Smart connector

Abstract

Smart connector (1) for the connection of a sensor array (2), to a processing unit (3), the connector (1) comprising a conversion unit (4) for receiving signals from a plurality of signal sources (9), each received signal having a signal intensity, and generating digital data signals, a controlling unit (5) for receiving the digital data signals and performing a first processing and data compression, the controlling unit (5) comprising a filtering module (11) to select a set of data signals based on the signal intensity of the signals received from the plurality of signal sources (9), wherein the selected set of data signals comprises only digital data signals originating from signal sources (9) that generated a signal having a signal intensity above a predetermined threshold, and a digital bus interface (6) connected to the controlling unit (5) for forwarding the selected set of data signals to the processing unit (3) for a second processing.

Description

[0001] SMART CONNECTOR

[0002] The present invention relates to a smart connector for the connection of a sensor array to a processing unit. Also, the present invention relates to a method for transmitting data between a sensor array and a processing unit. Furthermore, the present invention relates to a network, in particular a network of smart connectors, a device, in particular a robotic device, comprising the smart connector and a system, in particular a robotic system, comprising said device.

[0003] The increasing recent developments in artificial intelligence show growing interest in robotics and in particular in the design and employment of robots having human characteristics. These can be used in different fields of technologies, such as mechanics, medical science, space science, agriculture, etc..

[0004] The main aim of researchers and engineers working in the field of robotics is that the robots can operate in social spaces with “human-like” capabilities. Among the different senses used for interacting with the external environment, the tactile stimulation is one of the most important. Therefore, it is fundamental in robotics to develop a method for touch identification that can be active over all or most of the robot’s surface area; this could be achieved using an artificial skin. In prior art, an artificial skin is usually considered to be a flexible, stretchable array of sensors that fits onto curved robot surfaces of substantial extent. It may have the ability to sense tactile information such as pressure, texture, and temperature. In some cases, multiple layers of heterogeneous sensors can be used. Additionally, soft materials have been used to cover the sensors, to improve wettability and friction properties and to increase the contact area. It is noted that smart fabrics can be used in almost all areas of mechanical engineering in the future and not only in robotics; this to make surfaces “functional”. For example, an artificial skin can advantageously be used for monitoring elements or devices that are subject to mechanical stress and impacts, such as the wings of wind turbines (e.g. to monitor objects hitting the blades).

[0005] The employment of touch sensor arrays such as conductive fibers, e.g. piezoresistive or piezoelectric fibers, is known in the art for producing an artificial skin. In case of piezoresistive fibers, the strain or pressure variation is measured for each sensor thereby determining a change of the electrical resistance having a particular value that can be converted in an electrical signal having a particular amplitude or intensity based on the force exerted on the sensor array. In case of piezoelectric fibers, strain or pressure variation would generate a charge, i.e. a signal having a particular amplitude or intensity based on the force exerted on the sensor array.

[0006] A relevant application of these conductive fibers or smart textiles are in so called artificial skin of robotics tools such as “robot-hand” or robotic grabber. The challenge in robots is however the very limited space for connectivity and limited capacity of interconnecting busses and a processing unit (CPU). In particular, the challenge relates to the handling of big amount of sensor signals coming from the sensor array (e.g. smart textiles) that need to be forwarded to the processing unit by using a limited number of connecting elements (connection bus).

[0007] Examples of the present disclosure seek to address or at least alleviate the above problems.

[0008] In a first aspect, there is provided a smart connector for the connection of a sensor array to a processing unit, the connector comprising: a conversion unit for receiving signals from a plurality of signal sources, each received signal having a signal intensity, and generating digital data signals; a controlling unit for receiving the digital data signals and performing a first processing and data compression, the controlling unit comprising a filtering module to select a set of data signals based on the signal intensity of the signals received from the plurality of signal sources, wherein the selected set of data signals comprises only digital data signals originating from signal sources that generated a signal having a signal intensity above a predetermined threshold; and a digital bus interface connected to the controlling unit for forwarding the selected set of data signals to the processing unit for a second processing.

[0009] In a second aspect there is provided a network comprising a plurality of smart connectors according to the first aspect, wherein each smart connector of the network is configured to connect one or more signal sources of the sensor array to a processing unit.

[0010] In a third aspect there is provided a device, in particular a robotic device, comprising: a sensor array including a plurality of signal sources, the sensor array covering at least in part a portion of a frame of the device; and at least a smart connector according to the first aspect or a network according to the second aspect for the connection of the sensor array to a processing unit.

[0011] In a fourth aspect there is provided system, in particular a robotic system, comprising: a device according to the second aspect; and a processing unit connected to the device through the smart connector.

[0012] In a fifth aspect there is provided a method for transmitting data between a sensor array and a processing unit, the method comprising: receiving signals from a plurality of signal sources, each received signal having a signal intensity; generating digital data signals; receiving the digital data signals and performing a first processing and data compression; selecting a set of data signals based on the signal intensity of the signals received from the plurality of signal sources, wherein the selected set of data signals comprises only digital data signals originating from signal sources that generated a signal having a signal intensity above a predetermined threshold; and forwarding the selected set of data signals to the processing unit for a second processing.

[0013] Other aspects and features are defined in the appended claims.

[0014] Examples of the disclosure may provide a connector, a device, a system and a method able to optimize the managing of a big amount of sensor signals received from a sensor array, for example a smart textile, to be forwarded to a processing unit. Also, examples, of the disclosure may provide a connector, a device and a system that are compact and robust. It is noted that in standard sensor applications, only a few sensor signals are sent from a sensor (e.g. strain gauge) to a processing unit (e.g. a CPU). By using smart fabric, thousands of parallel, analogue signals, need to be collected, processed and transmitted. The examples of the present disclosure provide a solution for compressing these many parallel signals into a compact serial signal for the processing unit, without flooding the processing unit with irrelevant data (e.g. areas of the robotic hand which are not touching the object).

[0015] Examples of the disclosure will now be described by way of example only with reference to the accompanying drawings, in which like references refer to like parts, and in which:

[0016] Figure 1 A is a schematic representation of the smart connector according to an example;

[0017] Figure 1 B is a schematic representation of a network of smart connectors according to an example;

[0018] Figure 2 is a schematic representation of the components of the smart connector according to an example;

[0019] Figure 3 is a schematic representation of the sensor array according to an example; and Figure 4 is a flow chart describing the method according to an example.

[0020] A smart connector, a network, a device and a system as well as a method for transmitting data between a sensor array and a processing unit are disclosed. In the following description, a number of specific details are presented in order to provide a thorough understanding of the examples of the disclosure. It will be apparent however to a person skilled in the art that these specific details need not be employed in order to practice the examples of the disclosure. Conversely, specific details known to the person skilled in the art are omitted for the purposes of clarity in presenting the examples. Figure 1 A schematically illustrates a smart connector 1 for connecting a sensor array 2 to a processing unit 3 and therefore transferring data between the sensor array 2 and the processing unit 3. The sensor array 2 can be a touch sensor array, formed by a plurality of signal sources 9 arranged for example according to a two-dimensional or three-dimensional pattern. According to the figure, the sensor array 2 is represented by a sensor surface comprising a group of sixteen signal sources (i.e. sensors) arranged according to a 4x4 square. It is noted that any suitable arrangement can be used for the purpose of the present disclosure. The signal source 9 can generally be a conductive or sensing fiber / filament. In other words, the signal source 9 can be represented by smart fibers or in general by a smart textile. Each signal source 9 generates a signal, or output signal, having a signal intensity. The output signal is the response of the sensor to an input quantity of a physical phenomenon / parameter and the signal intensity is directly related to said input quantity. The greater the input quantity, the higher the signal intensity. The output signal can be directly generated upon the response of the sensor or can be the result of a conversion of a physical quantity variation by means for example of electrical components. In case of a touch sensor for example, by increasing the input touch on a sensor of the sensor array 2, the signal intensity for that particular sensor is determined to be higher. In a similar way, in case of a temperature sensor, the increase of the input temperature measured on one sensor of the sensor array 2 would determine a signal generated by that sensor having a higher signal intensity.

[0021] It is noted that the embodiment shown in figure 1A represents a general concept for the smart connector 1 , wherein all its elements can be encapsulated into a skin material. However, there are different embodiments that can be taken into account depending on what material is used (i.e. a silicon thin layer which is embedding the fibers can embed also the micro-electronics). Different physical configurations of the single components due to the specific cases are intended to be included in the present disclosure.

[0022] In examples, the signal sources 9 are conductive fibers, (e.g. piezo-resistors and / or piezoelectric sensors) that are activated, i.e. that generate a sensor signal, when a tactile contact occurs between the sensor array 2 and an object. In this way, a high resolution tactility can be achieved.

[0023] The connector 1 comprises a conversion unit 4. This unit 4 receives different signals from the sensor array 2, i.e. from the plurality of signal sources 9. According to figure 1A, the conversion unit 4 receives at least sixteen signals. As mentioned above, each received signals has a signal intensity related to the input quantity applied on the corresponding signal source 9. The conversion unit 4 serves to convert the received signals into digital data signals. The digital data signals are then transferred to a controlling unit 5 that performs a first processing and data compression. Also, the controlling unit 5 carries out a selection of the received data signals. This selection is performed by a filtering module 11 configured to select a set of data signals based on the signal intensity. In other words, the controlling unit 5, and in particular the filtering module 11 , receives the digital data signals from the conversion unit 4 and compares the signal intensity of each signal with a predefined threshold. The predefined threshold is a reference value of the signal intensity. The signals having a signal intensity above (or equal to) said reference value are part of the selected set of data signals. Signals having a signal intensity below said reference value are not part of the selected set of data signals. Only the selected set of data signals are then forwarded to the processing unit 3 through a digital bus interface 6 connected to the controlling unit 5. It is noted that the filtering module 11 is integrated into the controlling unit 5. According to some examples, the filtering module 11 corresponds to the controlling unit 5. In other examples, the filtering module 11 corresponds to a combination of electronic components of said controlling unit 5.

[0024] In this way, it is possible to avoid unnecessary overload of the digital bus interface 6 that is connecting the sensor array 2 with the processing unit 3. In particular, dynamic and real-time thresholds can be defined which have to be exceeded, before a signal is transmitted to the processing unit 3. Accordingly, the signal sources 9 of the sensor array 2 are “silent” as long as they are not sensing any physical phenomenon, e.g. as long as they are not in contact with an object. Thanks to the present smart connector 1 , it is possible to filter any unnecessary signal that would have reached the processing unit 3, thereby reducing the signal load on said processing unit 3. It is noted that the present connector 1 is suitable for the employment of smart fabrics that comprise several signal sources 9. However, the connector 1 can advantageously be used in any case where a very high signal quantity is generated, essentially by any type of sensor array.

[0025] In examples, the connector 1 is at least in part integrable in the sensor array 2 and / or in the processing unit 3. As shown in figure 1 A, the smart connector 1 can be part of a device 12, for example a robotic device. The robotic device 12 can be a robotic hand and the sensor array 2 can cover at least in part a portion of a frame of the device 12. For example, the sensor array 2 can be part of a smart skin in the form of a glove covering the robotic hand. In this case, the processing unit 3 can be remotely connected to the device 12 through the smart connector 1 , in particular through the digital bus interface 6 (wireless or via cable). For a robotic hand that is constantly moving it is also possible to integrate energy harvesting and connect it to the smart connector 1 . In this case, additional components and a battery could be required.

[0026] In an alternative example, the smart connector 1 can be part of a distributed connector system (e.g., connectors network). The plurality of signals coming from sensor array 2 can be handled by two or more connectors 1 The signals coming from conductive fibers can be split in groups and sent to a few connectors 1 . Miniaturization may benefit from this. This may accelerate the process of filtering and conversion as well. Of course, the number of connectors 1 depends on how large the sensing area is or how many sensing units are present. One node of the network can be represented by a single smart connector 1 . Each connector 1 of the network can have an ID same with the processing unit 3 so it is possible to know from which area the signals are coming. Figure 1 B illustrates an example of a network 20 comprising three smart connectors 1 . Each connector 1 is dedicated to receive signals from different areas of the sensor array 2 and from a different number of signal sources 9. For example, the sensor array 2 comprises eight signal sources 9 that are conductive fibers. A first connector 1 receives the signals from three conductive fibers 9 located on the top of the sensor array 2, a second connector 1 receives a signal from a single conductive fiber 9 located in the center of the sensor array 2, and a third connector 1 receives the signals from four conductive fibers 9 located on the bottom of the sensor array 2. Accordingly, the signals can be forwarded to the processing unit 3. Thanks to the network 20, it is therefore possible to map the sensing on the sensor array 2 in a more reliable way.

[0027] In an alternative configuration, the processing unit 3 can be part of a system 13 together with the device 12, i.e. together with the sensor array 2 and the smart connector 1 . In this case, the system 13 can be a robotic system including a robotic hand, a processing unit 3, for example a CPU, managing at least the robotic hand, and a smart connector 1 connecting the robotic hand to the CPU. Other possible systems 13 can include all kinds of surfaces which are functionalized by smart fabrics. These can be the wings of wind turbines, which can be permanently observed and controlled for mechanical stress and impact. Another example are surfaces of seats (e.g. car seats), furniture or hospital beds, which sense the pressure of a human body and adapt surface properties like shape or softness to this pressure.

[0028] Figure 2 illustrates a possible electronic circuitry defining the smart connector 1. In examples, the conversion unit 4 comprises at least a parallel to serial converter 7 and at least an analogue-to-digital converter 8 connected to said parallel to serial converter 7. Alternatively, the analogue-to-digital converter 8 can be integrated into the parallel to serial converter 7. As shown in the figure, the sensor array 2 is arranged on a bidimensional surface, such as a smart fabric. The fabric has vertical threads and horizontal threads (i.e. conductive fibers). It is noted that the vertical threads are connected to a parallel to serial converter 7, whereas the horizonal threads are connected to a different parallel to serial converter 7. One end of the fabric is connected to the ground.

[0029] It is noted that figure 2 only illustrates a schematic representation of the connector 1 and the elements connected to the connector 1 , such as the sensor array 2 and the CPU 3. The sensor array 2 is shown as a grid just to give an idea on how the smart connector 1 can be coupled to a smart fabric. Since the skilled person is aware of the technical details of a smart fabric, these details are not repeated in the present disclosure. For example, the skilled person would understand that the ground of the fabric depends on how the fibers are arranged and that the fibers would not touch if they are placed in a grid as in the figure. If they touch, the fibers can be coated to be insulated so that at the intersection points, there is no signal interference, and the overall system is not losing the signal directly to the ground.

[0030] Specifically, the parallel to serial converter 7 is used to convert more than 1 ,000 input signals. The output of each parallel to serial converter 7 is connected to the input of an analogue- to-digital converter 8 and the output of each analogue-to-digital converter 8 is input to the controlling unit 5. The analogue-to-digital converter 8 are converting the incoming analogue signals to digital signals and uploads them to the controlling unit 5. In particular, the connector 1 can comprise a cyclic buffer 10 for storing the received digital data signals in the controlling unit 5. In particular, the complete x- and y-axis of the incoming digital signals can be stored in a circular buffer 10 of variable depth. Out of these signals the controlling unit 5 calculates the position and strength of the touch signal and compares it to the given threshold. The threshold can vary over time and the surface, as it is dynamically adjusted by the processing unit 3. The result is a three- dimensional matrix, where the x- and y-axis of the matrix is representing the touch area and the z-axis the strength. The format in which this matrix is transmitted to the processing unit 3 can be pre-programmed and thereby adjusted to the application. For example, vector representations are defined, which allow for some degree of data compression by omitting the areas of the surface which have no relevant touch signal.

[0031] In examples, only signals that pass the threshold are stored as considered “of interest”. The threshold can filter out noise signals for instance (vibrations are sensed by fibers for instance and may interfere with the tactile function). However, depending on the situation, at the beginning all data can be stored in the buffer.

[0032] It is pointed out that there could be different ways of integrating the elements described in figure 2 depending on the requirements of specific applications. The highest level of integration would be to integrate the parallel to serial converter 7 with the analogue-to-digital converter 8 and the controlling unit 5 into one piece of semiconductor, thereby achieving the most lightweight and robust solution. An alternative solution would be the implementation of each element in different, separate semiconductors, which are fulfilling the respective functions. In both cases the challenge is to achieve a good scaling capability, because the amount of incoming analogue signals from the smart fabric can vary from small pieces of fabric (e.g. the fingertip of a robotic hand) to big areas, like the complete surface of a wind turbine. Under all circumstances, the length of the analogue lines between the smart fabric and the input connectors of parallel-to-serial converter 7 must be short (a couple of few centimeters) in order to limit disturbing effects (e.g. EMC, drop of signal strength etc.). Additionally, the input connector of parallel to serial converter 7 must be able to handle the high amount of thin input lines of the smart fabric in a robust way and provide a good and stable galvanic connection.

[0033] In examples, the signal sources 9 are conductive fibers, in particular piezo-resistors and / or piezoelectric elements. In this way, any pressure or strain variation at the signal sources 9 can be detected and (directly or indirectly) converted into an electrical signal. The intensity of the signal can vary based on the quantity of the pressure or strain detected.

[0034] In order to get a real time behavior in response to any sensor signal, the signals received from the plurality of signal sources 9 can have a latency time comprised between 100 and 400 msec.

[0035] In examples, the controlling unit 5 is a microcontroller configured to be reprogrammable in the field. In this way, the predefined threshold can be set in advance or modified on real time based on external conditions. It is noted that in robotics the applications of the robot are changing all the time. Therefore, a fixed and preprogrammed data processing and compression is not useful. Hence, the present controlling unit 5 can be reprogrammed if / when needed by the application.

[0036] The controlling unit 5 can be reprogrammed using the artificial intelligence (Al). Algorithms can be designed to learn based on experience and self-reprogram the threshold. This means that the smart skin starts learning based on training data and then progressively adds new experiences and adjusts the threshold based on these experiences. At the beginning there is no experience, and it is hard to set a threshold. The “calibration” process can be part of Al algorithms that are initially trained to recognize pressure sensing from vibration sensing. Different experiences will also recognize different signal sources (temperature, pressure etc.). The setting of a threshold will be different according to different sensing sources. For this reason it would be better to separate signal sources 9 (i.e. conductive fibers) for each sensing function.

[0037] In examples, the digital bus interface 6 is configured to allow a two way communication between the controlling unit 5 and the processing unit 3. In particular, the digital bus interface 6 can be a single pair ethernet (SPE) bus. Any type of bus connection between the controlling unit 5 and the processing unit 3 can be used based on the particular technical application.

[0038] The schematic circuit in figure 2 is related to the use of piezoresistive fibers as signal sources 9. Accordingly, the physical change of the sensor (under pressure or strain) brings a change of the electrical resistance, thereby meaning that additional components are needed, for instance a Wheatstone bridge circuit, to convert small changes in resistance to an output voltage. It is noted that in the case of piezoelectric fibers there is no need of extra components between the parallel to serial converter 7 and the ADC elements 8. However, a voltage source can still be useful since the piezoelectric fibers can act as actuators. In this case, the presence of a bus interface 6 having a two way communication can be used to send signals from the CPU 3 to control the voltage source so that an electric impulse can be sent to the fibers and stretch them for instance or can activate heating.

[0039] As mentioned above, the sensor array 2 can be a smart textile, i.e. it can be represented by an homogenous distribution of sensors (e.g. signal sources 9) on a fabric material. However, the sensor array 2 can alternatively be represented by a plurality of patches of smart areas 14 as illustrated in figure 3. Each smart area 14 is a sensing region representing a signal source 9 (or a group of identical signal sources 9). Figure 3 shows a plurality of smart areas 14 having all the same squared shape grouped in two main sizes. However, the smart areas 14 can be of different shapes and / or different sizes and can be arranged on a substrate 15 (for example a flexible substrate) according to a precise pattern or randomly. Specifically, the smart areas 14 can be made of conductive fibers 16, but represent areas of the substrate 15 that are sensing. It means that a first portion 17 of the length of the conductive fibers 16 is sensing, whereas a second portion 18 of the length of the conductive fibers 16 is not sensing. The smart areas 14 are connected to the conversion unit 4 of the smart connector 1 , as described for example in figure 1 A, through dedicated connector elements 19. One end of the substrate 15 is connected to the ground.

[0040] Similarly to figure 2, also the sensor array 2 of figure 3 is schematically illustrated. Although not shown in detail, from each smart area (patch) 14 one filament needs to be conducted to ground (can be gathered with all ground filaments as well) and another would carry the signal. Also, the ground line would not touch the smart areas 14 (patches).

[0041] It is noted that a sensor array 2 as shown in figure 3 can be coupled with a single smart connector 1 or with a connector system, i.e. a network of connectors as described above in figure 1 B. This can be extremely useful for mapping the sensing.

[0042] As shown in figure 4, a method 100 for transmitting data from a sensor array 2 to a processing unit 3 is provided.

[0043] At step S101 , signals are received from a plurality of signal sources 9. For example, signals are received from a plurality of piezo-resistors (or conductive fibers in general) arranged on a smart fabric. However, different types of sensors can be used for any “smart fabric”, with the condition that many output signals are to be treated at the same time. As a matter of fact, the smart fabric might be based on different coating processes, which might make the fabric piezo resistive, but also piezo electric. Depending on the chemical process of producing such smart fabric, it will become touch sensitive, temperature sensitive or even sensitive for electro-magnetic waves or chemical substances. As already mentioned, each received signal has a signal intensity based on the nature of the sensor and on the quantity of the input on said sensor. At step S102, digital data signals are generated at the conversion unit 4. The parallel to serial converter 7 has an high amount of input lines (several thousands) to collect all incoming parallel signals.

[0044] At step S103, digital data signals are received by the conversion unit 4, i.e. by the analogue-to-digital converter 8, and a first processing and data compression is performed.

[0045] At step S104, a set of data signals are selected based on the signal intensity of the signals received from the plurality of signal sources 9. In particular, the selection is related to a comparison between the signal intensity and a reference value (threshold value). If the signal intensity is above (or equal to) the reference value, then signal is selected, i.e. it is considered part of the set of data signals and it is forwarded to the processing unit 3 (at step S105). On the other hand, no data signal is forwarded to the processing unit 3, if the signal intensity of the signals received from the plurality of signal sources 9 is below the predetermined threshold (at step S106). In this way, the processing unit 3 shall only be informed for a subsequent data processing when relevant mechanical input is received by the smart fabric, e.g. the “hand” of the robot is touching an object.

[0046] It is noted that the predetermined threshold can be varied and adapted based on the application. For example, in case of piezo-resistors arranged on a smart fabric, the threshold is set to determine any single touch on the fabric. In other words, the reference value can be Po (zero intensity) so that only signals are selected that are generated by the slightest pressure variation on the fabric. On the other hand, the reference value can be Pi>Po (greater than zero) so that only signals are selected that are generated by a certain value of pressure exerted on the fabric. Signals generated by lower values of pressures are neglected and not sent to the processing unit 3. In a similar way, in case of other type of sensors, such as temperature sensors, the reference value can be Toor Ti>T0or TKT0SO that only signals are selected that are generated by a predefined temperature variation.

[0047] The value of the threshold can therefore determine the response of the entire system. This response can be adjusted or adapted based on the characteristics of the processing unit 3 and / or based on the specific technical application, i.e., the tactile application of a robotic hand covered by a smart skin or the monitoring of an object covered with a smart fabric.

[0048] Accordingly, as already mentioned, it is advantageous that the bus-connection between the controlling unit 5 and the processing unit 3 allows for two way communication, making it possible to reprogram the controlling unit 5 in the field. In this way, the data compression can be modified according to the applications. It is noted that the two way communication can be advantageously employed for a processing unit 5 using Al algorithms. In case of flexible piezoelectric fibers used as actuators, the smart skin (i.e. the sensor array 2) can be considered both passive and also active. The processing unit 3 can send commands back to the smart skin after the processing of the sensing data (in this case, a heating function can be added, so that the conductive fibers can heat the sensor array 2 and therefore transfer heat to an object that is in contact). In this case, a parallel model is needed. When the command goes back to the signal sources 9 from the processing unit 3 through the digital bus interface 6, it does not need a filter but the controlling unit 5 has the task to control a power supply to send “adjusted” electrical current in order to heat the conductive fibers up (the current needs to be controlled to control the temperature).

[0049] In examples, the method 100 further comprises transferring information data from the processing unit 3 to the sensor array 2 through the digital bus interface 6 (S107 in figure 4). In particular, the processing unit 3 can transfer data to one or more signal sources 9 of the sensor array 2 in response to digital data signals received at the processing unit 3. This step is optional and can be carried out by employing a two-way communication.

[0050] In examples, the method 100 comprises defining the predetermined threshold by the processing unit 3 before the selected set of data signals are forwarded to the processing unit 3. In this way, the signal transmission to the processing unit 3 can be controlled and the signal load from the sensor array 2 to the processing unit 3 can be reduced.

[0051] According to an example, the digital data signals received at the controlling unit 5 are first processed and data compressed based at least on the subsequent second processing at the processing unit 3. In other words, the selection of the digital data signals can be adapted based on the type of data processing that will be carried out in the processing unit 3. In particular, the digital data signals received at the controlling unit 5 are first processed and data compressed based also on an application software analyzing the signals originating from signal sources 9. For example, the processing and data compression of the digital data signals depends on the software which is dealing with the “touch signal” of the smart fabric. As a matter of fact, it is important to have the capability to adapt the threshold over time and the area of the smart skin. It is obviously a difference whether a robot is picking up a small and sensitive item like an egg as compared to a heavy and robust item like a hammer.

[0052] In examples, the method 100 further comprises generating vector data representing the intensity and the position on the sensor array 2 of each signal generated by the signal sources 9. In particular, a “vector graphic” representation can be envisaged, where the controlling unit 5 sends a vector to the processing unit 3 describing where and at which strength the touch-signal was sensed. It is noted that any other data representation shall be possible depending on the application. By employing data representations which are used in video systems, like e.g. the vector graphics used in computer games it is possible to achieve a sensor fusion between the tactile and video inputs for example of a robot. According to the present method, it is possible to achieve a high level of data compression from the thousands of incoming lines of the smart fabric to e.g. a vector informing the processing unit 3 that a surface is being touched with the tip of finger with a determined pressure generating a signal with a certain intensity.

[0053] Depending on the application, a complete robotic device or parts of it shall be integrated into the smart connector 1 for achieving a compact and light weight implementation.

[0054] The present connector 1 can be advantageously used when a huge amount of input signals needs to be analyzed by a processing unit 3. Specifically, the present connector 1 is suitable for the application of smart fabrics. In particular, the present connector 1 is to be located between the sensor array 2 (i.e. the smart fabric) and a processing unit 3 in order to determine a signal filtering. In this way, the processing unit 3 is not overloaded with thousands of information data and the signals elaboration results to be quicker and less prone to errors. Also, by using the present connector 1 , it is possible to achieve a dynamic data compression by programmable thresholds integrated into the connector 1 .

[0055] In addition, the present connector 1 is configured to connect a sensor array 2 to a processing unit 3, thereby making any device 12 and / or system 13 including said connector 1 more compact and efficient. This is extremely useful in the robotics field and in any technological application where space is an issue and miniaturization is required.

[0056] As already mentioned, smart fabrics can be used in almost all areas of mechanical engineering in the future. For example, the wings of windmills are subject to mechanical stress and impacts. Currently, a complex and expensive inspection by maintenance staff is need in order to assess whether significant damage has been inflicted. When the technical object (here the wing) is covered with a smart fabric, the information about the impact and the extent of possible damage is available immediately. Therefore, maintenance activities can be targeted more efficiently and greater consequential damage can be prevented by e.g. stopping the wind turbine. Therefore, the present connector 1 and method 100 can be advantageously used in fields of technologies other than robotics.

[0057] Although a variety of techniques and examples of such techniques have been described herein, these are provided by way of example only and many variations and modifications on such examples will be apparent to the skilled person and fall within the spirit and scope of the present invention, which is defined by the appended claims and their equivalents.

Claims

CLAIMS1 . Smart connector (1 ) for the connection of a sensor array (2), to a processing unit (3), the connector (1 ) comprising: a conversion unit (4) for receiving signals from a plurality of signal sources (9), each received signal having a signal intensity, and generating digital data signals; a controlling unit (5) for receiving the digital data signals and performing a first processing and data compression, the controlling unit (5) comprising a filtering module (11 ) to select a set of data signals based on the signal intensity of the signals received from the plurality of signal sources (9), wherein the selected set of data signals comprises only digital data signals originating from signal sources (9) that generated a signal having a signal intensity above a predetermined threshold; and a digital bus interface (6) connected to the controlling unit (5) for forwarding the selected set of data signals to the processing unit (3) for a second processing.

2. Smart connector (1 ) of claim 1 , wherein the conversion unit (4) comprises at least a parallel to serial converter (7) and at least an analogue-to-digital converter (8) connected to said parallel to serial converter (7), wherein in particular the parallel to serial converter (7) is used to convert more than 1 ,000 input signals.

3. Smart connector (1 ) of any one of claims 1 to 2, wherein the signal sources (9) are conductive fibers, in particular piezo-resistors and / or piezo electric elements.

4. Smart connector (1 ) of any one of claims 1 to 3, wherein the signals received from the plurality of signal sources (9) have a latency time comprised between 100 and 400 msec.

5. Smart connector (1 ) of any one of claims 1 to 4, further comprising a cyclic buffer (10) for storing the received digital data signals in the controlling unit (5).

6. Smart connector (1 ) of any one of claims 1 to 5, wherein a. the controlling unit (5) is a microcontroller configured to be reprogrammable in the field; and / orb. the digital bus interface (6) is configured to allow a two way communication between the controlling unit (5) and the processing unit (3); and / or c. the digital bus interface (6) is a single pair ethernet, SPE, bus.

7. Smart connector (1 ) of any one of claims 1 to 6, wherein the connector (1 ) is at least in part integrable in the sensor array (2) and / or in the processing unit (3).

8. Network (20) comprising a plurality of smart connectors (1 ) according to any one of claims 1 to 7, wherein each smart connector (1) of the network (20) is configured to connect one or more signal sources (9) of the sensor array (2) to a processing unit (3).

9. Device (12), in particular a robotic device, comprising: a sensor array (2) including a plurality of signal sources (9), the sensor array (2) covering at least in part a portion of a frame of the device (12); and at least a smart connector (1 ) according to any one of claims 1 to 7 or a network (20) according to claim 8 for the connection of the sensor array (2) to a processing unit (3).

10. System (13), in particular a robotic system, comprising: the device (12) according to claim 9; and a processing unit (3) connected to the device (12) through the smart connector (1 ).11 . Method (100) for transmitting data between a sensor array (2) and a processing unit (3), the method (100) comprising: receiving (S101 ) signals from a plurality of signal sources (9), each received signal having a signal intensity; generating (S102) digital data signals; receiving the digital data signals and performing (S103) a first processing and data compression; selecting (S104) a set of data signals based on the signal intensity of the signals received from the plurality of signal sources (9), wherein the selected set of data signals comprises only digital data signals originating from signal sources (9) that generated a signal having a signal intensity above a predetermined threshold; and forwarding (S105) the selected set of data signals to the processing unit (3) for a second processing.

12. Method (100) according to claim 11 , wherein no data signal is forwarded (S106) to the processing unit (3) if the signal intensity of the signals received from the plurality of signal sources (9) is below the predetermined threshold.

13. Method (100) according to any one of claims 11 to 12, further comprising transferring (S107) information data from the processing unit (3) to the sensor array (2) through a digital bus interface (6).

14. Method (100) according to any one of claims 11 to 13, further comprising defining the predetermined threshold by the processing unit (3) before the selected set of data signals are forwarded to the processing unit (3).

15. Method (100) according to any one of claims 11 to 14, wherein the digital data signals received at the controlling unit (5) are first processed and data compressed based at least on the subsequent second processing at the processing unit (3), wherein in particular the digital data signals received at the controlling unit (5) are first processed and data compressed based also on an application software analyzing the signals originating from signal sources (9).

16. Method (100) according to any one of claims 11 to 15, further comprising generating vector data representing the intensity and the position on the sensor array (2) of each signal generated by the signal sources (9).

Images