"System and Method for access control in open restricted areas"

Technical field

This application relates to the field of access control systems for accessing open restricted areas.

Background art

Generally there are 2 main methods for controlling the access of people to restricted areas, i.e. areas restricted to people having a valid ticket or restricted to people having a valid authorization:

— Physical barrier access control methods, from now on referred as physical barriers: the entry or both the entry and the exit channels are physically blocked with, for example, barriers, turnstiles, gates or doors; people can only enter or leave the restricted area after some kind of mandatory validation of their credentials at entrance (ticket, card, biometric identification, etc.) which opens the physical barrier and allows the passage of the person one at a time in each direction. In physical barrier systems, when a person does not validate his/her credential the physical barrier blocks his/her access to the restricted area.

— Open access control methods, also known as "proof-Of- payment" in public transport systems, from now on referred as open access: in this type of solution the entry and exit channels are open and have no physical barriers, so multiple persons can freely enter and exit the restricted areas simultaneously. To comply with access rules, users must perform a voluntary validation of their credentials (ticket, card, biometric identification, etc.) before entering the restricted area. This validation can be done by a human operator, but is most commonly performed by an electronic validation system (validator, reader or a biometric sensor, for example) . In open access systems there are no physical means to prevent those who did not voluntarily validate their credential from accessing the open restricted area.

Therefore, open restricted areas are characterized by:

i) Not having physical barriers to block the access to the restricted area.

ii) Multiple persons being able to enter and leave the restricted area simultaneously through the open entry and exit channels; and

iii) Authorized persons must voluntarily validate their credentials before entering the restricted area, usually on an electronic device.

By their own nature, these rules are often hard to control and enforce. Detection of violations (i.e. persons who enter the restricted area without previous credential validation) is usually performed through random control by human inspectors. However, this method is highly ineffective .

Considering the specific scenario of the public transport systems as a non-limiting exemplary context, fare evasion due to a non-efficient access control of passengers represents a major issue with millions of euros in losses every year. To reduce it, public transport systems implement different types of electronic access control systems which can be included in the physical barriers or open access categories.

Physical barrier method ensures a much higher effectiveness of access control, preventing anyone without a valid ticket from entering the restricted area. However, it requires more space to implement, is more expensive, constitutes an emergency evacuation obstacle and has a lower flow rate of people at entry and at exit. In station confined and heavy passenger load transport systems, such as subways, the fare evasion issues are usually tackled using physical barriers access control.

In other transport systems, especially where access control is performed on-board the transport vehicles, such as bus, tram, bus rapid transit, light rail transit, train or other types of surface, non-station transport systems, the introduction of physical barriers is more expensive, harder to install due to space limitations, reduces the passenger flow rate and constitutes an obstacle to fast evacuation in case of emergency. As such, most of these transport systems have implemented open access methods through electronic ticket validator systems. In fact, open access control systems, especially if implemented through electronic systems, require less space, are less expensive, do not block the exit in case of emergency evacuation and allow a higher flow rate of people at entry and at exit.

In the state of the art open access control systems ([1], [2]), passengers approach their tickets to the validator device which detects the ticket and checks if the ticket is valid. For this purpose, state of the art open access control systems use any form of machine readable electronic tickets, such as magnetic stripe, proximity contactless smart card (RFID) or near field communications tickets.

The drawback of open access control systems is their lower effectiveness, as they only control those people that make a voluntary validation at entry. Therefore, they do not detect users that bypass the validator device and enter the restricted area without a valid ticket or authorization. In fact, most of the technological efforts to tackle fare evasion on state of the art open access control systems are directed towards the ticketing system itself, including tampering-proof systems and hard to fake tickets.

Considering those limitations, public transport systems, for example, complement the open access method with inspection teams that check passengers' tickets. However, these inspections are made in a random and sampled manner, therefore having a limited effectiveness in detecting fare evasion passengers and in preventing fraud, due to the public perception of sample and randomness - i.e., the offender believes he/she will not be caught "this time".

In public transport systems equipped with physical barriers, the most common form of unauthorized entry is walking in right behind someone (tailgating) , taking advantage of the time lapse before the barriers close again and the safety measures that prevent the barriers to close whenever someone is still crossing them. To cope with this, some physical barrier systems ([3]) have included ways to detect and count the number of people that cross the barrier by means of photoelectric or optical sensors. However, these solutions present some drawbacks: — they are only applicable to specific narrow and unidirectional physical barrier methods specifically designed for people passing one at a time after a mandatory validation of credentials; therefore they are not applicable in open access situations characterized by an open entry and exit channel in which multiple persons enter and leave the restricted area simultaneously in both directions and in which there are no physical barriers to prevent those who did not voluntarily validate their credential from accessing the open restricted area;

— they detect the presence of more than one person entering at the same time but cannot differentiate between persons who validated their tickets and those who didn't; therefore, they cannot identify which persons are in an irregular situation.

In the view of the foregoing difficulties with the prior art, it is herein proposed a new system and respective method to cope with control access in open restricted areas .

Finally, it is worth to mention that there are several algorithms available to perform real time detection and tracking of people in space and time. In the following list several examples of such algorithms are presented, which can be classified into three main categories:

— Algorithms using region or blob-based tracking, sometimes with additional classification schemes based on color, texture or other local image properties [5- 7, 9, 10] . — Algorithms using 2D appearance models of human beings

[4, 8] .

— Algorithms using full 3D modelling of human beings

[11, 12] .

Re erences

[1] Transport Ticketing Security and Fraud Controls, Keith E. Mayes et al .

[2] The security of mass transport ticketing system, Marc Sel et al .

[3] Transport System with Artificial Intelligence for Safety and Fare Evasion,

(https : / /ec . europa . eu/easme/en/sme/ 5840/transport-system- artificial-intelligence-safety-and- fare-evasion)

[4] I. Haritaoglu, D. Harwood, and L. S. Davis, "W4 : Real ^¬ time surveillance of people and their actions," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 22, pp. 809-830, August 2000.

[5] Q. Cai, A. Mitiche, and J. K. Aggarwal, "Tracking human motion in an indoor environment," in Proceedings of the 2nd International Conference on Image Processing (ICIP'95), pp. 215-218, 1995.

[6] C. Wren, A. Azarbaye ani , T. Darrell, and A. Pentland, "Pfinder: Real-time tracking of the human body," Tech. Rep. 353, MIT Media Laboratory Perceptual Computing Section, 1995. [7] F. Br ^' emond and M. Thonnat, "Tracking multiple non- rigid objects in a cluttered scene," in Proceedings of the 10th Scandinavian Conference on Image Analysis (SCIA '97), Lappeenranta, Finland, June 9-11, 1997, vol. 2, pp. 643- 650, 1997.

[8] A. M. Baumberg, Learning Deformable Models for Tracking Human Motion. PhD thesis, School of Computer Studies, University of Leeds, October 1995.

[9] A. J. Lipton, H. Fujiyoshi, and R. S. Patil, "Moving target classification and tracking from real-time video," in Proceedings of the DARPA Image Understanding Workshop

(IUW'98), Monterey, CA, November 1998, pp. 129-136, 1998.

[10] S. Khan, 0. Javed, Z. Rasheed, and M. Shah, "Human tracking in multiple cameras," in Proceedings of the 8th IEEE International Conference on Computer Vision (ICCV 2001), Vancouver, Canada. July 9-12, 2001, pp. 331-336, July 2001.

[11] D. M. Gavrila and L. S. Davis, "Tracking of humans in action: A 3-D model-based approach," in ARPA Image Understanding Workshop, (Palm Springs), pp. 737- 746, February 1996.

[12] H. Sidenbladh, M. J. Black, and D. J. Fleet, "Stochastic tracking of 3D human figures using 2D image motion," in ECCV 2000, 6th European Conference on Computer Vision (D. Vernon, ed.), pp. 702-718, Springer Verlag, 2000. General Description

The present application intends to solve the problem of controlling the access to open restricted areas, which have no physical barriers so multiple persons can freely enter and leave the restricted area simultaneously, of both people who voluntarily follow the rules that require validation of their credentials on a validator unit and also of those who do not follow those rules, i.e., do not validate their credentials before entering the restricted area; thus assuring that only authorized people are allowed inside the restricted areas.

It is disclosed a system and the respective method of operation which is intended to provide control of access to an open restricted area according to a completely different approach from the one known from the state of the art.

The proposed system falls under the category of open access control systems, respecting therefore its basic principle of operation which requires a prior validation step of the user before entering the area to which the access is restricted .

However, unlike the prior art systems already mentioned, the approach now followed focus its operation in the person, allowing to determine those who deceitfully, by not performing the voluntary validation step, enter the open restricted area.

According to principles described herein, the proposed system comprises a central control unit (CCU) having a special purpose processor module responsible for all processing tasks such as control, logical and mathematical computations, which handles people detection and validation status management. The system also comprises one or more sensor units, one or more validator units, an interface unit and at least one alarm unit.

The sensor units are responsible for capturing spatial data that, together with the execution of a real-time people detection and tracking algorithm in the CCU, can detect and track people in both the restricted area and the approach area where the validator units are located.

In that sense, the spatial data collected must have enough detail and frame rate to differentiate people from background, which, in moving vehicles, for example, is extremely subject to changes both in position (vibrations) and light conditions (e.g., sudden sun incidence) making the technique of background subtraction a problematic approach. At the same time, said collected data must have a low enough data rate to allow processing in real-time with reasonable computer processing capacity.

The collected data could therefore be, for example:

— A 2D image obtained by an optical sensor;

— A pseudo 3D image, also known as a range image (2D image + depth map or disparity map) , obtained by a special type of sensors called "range cameras"; these can be based on several technologies, including stereoscopy, structured light, time-of-flight , etc.

It should be noted that the aforementioned list is merely illustrative and does not intend to limit the scope of sensor technologies that can be used to this purpose. The validator units are provided with means for reading, processing and validating the person's credentials. The interface unit allows the connection to any type of validation units, being able to decode the validator's communication protocol and to convey the data to the CCU.

The alarm unit has the purpose of alerting whenever a non- authorized person has entered the restricted area.

The proposed system constantly monitors all validation events and all persons' positions.

The operation method of the proposed system is based on the assumption that at the time of a given validation event the person closest to the validator, who hasn't yet performed a validation, is the one who has executed the operation. This assumption being true, such person is then assigned a "VALID" status that indicates his or hers right to access the restricted area.

The system operates by constantly keeping track of all persons' positions and all validation events and then matching every validation event with the person closest to the relevant validator, assigning that person a "VALID" identification tag, that certifies that from then on such person is clear to enter the restricted area and also will not be considered in future validation events matchings .

Every time a person not bearing a "VALID" or "UNCERTAIN" - which indicates an exceptional situation where the system was unable to guarantee the matching accuracy - status is inside the restricted area, the alarm

Brief description of drawings

For easier understanding of this application, figures are attached in the annex that represent the preferred forms of implementation, which nevertheless are not intended to limit the technique disclosed herein.

Figure 1 illustrates the conceptual operation of the system divided in three moments in time (steps 1, 2 and 3) ; in each of these steps reference numbers represent:

1- Central control unit (CCU) ;

2- Sensor unit;

3- Validator unit;

4- Interface unit;

5- Alarm unit ;

6- Person with a valid access authorization to enter the open restricted area;

7- Person without an access authorization to enter the open restricted area;

8- Approach area;

9- Restricted area.

Figure 2 illustrates a conceptual illustration of the system disclosed, in which reference numbers represent:

1- Central control unit (CCU) ;

2- Sensor unit;

3- Validator unit;

4- Interface unit;

5- Alarm unit . Figure 3 illustrates a conceptual flowchart of the method for access control in open restricted areas using the system disclosed in each of the three steps identified in figure 1, in which reference signs represent:

A- Functional action of person entering the approach area;

B- Process action of data gathering by sensor unit(s); C- Process action of central control unit calculations to detect people in time and space;

D- Process action of central control unit creating a person record on person database with an authorization status = "NOT VALID";

E- Functional decision of person' s voluntary validation or no validation of an access authorization;

F- Functional action of person validating the authorization on the validator unit "X";

G- Process action of validator unit "X" reading and processing the person's authorization and validating it;

H- Process action of interface unit receiving and decoding the validation signal received from the validator unit "X";

I- Process action of central control unit lookup detected person in person database and checking his/her authorization status;

J- Process action of central control unit assigning an authorization status = "VALID" or "UNCERTAIN" to the detected person (s) with an authorization status

= "NOT VALID" or "UNCERTAIN" located nearest to validator "X";

K- Functional action of person entering the open restricted area; L- Process decision on what is the person' s authorization status;

M- Process action of triggering the alarm unit;

And wherein the meanings of graphical symbols are as follows :

— Full arrows: PROCESS FLOW;

— Dotted arrows: ONE PERSON FLOW;

— Boxes (rectangles) : ACTIONS;

— Diamonds: DECISIONS.

Figures 4 shows a conceptual illustration of the system implemented onboard a public transport bus in which reference numbers represent:

1- Central control unit;

2- Sensor unit;

3- Validator unit;

4- Interface unit;

5- Alarm unit ;

8- Approach area;

9- Restricted area;

10- Power supply;

11- Driver's manual control button.

Figure 5 shows another conceptual illustration of the system implemented onboard a public transport bus in which reference numbers represent:

3 - Validator Unit;

6- Person with a valid access authorization to enter the open restricted area;

7- Person without an access authorization to enter the open restricted area;

8- Approach area; 9- Restricted area.

Detailed Description

Now, preferred embodiments of the present application will be described in detail with reference to the annexed drawings. However, they are not intended to limit the scope of this application.

In the following description, for the purpose of explanation, numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, that the present invention may be practiced without these specific details.

Referring now to figure 1, the proposed system falls under the category of open access control systems to a restricted area (9) - in which the entry channel is open, having no physical barriers, so that multiple people can freely enter and exit the restricted area (9) simultaneously, respecting therefore its basic principle of operation which requires that when a given person approaches (8) the restricted area -step 1- he or she must perform a voluntary validation - step 2- of the access credential on a validator unit (3) before entering -step 3- the open restricted area (9) .

Unlike the prior state-of-the-art open access control systems already mentioned, which have a focus on the credentials (card, ticket, etc.), the approach now followed focus its operation on the persons throughout the three steps, detecting those who wrongly, by not voluntarily performing step 2, enter the open restricted zone (9) without authorization. The method of operation of the access control system disclosed consists in constantly monitoring all persons' positions and all validation events.

Whenever someone enters the approach zone -step 1- the real time people detection and tracking algorithm run by the central control unit, based on the data gathered by the sensor unit, will detect and track the location of that person and a new record is created and assigned an initial "NOT VALID" validation status.

Whenever a validation event occurs -step 2-, the central control unit decides it should be allocated to the detected person with a "NOT VALID" or "UNCERTAIN" validation status whose location, obtained by the people detection and tracking algorithm, is nearest to the validator unit (3) coordinates and assigns a "VALID" validation status to that person. If two or more persons are located near the validator unit at a similar distance, an "UNCERTAIN" status is assigned to them.

When a person (6) enters the restricted area (9) in step 3 carrying a "VALID" or "UNCERTAIN" validation status, the alarm (5) is not triggered, by applying the " n dubio pro reo" principle.

In the opposite scenario, if a person (7) does not perform the validation operation in step 2, he or she will keep the "NOT VALID" validation status initially assigned to him/her when such person was detected for the first time in step 1. Since the persons' position is continuously monitored by the system through the people detection and tracking algorithm, based on the data gathered by the sensor unit, whenever that person enters the restricted area (9) in step 3, the "NOT VALID" validation status will cause the alarm to be triggered (5) .

It should be noted that some initial uncertainty situations can be resolved later in time by the decision algorithm. One of such situations is now described as a non-limitative example: 1. At the time of a validation event, two persons (PI and P2) carrying the "NOT VALID" status are the closest to the relevant validator, at very similar distances; therefore, the system assigns them both a new "UNCERTAIN" status. 2. Later, a new validation event occurs; this time, either person PI or person P2 is the closest to the validator or the same initial situation is repeated (both persons PI and P2 are at the same distance) . Now, this can allow the system to safely change the status of both persons PI and P2 to "VALID".

This is important because, although both "VALID" and "UNCERTAIN" statuses will allow access to the restricted area, since neither triggers the alarm, they are accounted differently in terms of statistics and system accuracy metrics .

Referring now to figure 2, the system for access control in restricted areas shown comprises a main central control unit (1), one or more sensor units (2), one or more validator units (3), an interface unit (4) and one or more alarm units (5) . All units are connected by means of high speed transmission links. The necessary bandwidth will depend on the type and resolution of the sensors used and of the frame capture rate.

In one embodiment, the central control unit (1) comprises a processing module provided with computational requirements and is responsible for managing the operation of the system, implementing a finite state machine that controls all other technical units that are integrated to it, and performing, in a different process that runs concurrently, the intensive mathematical computations needed to detect people, in real time, from spatial data gathered from the sensors .

The central control unit (1) also has the ability to temporarily store data in a person' s record regarding each detected person and to perform the association of each validation event to the detected person who executed it in the validator unit (3) by assigning a positive validation status on that person's record. The central control unit (1) triggers the alarm unit (5) whenever a person with a "NOT VALID" validation status is detected inside the restricted area.

The sensor units (2) position will depend on the area being covered and other factors, such as the lenses field-of-view whenever optical sensors are used, but will generally be located above the approach and restricted areas in order to properly be able to cover said areas . The sensors gather spatial data of the approach and restricted areas with an equilibrium of i) a high enough detail and frame-rate, to allow an accurate representation of space; and ii) a low enough data rate, which is necessary for the operation of the real time people detection and tracking algorithms running in the CCU (1) with reasonable computer processing capacity. For such purpose the said sensors can be implemented through 2D cameras, stereoscopic, structured- light, time-of-flight or other range cameras, thermal or ultrasound technologies, etc. It should be noted that the aforementioned list is merely illustrative and does not intend to limit the scope of sensor technologies that can be used to this purpose.

One or more validator units (3) are provided with means for reading, processing and validation of the person' s credentials and transmitting the validation event data to the central control unit (1) through the interface unit

(4) . The validator unit (3) use any form of machine readable electronic validation system, such as, for example, magnetic stripe, proximity contactless smart card

(RFID) , near field communications systems or biometric readers .

The interface unit (4) is responsible for handling and processing low-level communication protocols between the validator units (3) and the central control unit (1) . It is a general-purpose interface, specifically designed to be compatible with all types of validator units (3), which can be configured, at software and hardware levels, as needed on each particular system installation.

The alarm unit (5) is controlled by the central control unit (1) and emits a visual alert, an audio alert, a signal alert or other types of alert whenever the central control unit (1) detects a person (7) carrying a "NOT VALID" validation status inside the restricted area. The alarm unit (5) can consist, for example, of a luminous alarm, an audible alarm, a signal transmitted alarm or any other type of external or internal (computer-generated) alarm. In a specific and preferred embodiment, the alarm includes an external red luminous portal and an internally-stored multimedia file. The sound (or other multimedia content) is output through the central control unit (1) dedicated connectors and reproduced through optional speakers/monitors. The file contents can include both buzzer/horn sounds and a recorded voice message. It should be noted that the aforementioned list of alarm types is merely illustrative and does not intend to limit the scope of alarm technologies that can be used to this purpose.

It is also disclosed a method for access control in restricted areas using the system described. The method comprises two different phases: a setup phase and an operation phase.

The setup phase is performed once at every location before the system can operate. It consists of:

— sensor unit calibration: process where the variable sensor parameters, those which are sensor-specific and, for manufacturing reasons, have slight variations from one unit to the next one, are measured and registered in a calibration file so that the central control unit can, from then on, compensate errors caused by those variations; for example, when optical cameras are used, lenses distortion must be measured and stored by means of a checkered target placed in various positions;

— definition of each validator unit location -x, y, z coordinates- in the sensor unit referential to enable the central control unit to calculate the distance between the validator unit and each detected person;

— definition of the restricted area -x, y, z coordinates- in the sensor unit referential;

— definition of the approach area -e.g., a vehicle entrance door zone- in the sensor unit referential -x, y, z coordinates- where new persons can be detected by the system for the first time. Any person who is detected for the first time out of the approach area is disregarded by the CCU in order to prevent erroneous detections caused by noise or people coming back from inside the restricted area.

Referring to figure 3, in the operation phase, when a person enters the approach area (A, 8) the 2D or 3D data is permanently being gathered by the sensor unit (B,2) to allow a representation of the space in order to detect and determine the location of every person in the approach and restricted areas. Said data is transmitted to the central control unit in order to create a model of the analyzed space in real-time.

The gathered data is then mathematically processed by the processing module of the central control unit (C,l), using adequate algorithms to perform real time detection and tracking of people in space and time. There is a wide range of people detection and tracking algorithms that can be used such as:

— Algorithms using region or blob-based tracking, sometimes with additional classification schemes based on color, texture or other local image properties;

— Algorithms using 2D appearance models of human beings; — Algorithms using full 3D modelling of human beings;

It should be noted that the aforementioned list is merely illustrative and does not intend to limit the scope of people detection and tracking algorithms that can be used to this purpose.

For each identified person, the CCU creates a temporary record (D) in a person's database -ensuring the fulfillment of legal privacy protection requirements- and assigns to it a unique ID tag, which must be kept throughout time until the said individual leaves the analyzed area. Initially, each ID tag is also assigned an authorization status = "NOT VALID" flag. While inside the approach area, each person may or may not voluntarily validate his access authorization (E) .

If the person validates his access authorization (F) on a given validator unit "X", the validator unit reads, processes and validates the person's authorization (G) . The interface unit receives and decodes (H) the signal transmitted by the validator "X", which is then conveyed to the central control unit.

Simultaneously, data continuously being gathered by the sensor unit (B) is mathematically processed by the processing module of the central control unit (C) in order to detect in real-time each person' s position in time and space .

The central control unit looks up the detected persons in the database and checks the valid authorization status of each ( I ) . The central control unit computes the distance to validator "X" of each person with a validation status = "NOT VALID" or "UNCERTAIN" and changes the validation status of the nearest person to "VALID" (J) . Or, if two or more persons located nearest to validator "X" have a similar distance (within a user definable range) an "UNCERTAIN" status is assigned to them.

Statuses marked as "UNCERTAIN" may remain as such until the person enters the restricted area or yet be upgraded to "VALID" if certain future conditions are met; for example, a validation event received while there are no candidates to have performed that operation other than a person marked as "UNCERTAIN" means that person' s status should have been "NOT VALID" before and will from then on be "VALID" with 100% certainty.

The data continuously being gathered by the sensor unit (B) is mathematically processed by the processing module of the central control unit (C) in order to detect in real-time each person's position in time and space. When a person is detected inside the restricted area (K) , because he or she just entered it, the central control unit looks up the detected person' s record in the database and checks its authorization status (I) .

If the person's authorization status = "VALID" or "UNCERTAIN" (L) , the alarm is not triggered, allowing the application of the " n dubio pro reo" principle. On the contrary, if the person' s authorization status = "NOT VALID" the central control unit will trigger the alarm unit (M) .

Figures 4 and 5 intend to illustrate the application of the system and method described in a public transport bus, as a non-limiting implementation scenario.

In fact, on public transport buses, the implementation of the proposed system and method presents additional challenges, namely due to:

— An onboard moving scenario;

— A tight entrance area making passengers walk very close to each other which poses added difficulty to the people detection algorithms;

— A limited ceiling height, reducing the area covered by the sensors;

— Continuously changing lighting conditions, including intermittent direct sunlight;

— Significant vibrations coming from the bus normal operation amplified by road imperfections;

— Electrical noise.

The preferred embodiment of the system proposed under these conditions require a compact size central control unit (1) specially designed for onboard utilization. Sensor unit (2) cannot be affected by sunlight and must have a high field of view angle. One possibility is to use a stereoscopic optical sensor, with megapixel class cameras, which allows a 3D representation of the space, but other sensor approaches can be used, dependent on the type of real time people detection and tracking algorithm to use. All components must be resistant to vibrations and protected from electrical noise through a DC/DC regulated power supply (10) .

Moreover, as required by most public transport companies, a manual control unit (11) can also be used to allow human intervention - for example, in a bus, the driver must be able to manually override the automatic operation of the system to cope with exceptional or unexpected situations, such as an on-board manual ticket sale, an inspection team boarding the bus, a system malfunction, etc. The manual control unit includes one or more buttons/switches that allow a human to perform one or more pre-defined functions, which are transmitted to the central control unit, overriding the automatic operation of the system. These can be configured differently on each installation. The buttons/switches are binary (on/off) and so is the transmitted signal, which is also handled by the I/O interface .

This description is of course not in any way restricted to the forms of implementation presented herein and any person with an average knowledge of the area can provide many possibilities for modification thereof without departing from the general idea as defined by the claims. The preferred forms of implementation described above can obviously be combined with each other. The following claims further define the preferred forms of implementation.