating status results z _t * .1 , z _t * .2 do not match, if for example not based on the first data z _t * .1 = z ₁ * but for example z _t * .1 = z ₂ * and based on the second data z _t * .2 = z ₁ * is determined.

Another error message appears FM2 output when on the one hand the first operating status result z _t * .1 a predetermined operating mode BA of the traffic area 1 corresponds, when they are present, at least one selected object o ₁ in the traffic area 1 is to be expected, and if, on the other hand, in a predetermined time interval after the determination of the first operating state result z _t * .1 no predetermined object result o ₁ * is determined which corresponds to the expected object o ₁ . If the level crossing is secured, for example - indicated here by the reference symbol BA , then a train o _{1 must cross} the level crossing in the specified time interval.

Application-specific information - for example, that certain objects must be recognized in certain operating modes, which are harmless in this situation and almost always occur - can be used to reveal the failure of the method. Like the case described, where after the level crossing has been secured, a train usually crosses the level crossing. This train must be recognized and correctly classified by the machine learning algorithm. If this is not the case, there is an operational exception or an error in the railway safety system. In any case, the error message FM2 and a review of the railway safety system. In the same way, objects, cars, pedestrians, etc. should generally be recognized at non-secured level crossings, i.e. when the barriers are open. In known low-traffic times, it can be advantageous to deactivate the monitoring in order to avoid false alarms.

A monitoring message UEM is output when the two operating status results z _t * .1 , z _t * .2 coincide if moreover that to the operational situation s _t determined operating situation result s _t * to the given situation s _T determined predetermined situation result s _T * and if the given situation result ( s _T * ) as danger ( G ) is specified. So if the car o ₂ is between the barriers, this is dangerous. The situation result is as s _T = s _n +1 so is as a danger G given. Of course, a plurality of further predefined situation results are also predefined as a hazard, but this is not shown further here.

The given situation result s _T * is a combination of the given status result z _T * and the given object result o _T * .

This is shown here because z _t * .1 = z _t * .2 = Z ₁ * and ot * = ο ₁ *.

The monitoring message UEM is output when it is recognized that the operating situation s _t of the traffic area 1 the given situations s _T of the traffic area. Here as shown

The first computer 6th of the computer system 5 receives the first data for this purpose DS1 _t , DS2 _t , DS3t from the sensor device 3 and determines the first operating status result z _t * .1 and the object result o _t * using the machine learning algorithm MLA .

The second computer 7th of the computer system 5 receives the second data DP1 _t , DP2 _t , DP3 _t from a state parameter data provision device 4th and determines the second operating status result z _t * .2 using the state model algorithm ZMA .

Alternatively it can be provided that the second computer the 7th Training data DP1 _T , DP2 _T , DP3 _T receives and the second operating state result by means of a further machine learning algorithm, for example a so-called Random Forests algorithm, determined.

Thus, an algorithm based on a neural network could be used as the machine learning algorithm (main algorithm). And as the alternative further machine learning algorithm (secondary algorithm), an algorithm based on decision trees (so-called random forests) could be used.

In operation, you get two independent assessments of the true condition z _T : Using the state model algorithm, the state parameter data (environment information) ZMA made an initial assessment, as a result of which z _t * .1 is determined. And using the machine learning algorithm MLA a second assessment is made, as a result of which z _t * .2 is determined.

If z _t * .1 and z _t * .2 do not match, an error response is initiated in the form of error message FM1. If they match, it can be evaluated and the result is more reliable than the machine learning algorithm alone.

The method according to the invention thus ensures traffic area or danger area monitoring with state-based pattern recognition.

Commercial sensors for detecting the presence of hazardous areas (GFM) at the level crossing or the platform in a train station can be used, especially in simple operating conditions.

A solution is outlined which - apart from commercial sensors and monitoring mechanisms - does not require additional security technology and which still offers reliable detection of obstacles and dangers.

It is known that very reliable results can be achieved with commercial sensors such as cameras, lidar, radar ... in combination with pattern recognition, but this can be difficult to prove if, for example, algorithms such as the machine learning algorithm mentioned are used as pattern recognition MLA , which can be based on a neural network, for example.

With the method according to the invention, the results of the machine learning algorithm MLA supported in an advantageous manner. So there are demonstrable results.

The method according to the invention and the room surveillance system according to the invention 2 thus guarantee a simple support of the machine learning algorithm by environmental information as well as an improved failure disclosure. The main advantage lies in the increased reliability of the decisions as to whether there is an error or whether and what danger there is.

List of reference symbols

1

Traffic area

2

Room surveillance system

3

Sensor device

S1, S2, S3

Sensor units of the sensor device

W1, W2, W3

Sensor mechanisms

DS1 _T , DS2 _T , DS3 _T

Amounts of first data (amounts of first sensor data)

K11, K12, K13

first communication paths

4th

State parameter data provision device

P1, P2, P3, P4

Units of the state parameter data provision device

DP1 _T , DP2 _T , DP3 _T , DP4 _T

Sets of second data (sets of parameter data)

K21, K22, K23, K24

second communication paths

5

Computer system

6, 7

Computer system computer

ZMA

State model algorithm

K

Communication path between the computers 6th and 7th

MLA

Machine learning algorithm

S _T = {s _T | s _T = s ₁ or s _T = s ₂ ... or s _T = z _x }

Set of predefined situations in the traffic area for which the traffic area is monitored

Z _T = {Z _T | Z _T = z ₁ or z _T = z ₂ ... or z _T = z _n }

Set of given traffic space conditions

O _T = {o _{T |} o _T = o ₁ or O _T = o ₂ ... or O _T = O _m }

Set of given objects

S _T * = {S _T * | s _T * = s ₁ * or s _T * = s ₂ * ... or s _T * = z _x *}

Set of given situation results

Z _T * = {z _T * | z _T * = z ₁ * or z _T * = z ₂ * ... or z _T * = z _n *}

Set of given status results

O _T * = {O _T * | O _T * = o ₁ * or ο _T * = o ₂ * ... or ο _T * = o ₂ * ... or O _T * = O _m *}

Set of given object results

G

danger

H

obstacle

s _t

Operational situation of the traffic area 1

z _t

Operating condition of the traffic area 1

o _t

Object in the traffic area, which in combination with the operating status z _t the operational situation s _t or part of the forms

BA

specified operating mode of the traffic area ( 1 )

DS1 _t , DS2 _t , DS3 _t

to the operational situation s _t captured

DP1 _t , DP2 _t , DP3 _t , DP4 _t

first data (first sensor data) on the operating situation s _t recorded second data (parameter data)

s _t *

to the operational situation s _t determined operating situation result

z _t * .1

to the operating state z _t determined first operating state result

z _t * .2

to the operating state z _t determined second operating state result

O _t *

To the object o _t determined object result

FM, FM2

Error messages

UEM

Monitoring message

QUOTES INCLUDED IN THE DESCRIPTION

This list of the documents listed by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• DE 10341426 A1 [0004, 0006]
• DE 102017126553 A1 [0005]

Claims (16)

Room Monitoring method for monitoring a transport space (1), - in which an operating situation (s _t) of the transport space (1) is recognized as a predetermined situation _(T s) of the transport space (1) corresponding to when a reference to first data (DS1 _t, DS2 _t, DS3) _t *) a corresponding to the given situation _(T s) determined predetermined position result (s _T *) to the operating situation (s _t) unidentified operating situation result (s _t, - wherein the predetermined position result (s _T *) or A part of the given situation result (s _T *) is a combination of a given state result (z _T *) and a given object result (o _T *), because it is characterized by the s - with regard to the operational situation (s _t ) of the traffic area (1) based on the first data (DS1 _t , DS2 _t , DS3 _t ) a first operating state result (z _t * .1) and based on second data (DP1 _t , DP2 _t , DP3t, DP4 _t ) a second operating state result (z _t * .2) ermit and - the two operating status results (z _t * .1, z _t * .2) are checked for agreement. Room surveillance procedure according to Claim 1 , because I can denote that with regard to the operating situation (s _t ) of the traffic area (1) based on the first data (DS1 _t , DS2 _t , DS3 _t ) an object result (o _t *) is determined, which in combination with the first Operating status result (z _t * .1) forms the operating situation result (s _t *) or a part of the operating situation result (s _t *). Room surveillance procedure according to one of the Claims 1 or 2 Because I mean that an error message (FM1) is output if the two operating status results (z _t * .1, z _t * .2) do not match. Room surveillance procedure according to one of the Claims 1 to 3 , because I mean that a further error message (FM2) is output if, on the one hand, the first operating status result (z _t * .1) corresponds to a specified operating mode (BA) of the traffic area (1), if at least one selected object ( o ₁ ) is to be expected in the traffic area (1), and if, on the other hand, no predefined object result (o ₁ *) which corresponds to the expected object (or similar) is determined in a predefined time interval after the determination of the first operating state result (z _t * .1) ₁ ) corresponds. Room surveillance procedure according to one of the Claims 1 to 4th As major chgekennze I net that a monitoring message (UEM) is output if (* z _t * .1, z, _t .2), the two operating state results agree further when arriving at the operating condition (s _t) determined operating situation result (s _t *) corresponds to the given situation result (s _T *) determined for the given situation (s _T ) and if the given situation result (s _T *) is given as a hazard (G). Room surveillance procedure according to one of the Claims 1 to 5 Since I can identify that a computer system (5) of a room surveillance system (2), in particular a first computer (6) of the computer system (5), the first data (DS1 _t , DS2 _t , DS3 _t ) from a sensor device (3 ) and the first operating status result (z _t * .1) and the object result (o _t *) are determined by means of a machine learning algorithm (MLA). Room surveillance procedure according to one of the Claims 1 to 5 , as I can tell that the computer system (5), in particular a second computer (7) of the computer system (5), receives the second data (DP1 _t , DP2 _t , DP3t) from a state parameter data provision device (4) and that second operating state result (z _t * .2) determined by means of a state model algorithm (ZMA) or by means of a further machine learning algorithm. Room monitoring system (1) for monitoring a traffic area (1), with a computer system (5) which is set up to correspond to an operating situation (s _t ) of the traffic area (1) as a predetermined situation (s _T ) of the traffic area (1) recognize when a first data (DS1 _t, DS2 _t, DS3 _t) to the operating situation (s _t) unidentified operating situation result (s _t *) a corresponding to the given situation _(T s) determined predetermined position result (s _T *), - wherein the predetermined situation Result (s _T *) or a part of the given situation result (s _T *) is a combination of a given state result (z _T *) and a given object result (o _T *), because it denotes that the computer system (3rd ) is set up, - with regard to the operating situation (s _t ) of the traffic area (1) based on the first data (DS1 _t , DS2 _t , DS3 _t ) a first operating status result (z _t * .1) and based on second data (DP1 _t , DP2 _t , DP3t, DP4 _t ) to determine a second operating status result (z _t * .2) and - to check the two operating status results (z _t * .1, z _t * .2) for correspondence. Room surveillance system according to Claim 8 Since I can tell that the computer system (5) is set up to determine an object result (o _t *) based on the first data (DS1 _t , DS2 _t , DS3t) with regard to the operating situation (s _t ) of the traffic area (1) which, in combination with the first operating status result (z _t * .1), forms the operating situation result (s _t *) or a part of the operating situation result (s _t *). Room surveillance system according to one of the Claims 8 or 9 , because I mean that the computer system (5) is set up to output an error message (FM1) if the two operating status results (z _t * .1, z _t * .2) do not match. Room surveillance system according to one of the Claims 8 to 10 , characterized in that the computer system (5) is set up to output a further error message (FM2) if, on the one hand, the first operating state result (z _t * .1) corresponds to a predetermined operating mode (BA) of the traffic area (1), if it is at least a selected object (o ₁ ) is to be expected in the traffic area (1), and if, on the other hand, in a predetermined time interval after the determination of the first operating state result (z _t * .1), no predetermined object result (o ₁ *) is determined which corresponds to the expected object (o ₁ ). Room surveillance system according to one of the Claims 9 to 11 , because I can tell that the computer system (5) is set up to output a monitoring message (UEM) if the two operating status results (z _t * .1, z _t * .2) match and if this also relates to the operating situation (see Sect _t) unidentified operating situation result (s _t *) which corresponds to the given situation _(T s) determined predetermined position result (s * _T). Room surveillance system according to one of the Claims 8 to 12 , because I net that the computer system (5), in particular a first computer (6) of the computer system (5), is set up to send the first data (DS1 _t , DS2 _t , DS3 _t ) from a sensor device (3) receive and determine the first operating status result (z _t * .1) and the object result (o _t *) by means of a machine learning algorithm (MLA). Room surveillance system according to one of the Claims 8 to 13th As a result, I denote that the computer system (5), in particular a second computer (7) of the computer system (5), is set up to receive the second data (DP1 _t , DP2 _t , DP3 _t ) from a state parameter data provision device (4 ) and to determine the second operating state result (z _t * .2) by means of a state model algorithm (ZMA) or by means of a further machine learning algorithm. Computer program with program instructions which, when the computer program is executed by a computer system (3), cause the computer system (3) to carry out the steps of the method according to one of the Claims 1 to 7th execute. Provision device for the computer program according to Claim 15 , wherein the provision device stores and / or provides the computer program.

Images