The present invention is directed to monitoring system for mechanical systems. BACKGROUND OF THE INVENTION

The monitoring of certain mechanical structures during operation can be demanding. For example automatic doors (swing doors, sliding doors as well as roller doors) have to be monitored when opening and closing in order to prevent them from doing any harm to a user as well as from being damaged when colliding with any obstacle. In order to detect such irregular incidents during operation of a mechanical structure, the external forces acting on such a structure as well as accelerations may have to be detected and analyzed. From the prior art a variety of sensors is known in order to provide such monitoring. These sensors include infrared and ultrasonic detectors, pressure sensitive safety bumpers, mats, plates, edges or the like, as well as single photoelectric sensors and light curtains.

However, the systems known from the prior art usually have a rather limited functionality, allowing only to detect obstacles and/or force effects in a specified spatially limited region of a mechanical structure. Consequently in many cases a multiplicity of different sensors has to be applied in order to allow comprehensive monitoring which is needed to detect a variety of potential incidents during operation. This especially holds true if multiple degrees of freedom are allowed in the operation. Hence complex and relatively expensive monitoring systems may be needed. At the same time, many of the sensors known from the prior art are susceptible to errors. As such, photoelectric sensors may be affected by bright ambient light or pollution. The same holds true for the functionality and reliability of other sensors, such as ultrasonic and infrared sensors which can also be affected by changes in the sensors' environment or e.g pressure sensitive safety bumpers which as mechanical devices may be prone to malfunction themselves. Therefore building up a reliable monitoring system can become very demanding, especially if relatively big and/or complexly shaped mechanical structures have to be monitored.

One example for a type of mechanical structure which needs comprehensive monitoring are adjustable kitchens, as e.g. disclosed in EP1077624 and US7743716. In adjustable kitchens usually multiple components of the kitchen are adjustable in height and/or depth. As such, the countertop as well as upper and lower cupboards may be adjusted in order to adapt the kitchen layout to a variety of users in order to allow an ergonomic use of the kitchen. Hence, the different components of the furniture may be displaced relative to the room as well as relative to each other, which typically is done automatically using relatively strong actuators. This poses a significant threat to the user and consequently makes a comprehensive pinch protection important. Yet, as soon as many different components can be adjusted independently from each other and in various directions, pinch protection for all potentially dangerous spots by using monitoring systems known from the prior art becomes complex. EP 1 657 022 B1 was filed on 08.1 1 .2005 on behalf of Fanuc Ltd. and describes an impact detection device for detecting when an object collides with another object. The document is in particular directed to the detection of a collision between a spindle of a machine tool and another object. According to the document a vibration sensor is attached to an object to be monitored and an output signal of the vibration sensor is sent to a receiving unit which comprises a means for holding the largest value in a continuous output signal as peak value. The impact detection device further comprises a collision judging means that is calculating an average output level of output signals received from the vibration sensor via the receiving unit. According to this document, the peak value is compared with an averaged output signal received from the vibration sensor over a predetermined time. The collision judging means assumes a collision if the difference between the peak value and the averaged output signal exceeds a predetermined value. The peak value is output as a digital signal by an output unit.

DE 10 2012 021 272 A1 was filed on 29.1 0.201 2 on behalf of Keihin Corporation and discloses a device to detect a collision of a vehicle. The document is in particular directed to be used for a supplemental restrain system airbag system. According to the document, a vibration sensor is used in order to detect vibrations in a high frequency band as well as in a low frequency band, the bands being above respectively below the acoustic frequency band. The device is directed to be used for a vehicle that comprises an extension frame with a plurality of impact absorbing zones. Using calculating means the change of energy of the high frequency vibrations and the integral value of the low frequency vibrations are com- puted. These two values are used in order to determine if one of the impact absorbing zones breaks and hence if a specific threshold which is used to for collision detection has to be adapted.

US4616887 was filed on 23 .04.1 984 on behalf of Dirk J. Oudman and shows an adjustable work top attachable to a wall. The system shown includes at least one means for vertically adjusting said work top that is provided with a horizontally extending carrier arm for carrying said work top. The publication further shows a safety arrangement using a light source or infrared source and a photosensitive or infrared-sensitive cell, which in case of an interruption of the light beam causes the system's motor to be switched off at least during a downward movement of the work top.

DESCRIPTION OF THE INVENTION

It is therefore one object of the invention to provide an improved monitoring system that allows comprehensive monitoring of a mechanical structure during operation. Another object of the invention is directed to a collision detection device that can be used in order to prevent a user from being harmed by a mechanical structure, as well as to prevent the mechanical structure from being damaged during operation.

It is a basic idea of the present invention to use the existing load paths of a mechanical structure to obtain a comprehensive monitoring of the whole structure during operation by using at least one sensor arranged. The underlying principle may be seen as that a force, torque, local change in temperature or other physical incident that takes place at a certain location of a mechanical structure must have a detectably effect on the structure. Consequently the physical incident can be detected remotely on the structure by using existing load paths and the system can be learned to said physical incident and its qualification. For example a force acting on one side of a large framework will not only cause local deformation in the vicinity of its point of action, but will also result in remote deformation of the structure, hence e.g. evoke a different strain field in remote parts of the frame ^¬ work. For many materials and/or types of structures such changes in the local geometry - respectively strain - occur almost simultaneously with the application of such an external force. In addition, other physical phenomena, such as structure-borne sound propagates at a very high speed. Therefore local incidents resulting in local changes of a mechanical structure can virtually be detected instantaneously at any location of said mechanical structure.

It is a concept of the present invention to provide a system for detecting mechani ^¬ cal impacts or collisions on a mechanical structure during operation which com ^¬ prises at least one sensor that is interconnected to the mechanical structure in the area of a load path of the structure. For some applications good results may be obtained if the sensor is interconnected rigidly to the mechanical structure, however the invention is not limited to such types of interconnections. The at least one sensor senses (measures) physical characteristics of the mechanical structure and converts them into at least one sensor output signal. During the operation said sensor senses, respectively registers, changes in the physical characteristics of the mechanical structure caused by at least one mechanical incident. The physical characteristics may for example be strain, stress, deformation/def lection, forces, temperature, combinations thereof and others. Incidents causing such changes may be both spatially nearby and remote from the sensor. The at least one sensor output signal is transmitted to at least one controller.

In a variation of the invention, during operation said at least one controller may perform at least two steps iteratively. In a first step, the controller computes an expected sensor output signal for a subsequent time interval. The expected sensor output signal can also be seen as an estimation of a regular sensor output signal that would be expected within the subsequent time interval. This computation will typically base on the actually measured sensor output signal during an antecedent time interval. In order to compute the expected sensor output signal a variety of algorithms may be applied, as will be described hereafter. In a second step, during the subsequent time interval the controller compares the actually measured real at least one sensor output signal with the expected sensor output signal. Such a comparison may be used in order to decide whether an irregular incident during operation has occurred. At the latest when this time interval is over, a new expected sensor output signal is computed for the following (subsequent) time interval. The different steps do not have to be processed by the same controller, hence they may also be processed by an assembly of controllers.

Due to the fact that in such a variant of the invention the expected sensor output signal is constantly adjusted based on the antecedent sensor output signal, regular changes in the mechanical structure which will typically occur in a rather steady way can be sufficiently predicted and hence be identified. In contrast to this, irregular incidents (anomalies), which will typically occur in a more abrupt way will lead to a clear deviation between the sensor output signal as measured and the expected sensor output signal. It is a main advantage of the present in ^¬ vention that the system automatically adapts to regular changes in the mechanical structure. In adjustable kitchens such regular changes may e.g. be induced by ad ^¬ justments in depth and height of upper cupboards, which will cause changes in the load distribution within the whole mechanical structure. The same holds true if the cupboards and/or the countertop is loaded by kettles, pans and tableware in various ways, which will lead to various standard load cases.

Thus - in contrast to monitoring systems known from the prior art - a monitoring system according to the present invention is capable to adjust automatically to changes in its environment based on a certain mindset made available to it, hence could also be referred to as a self-learning system.

In a variation of the invention, the controller surveys if the sensor output signal lies within a certain band described by an upper and a lower threshold. Hence irregular incidents can be detected and supercritical changes in the mechanical structure can be detected. Within the context of the present invention, single or multiple upper and multiple lower thresholds may be used.

However, if appropriate, also only one or multiple upper thresholds respectively only one or multiple lower thresholds may be applied for signal processing. Depending on the application, the upper and/or lower thresholds may have constant values during operation and may be pre-set and fixed. In another variation of the invention, the values of the upper and/or lower thresholds vary during operation, depending e.g. on operation parameters and/or physical characteristics measured by a sensor.

If appropriate, an upper threshold may be given by a specified upper deviation from an expected sensor output signal and/or a lower threshold may be given by a specified lower deviation from an expected sensor output signal. Thus, for many applications irregular incidents in can be reliably detected as such events will typically occur abruptly and hence will result in a significant deviation between the actually measured sensor output and the expected sensor output signal. Consequently the measured sensor output signal will exceed the upper threshold or go below the lower threshold, which can both be used as trigger criteria for the controller in order to decide whether an irregular incident occurs or not.

For some applications, the upper deviation and/or the lower deviation may be constant during operation. Hence they may be pre-set, adjusted on the particular mechanical structure and/or the types of regular and irregular incidents that are likely to happen during operation. Therefore the values of these deviations will typically depend on the specific application of the monitoring system. Thus they may be parameterized during an initial calibration procedure where the upper and lower deviations will be set such that regular incidents during operation will not result in exceeding or undershooting of the thresholds. Hence the controller may be calibrated on a specific mechanical structure and the parameters determined during the calibration process can be copied to other controllers used for the type of mechanical structure. Consequently large-scale production of e.g. monitoring systems for adjustable kitchens becomes possible as a time consuming installation and calibration of a monitoring system can be omitted.

For some applications the controller may be integrated in the sensor. Thus, relatively small monitoring systems may be made.

For some special types of mechanical structures the upper deviation and/or the lower deviation may vary during operation. Such embodiments of the invention may e.g. be used for mechanical structures that show regular largely altering structural properties during operation. The deviations may be pre-defined, depending on the operating state of the mechanical structure. As for adjustable kitchens, the upper and lower deviations applied for adjustments of the counter- top close to the topmost and/or bottommost position may be different from the deviations at positions in between. However, the upper and/or lower deviation may also depend on the value of the expected sensor output signal or on the value of the sensor output signal in an antecedent time interval.

Alternatively or in addition, the upper and/or lower deviation may depend on at least one sensor output signal. As such, they may depend on the same at least one sensor output signal which is used in order to compute the expected sensor output signal. However it may also depend on another sensor output signal. Such a variation of the invention may e.g. be advantageous in order to compensate for changes in the ambient temperature which may affect the mechanical structure during operation.

Depending on the application, the expected sensor output signal during operation or at least part of it may be retrieved from a memory device and used as a reference signal. Such a variant of the invention may be advantageous for mechanical structures which always undergo the same type of operation in a highly reproducible manner.

Good results may be obtained if the positioning of the at least one sensor bases on real or virtual experiments where different types of operations and incidents are simulated in order to determine the load paths and hence find out where these incidents can be optimally captured. Virtual experiments may include computer simulations, such as finite element analyses.

For some applications, good results may be achieved if the monitoring system comprises at least one additional sensor which senses the same and/or an additional physical characteristics. Such a variation of the application may be advantageous in order to obtain a particularly reliable and comprehensive monitoring system which can also be used for very large or fragmented mechanical structures. Such additional sensors may be arranged spatially close to a first sensor or may be arranged at other areas, e.g. in order to capture multiple load paths in certain special mechanical structures. In one variation of the invention, the monitoring system may comprise at least one additional controller. Such a supplementary controller may be used in order to monitor additional (different) physical characteristics. Alternatively or in addition, an additional controller may serve as a backup controller in order to increase the reliability of a monitoring system. This will be described in further detail below.

Depending on the field of application a variety of controller types may be used, including hard wired transducers as well as programmable computers. For some applications, good results may be achieved using single-board computers (SBC) - respectively microprocessors - such as Raspberry Pi, Arduino or similar devices. Controllers may also comprise combinations of such devices. Thus, in contrast to other systems known from the prior art, for many applications a comprehensive monitoring can be obtained by applying relatively inexpensive controllers.

The present invention is not limited to a specific type of sensor. However good results may be obtained if the at least one sensor comprises a strain gauge and/or a piezoelectric sensor. Good results may be obtained using four-wire or six-wire strain gauges, as these allow capturing even complex changes in strain fields. Such types of sensors offer a reliable and relatively inexpensive way to measure local deformation ( respectively changes in the local strain field). In addition, such sensors can be interconnected to a mechanical structure using adhesives, allowing easy installation without affecting the structural competence of the mechanical structure. If appropriate, the at least one controller may transmit at least one controller output signal to at least one other device. Such a device may include controllers for the operation - such as a motion control device - an actuator or a signal device, such as an optical or acoustic alarm device. Thus for example a process can be stopped or even at least partially be reversed. As for an automatic door, a control ^¬ ler output signal may result in a closing operation being cancelled as a person may be trapped by the door. For some purposes a controller output signal may also result in an operation being at least partially reversed, such as in order to release a person or object trapped or pinched.

Alternatively or in addition, the controller may also analyze the shape of the sen ^¬ sor output signal curve (curvature) and generate different controller output sig ^¬ nals for different types of sensor output signals. Such a variation of the invention may be used in order to recognized different types of irregular incidents and generate a controller output signal that is appropriate for specific types of irregular incidents. Such a variation of a monitoring system may e.g. be used in order to control adjustments of a desk that is adjustable in height. As such a sensor output signal evoked by a user who is pressing with one finger on the bottom face of the table board may cause an actuator to elevate the table board, whereas pressing with one finger in a specific region of the top face may result in lowering the table board. Hence in a variant of a monitoring system according to the invention, the monitoring system may also be used as a controller.

Good results may be obtained if the shape of the sensor output signal curve (cur ^¬ vature) is only be analyzed if it exceeds the upper threshold or lies below the lower threshold. Hence signals caused by regular incidents during operation do not have to be analyzed, which allows applying relative simple algorithms for signal analysis.

According to one variation of the invention the monitoring system may comprise a unit that is monitoring the status of the sensor for errors, such as sensor break or short-circuit. This unit may be the controller or may also be a separate device. Hence an operation may be prohibited if reliable monitoring cannot be ensured.

Alternatively or in addition, the monitoring system may comprise at least one measuring amplifier which amplifies the sensor output signal. Such a variation may be advantageous in order to use the same controller and sensor for different applications.

For some applications the at least one controller and/or sensor output signal may be transmitted by a wired or by a wireless connection. Wireless transmission may be advantageous for relatively large mechanical structures or e.g. to monitor structures that operated in hermetically sealed volumes which should not be influ ^¬ enced by cables.

In a variation of the invention the monitoring system may comprise a data storage unit that saves data related to the usage of the monitoring system and/or physical characteristics captured by the sensor. Such data may be used for the mainte ^¬ nance monitoring system and/or the mechanical structure to be monitored. For some applications, the monitoring system may comprise a data storage unit, such as an EEPROM device, in which pre-set parameters, such as upper and/or lower deviation, are saved. Such a variation of the invention may be advantageous for monitoring systems that are not always supplied by power.

In a variation of the monitoring system the at least one controller and/or the at least one sensor are activated shortly before or as soon as operation starts. Such a variation of the invention may be used if the monitoring system is only used during operation, such as to provide user safety.

In another variation of the monitoring system, the at least one controller and/or the at least one sensor are always activated. Such a variation of the invention may be used if the monitoring system is also used as a controller, such as to start a certain operation as described above.

In a variation of the invention multiple thresholds may be used. As such for adjustable furniture, different upper and lower deviations may be pre-set, depending on speed and direction of the motion.

A monitoring system according to the present invention may also be used in order to monitor automatic doors, such as swing doors, sliding doors as well as roller doors or even pool covers, such as swimming pool covers. Hence they may e.g. be used in order to detect collisions with persons, animals or other obstacles during closing or opening processes. Alternatively or in addition they may be used in or ^¬ der to detect unwanted mechanical manipulation. In case of pool covers a moni- toring system according to the invention may also be used in order to detect e.g. people who fell on the pool cover and hence may be in danger of drowning.

A monitoring system according to the present invention may also be used for monitoring the operation of a mechanical structure during extended time intervals in order to make adjustments to the mechanical structure if needed. As such a monitoring system may be used in order to adjust the knife blades of harvesters which from time to time have to be re-adjusted (repositioned) due to the deposition of dirt during usage. Therefore according to the present invention a sensor may be positioned in some distance from the knife blades and still be able to detect misalignment of the blades.

A monitoring system according to the present invention may also be used as a safety device in order to detect when people slip and/or fall in a shower or bathtub. Therefore a sensor may be attached e.g. to a shower tray, respectively bathtub. As the impact force which a human body exerts on a shower tray or bathtub when falling is significantly higher than when standing, an impact can reliably be detected. In addition, the strain field in a shower tray or bath tub as induced by lying person is significantly different than the strain field induced by a person standing. Hence such a monitoring system may be used in order to detect slipping and falling of a person and consequently e.g. an alarm can be activated.

In addition, a monitoring system according to the invention may also be used for ski bindings. Thanks to the invention even highly complex load cases between ski binding and ski boot can be captured and assessed by the controller in order to decide whether the binding should release the boot or not. A monitoring system according to the invention may also be used in order to monitor structural steelwork in order to detect overloading or other incidents.

In order to compute an expected sensor output signal several approaches, respec ^¬ tively algorithms are possible. However, it is clear that the invention is not limited to these approaches.

In a first variation of the invention, the measured sensor output signal of an antecedent time interval is averaged or filtered in order to remove noise and then extrapolated in order to obtain the expected sensor output signal during the subse ^¬ quent time interval. In order to obtain a good estimate of the expected sensor output signal, the first and/or second order derivative of the measured sensor output signal may be obtained and used for the extrapolation. If appropriate, also higher order derivatives may be used. The length of the antecedent time interval which is taken into account in order to compute the expected sensor output signal, as well as the length of the subsequent time interval may depend on the mechanical characteristics of the mechanical structure and/or operation parameters, such as e.g. the speed of a certain motion.

One aspect of the invention is directed to a collision detection device for an adjustable kitchen having several degrees of freedom which provides comprehensive monitoring of all virtually all types of adjustments feasible with such a kitchen. By using a monitoring system according to the present invention, collisions between the kitchen and users or objects can reliably be detected. Therefore a sensor of said monitoring system may be interconnected to e.g. a framework which interconnects the various components (e.g. cupboards and countertop) of such a kitchen. By this, the whole kitchen, which can be regarded as a mechanical structure including framework, cupboards and countertop, can be monitored. Good results may be obtained if the sensor is positioned at an area which is important for the structural competence of the whole mechanical structure. This may for example be a beam which is connected to multiple other beams and hence be influenced by many load paths.

In a variant of such a collision detection device, a controller connected to said sensor will compare the sensor output signal with an expected sensor output signal, as described above. In accordance with the present invention it has been found that reliable collision detection can be obtained using upper and lower thresholds defined by fixed upper and lower deviations, relative to the expected sensor output signal. A measured sensor output signal that exceeds the upper threshold or goes below the lower threshold will be considered by the controller as a collision and may result in a variety of controller output signals. Such signals may be transmitted to the motion controller which controls actuators that cause the adjustments of the adjustable kitchen. In a variant of the invention, a collision detection device may transmit signals to said motion controller in order to stop and/or reverse a certain adjustment.

BRIEF DESCRIPTION OF THE DRAWINGS The herein described invention will be more fully understood from the detailed description of the given herein below and the accompanying drawings, which should not be considered as limiting the invention described in the appended claims.

Fig. 1 schematically shows a first embodiment of the detector system according to the invention;

Fig. 2 schematically shows a second embodiment of a detector system according to the invention;

Fig. 3 schematically shows a third embodiment of a detector system according to the invention;

Fig. 4 schematically shows a fourth embodiment of a detector system according to the invention;

Fig. 5 schematically shows a fifth embodiment of a detector system according to the invention;

Fig. 6 schematically illustrates the value of a sensor output signal over time together with an upper and a lower threshold;

Fig. 7 schematically illustrates the value of a sensor output signal over time together with an upper and a lower threshold; Fig. 8 schematically illustrates a possible signal curvature for a first type of irregular incident;

Fig. 9 schematically illustrates a possible signal curvature for a second type of irregular incident;

Fig. 1 0 schematically shows a possible sensor output signal over time; Fig. 1 1 schematically shows a possible sensor output during operation; Fig. 1 2 shows detail A of Fig. 1 1 ;

Fig. 1 3 shows an embodiment of an adjustable kitchen with a collision detection system according to the present invention;

Fig. 1 4 shows a support arrangement for an adjustable kitchen as shown in Fig.

1 3, a sensor of the collision detection system according to the present invention;

Fig. 1 5 shows detail B of Fig. 1 4. DETAILED DESCRIPTION OF THE INVENTION

The foregoing summary, as well as the following detailed description of the preferred embodiments, is better understood when read in conjunction with the appended drawings. For the purposes of illustrating the invention, an embodiment that is presently preferred, in which like numerals represent similar parts through- out the several views of the drawings, it being understood, however, that the in ^¬ vention is not limited to the specific methods and instrumentalities disclosed.

Figure 1 schematically shows an embodiment of a monitoring system according to the invention. The monitoring system comprises a sensor S 1 that is interconnected to a mechanical structure 1 which is to be monitored. The sensor S 1 senses physical characteristics of the mechanical structure 1 , such as e.g. deformation. Therefore the sensor may e.g. be a strain gauge or a piezoelectric sensor. The measured physical characteristics are converted into a first sensor output signal 1 1 . The first sensor output signal 11 is transmitted to a first controller C1 which compares said signal with at least one reference signal. Depending on the result of this comparison, a first controller output signal 1 may be sent from the first controller C1 to a first device D 1 . The first device D 1 may be a controller that controls an operation which affects the mechanical structure 1 . As such, the device may e.g. be a motion controller which controls actuators that act on the mechanical structure 1 .

Figure 2 shows another embodiment of monitoring system according to the present invention. In such a variant of the invention a first sensor S 1 is connected to a first and to a second controller C1 , C2. The first sensor S 1 sends a first sensor output signal 1 1 to the first controller C 1 and a second sensor output signal I2 to a second controller C2. The first and second sensor output signal 11 , I2 may be identical or may be different signals. The signals may e.g. be different if the first sensor S 1 is able to sense multiple types of physical characteristics, such as e.g. strain and spatial orientation. Subsequently, the first and second controllers C 1 , C2 compare the signals sent to them with reference signals and generate a first controller output signal 01 , respectively second controller output signal 02. The first and second output signals 01 , 02 are then transmitted to a first, respectively second device D 1 , D2. In Figure 3, another embodiment of a monitoring system is shown, this embodiment comprising three sensors S 1 , S2, S3 which are interconnected to a mechanical structure 1 . The sensors S 1 , S2, S3 may be interconnected to the mechanical structure 1 at virtually the same location or in some distance from one another. In the embodiment shown in Figure 3, the first sensor S1 may be a strain gauge, measuring strain in the adjacent mechanical structure 1 and converting the measured strain into a sensor output signal 1 1 . The second sensor S2 may be a torque sensor that measures a torque applied to the mechanical structure 1 . The measured torque is converted into a second sensor output signal I 2. The third sensor S3 may be a gyroscope which registers changes in the spatial orientation of the me- chanical structure 1 and converts them into a third sensor output signal I3. All sensor output signals 1 1 , I2, I3 are transmitted to the same first controller C 1 , which generates a first and a second controller output signal 01 , 02 that are sent to two different devices D 1 , D2. In order to obtain reliable fail-safe monitoring, the monitoring system comprises a second controller that works as a backup con- troller CB. Such an arrangement of controllers C1 , CB may work in different modes.

In a first mode, the sensor signals S 1 , S2, S3 are computed by the first controller C 1 only, while the backup controller CB is in standby mode. If the first controller C1 fails, the backup controller CB will take over the computing of the three signals S 1 , S2, S3 mode, the sensor signals 51 , S2, S3 are computed by the first controller C 1 as well as by the backup controller CB. The two controllers C1 , CB may compute the signals the same way, hence e.g. using the same algorithm, or they may compute the signals S 1 , S2, S3 using different approaches. In this mode the results of the two controllers are compared before controller output signals 01 , 02 are sent to the devices D 1 , D2. Hence malfunctions in signal processing can be detected and appropriate controller output signals 01 , 02 can be sent to the devices D 1 , D2. Such controller output signals 01 , 02 may e.g. be a command to immediately stop and/or reverse a certain type of operation.

It is clear that within scope of the present invention also other concepts of redundancy may be applied - such as triple modular redundancy - and that these concepts may also be applied for other sensor-controller designs.

As shown in Figure 4, in some embodiments of the invention only one first controller C 1 may be used in order to process the signals 1 1 , I 2, I3 coming from three different sensor S 1 , S2, S3. Hence, for many applications backup systems can be omitted as the monitoring system according to the present invention is relatively simple if compared to other types of monitoring systems and thus not prone to malfunction.

Figure 5 shows another embodiment of the invention with a first sensor S 1 and a second sensor S2 being interconnected to a mechanical structure. In the embodiment shown the two sensors S 1 , S2 are different types of sensors, the two sen- sors' output signals 11 , 12 are processed by the controller C1 , which sends a first controller output signal 01 by a wired connection to a first device D 1 , while a second and a third controller output signal 02, 03 are sent by wireless connection to a second, respectively third device D2, D3.

Figure 6 illustrates a first variant of how a sensor output signal 2 (dotted line) could be processed by a controller over time. In the variant shown, an expected sensor output signal 3 is computed based on the sensor output signal 3 as measured in an antecedent time interval. The expected sensor output signal 3 will typically be an extrapolation of an average sensor output signal 3 during the antecedent time interval. Hence this averaged signal will have a more steady form than the actually measured sensor output signal 2. Figure 4 shows as well an averaged expected sensor output signal 4 which represents an average of the expected sensor output signal 3, averaged during multiple time intervals. In the variant shown in Figure 4, the upper threshold 5 and the lower threshold 6 are both set by a fixed upper, respectively lower deviation from the expected sensor output 3. Most of the time the measured output signal 3 stays in the band between the upper and lower threshold 5, 6. These signal values are regarded as being caused by regular incidents. As shown the sensor output signal can fluctuate, said fluctuation maybe being caused by The fluctuation of the signal may be caused by noise or by actuators used for the operation. However Figure 5 also shows three intervals in which the signal exceeds the upper threshold and two intervals during which the signal goes below the lower threshold. These intervals are highlighted by solid lines. Depending on the application, such intervals may be regarded as irregular incidents.

Figure 7 shows an alternative variant of how a sensor output signal 2 could be processed by a controller over time. In contrast to the approach illustrated in Figure 6, here the upper and lower threshold 5, 6 do not follow the expected sensor output signal 3 with a fixed upper and lower deviation, as the two deviations vary, based on e.g. at least one additional sensor output signal (not shown) which causes changes in the upper and lower deviation over time.

Two intervals in which the measured sensor output signal 2 exceeds a threshold 5 are shown in greater detail in Figures 8 and 9. As indicated, these intervals may be further analyzed, such as their duration t and maximum deviation from the threshold 5 may be determined. Such information may be used in order to identify what type of irregular incident has happened and hence e.g. to generate an appropriate controller output signal to a device (both not shown). Due to such a kind of signal analysis, a variant of the monitoring system can also be used as a control system for the operation of the mechanical structure. As such the monitoring system can e.g. detect where and how the structure is touched by an operator/user. Depending on the location, direction and strength of a user interaction, different types of sensor output signals will be generated, as schematically shown in Figure 10. Thus, a controller can generate predefined controller output signals which can be used to control the operation of the mechanical structure. An example of signal processing over time is shown in Figure 1 1. A measured sensor output signal 2 is compared to an upper and a lower threshold 5, 6. If the measured sensor output signal leaves the band defined by the upper and lower threshold 5, 6 the controller decides for an irregular incident 7.

In Figure 1 2 one can see that in this variant of the invention the upper and lower thresholds 5, 6 are defined based on an upper and lower deviation from an averaged expected sensor output signal 4.

Figures 13 to 1 5 show an adjustable kitchen 1 0 with a collision detection system according to the invention. The adjustable kitchen 1 0 as shown comprises a coun- tertop 1 1 which can be adjusted in height relative to the floor. In addition it comprises upper cupboards 1 2 which can be adjusted in height and depth and are shown in a partially lowered position. As well, the lower cupboards 1 3 can be adjusted in depth, which e.g. allows a person sitting in a wheelchair to move closer to the countertop 1 1 . In order to monitor all these types of adjustments a collision control device according to the invention virtually uses most of the kitchen 1 0 as a sensitive mechanical structure 1 .

Therefore a sensor S 1 is interconnected to a part of a framework 1 4 with which the countertop 1 1 as well as the upper and lower cupboards 1 2, 1 3 are mechanically interconnected, as shown in Figure 14 and 1 5. In the embodiment shown, the sensor S 1 is a strain gauge that measures strain in a beam which plays an important structural mechanical role in said mechanical structure 1 . The sensor S1 generates a sensor output signal which is transmitted to a controller C1 that is located in an upper cupboard 1 2.

LIST OF DESIGNATIONS

1 Mechanical structure S2 Second sensor

2 Sensor output signal S3 Third sensor

3 Expected sensor output signal C 1 First controller

4 Averaged expected sensor outC2 Second controller

put signal 20 CB Backup Controller

5 Upper threshold D 1 First device

6 Lower threshold D2 Second device

7 Irregular incident (critical value) 11 First sensor output signal 1 0 Adjustable kitchen I2 Second sensor output signal

1 1 Countertop 25 I3 Third sensor output signal

1 2 Upper cupboard 01 First controller output signal

1 3 Lower cupboard 02 Second controller output signa

1 4 Framework 03 Third controller output signal S 1 First sensor