Floor type detection

Abstract

A floor type detection system for a floor cleaner, wherein the floor type detection system comprises one or more light emitters, a light sensor configured to detect light having infrared wavelengths i

Description

B ACKGROUND Floor cleaners with motor driven rotating sweeping and / or mopping rollers are becoming increasingly popular. It has been found that the cleaning performance and energy efficiency of such floor cleaners can be improved by optimising the operating parameters of the floor cleaner depending on the type of floor surface to be cleaned. Some floor cleaners have sensors for identifying floor type. It is against this background that the present invention has been developed. SUMMARY The present invention provides a floor type detection system for a floor cleaner, wherein the floor type detection system comprises: one or more light emitters; a light sensor configured to detect light having wavelengths in a reference range of HOOnm to 3000nm and output one or more sensor signals indicative of the intensity of the light at a plurality of reference wavelengths within the reference range; and one or more processors collectively configured to: receive the one or more sensor signals; and determine a floor type in dependence on the one or more sensor signals. The present invention is advantageous as the floor type to be cleaned may be determined thereby allowing the floor cleaner to be operated to suit the floor type detected and from both a cleaning efficacy and energy efficiency perspective. Optionally the light sensor may be configured to detect light having wavelengths in a reference range of 1 lOOnm to 2500nm, or !400nm to 2500nm. Sensors able to detect light in these ranges are economically preferable. At least one of the reference wavelengths may optionally be in the range: 1440nm to 1460nm; or 1900nm to 2000nm. These ranges are beneficial for the detection of surface water. At least one of the reference wavelengths may be in the range 1440nm to 1470nm or 1840nm to 1870nm. These ranges are beneficial for the detection of laminate. At least one of the reference wavelengths may be in the range 1640nm to 1670nm or 1940nm to 1970nm. These ranges are beneficial for the detection of vinyl and wood respectively. In one example the plurality of reference wavelengths may comprise at least two, at least three or at least four different wavelengths. This is beneficial as it provides more data points for use in the determination of floor type to improve the accuracy of floor detection. Optionally at least three of the reference wavelengths are above 1400nm which is beneficial as it provides more data points at larger wavelengths which can improve accuracy of floor detection. The floor type detection system my optionally comprise: a visible light detector configured to detect light of a visible wavelength and output a visible light signal indicative of the intensity of the light of the visible wavelength, wherein the visible wavelength is in the range 400 to 750nm; and / or a near infra-red (NIR) light detector configured to detect light of an NIR wavelength and output an NIR signal indicative of the intensity of the light of the NIR wavelength, wherein the NIR wavelength is in the range 751 to 1099nm, wherein the one or more processors are configured to: receive the visible and / or NIR signals; and determine a floor type in dependence on the visible and / or NIR signals. In some circumstances, for example when a floor is wet, it is beneficial to obtain data points from a wider range of the light spectrum to improve accuracy of floor detection. Determining the floor type may comprise: using the sensor signals indicative of the intensity of the light at a plurality of reference wavelengths to determine a corresponding plurality of data points; comparing the plurality of data points with reference characteristic floor data; and determining the floor type in dependence on a correlation between the plurality of data points and the reference characteristic floor data. Comparison of detected data points to reference characteristic floor data has been found to provide a good level of accuracy for floor type detection. Optionally the reference characteristic floor data may comprise a continuous trace, or an array of discrete values, indicative of: absolute reflected light intensity across a spectrum of wavelengths; normalised reflected light intensity across a spectrum of wavelengths; or scaled reflected light intensity across a spectrum of wavelengths. The floor type detection system may comprise: determining a floor identifier in dependence on the correlation between the plurality of data points and the reference characteristic floor data; and determining a floor type in dependence on the floor identifier. This is advantageous as it facilitates correlation between a readily detectable feature of the floor (such as colour or surface finish) with an associated floor type (material). This helps to improve the accuracy of floor type detection. In another aspect the present invention provides a control system for controlling a floor cleaner, the control system comprising the floor type detection system described above, wherein the one or more processors are configured to: determine one or more operating parameters of the floor cleaner in dependence on the floor type or floor identifier; and output at least one control signal comprising one or more instructions to operate the floor cleaner in accordance with the one or more operating parameters. This is advantageous as it enables the floor cleaner to be operated to suit the floor type to be cleaned from a cleaning efficacy and energy efficiency perspective. Optionally the one or more processors may be configured to: output a detected floor type signal; and receive a user input signal indicative of: acceptance of the detected floor type; or rejection of the detected floor type, wherein if the user input signal is indicative of acceptance of the detected floor type, the control system is configured to: determine one or more operating parameters of the floor cleaner in dependence on the detected floor type. This is advantageous as it allows the floor type determined by the floor type detection system to be corroborated by a user to ensure that the operating parameters of the floor cleaner are appropriate for the floor to be cleaned from a cleaning efficacy and efficiency perspective. In one example, if the user input signal is indicative of rejection of the detected floor type, the control system may be configured to: receive a user-identified floor type input signal indicative of a user-identified floor type; determine one or more operating parameters of the floor cleaner in dependence on the user-identified floor type input signal. This is advantageous as it allows the floor type determined by the floor type detection system to be rejected if incorrect so that the operating parameters of the floor cleaner will be appropriate even if the detected floor type is incorrect. Optionally the one or more processors may be configured to: associate the one or more sensor signals with the floor type identified by the user; and issue an instruction to store the association between the one or more sensor signals and the floor type identified by the user. This is advantageous as the various detected light intensities at the various different wavelengths can be stored for future use, for example, to train the floor detection system. Optionally the one or more operating parameters may comprise: a first speed of rotation for a roller or agitator element of the floor cleaner; and / or a level of hydration to be applied to the floor to be cleaned and / or to a roller or agitator of the floor cleaner. The one or more processors may optionally be configured to: determine that surface water is present if the one or more sensor signals indicate that the intensity of the light at the least one of reference wavelength in the range 1440nm to 1460nm or 1900nm to 2000nm is below a predetermined threshold; and if it is determined that surface water is present: determine a second speed of rotation that is greater than first speed of rotation; or reduce the level of hydration applied to the mopping rollers; or reduce the level of hydration applied to the floor. This advantageously allows the amount of water present of the surface of the floor to be removed efficiently. In a further aspect the present invention provides a floor cleaner comprising the floor type detection system or the control system described above. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 shows a perspective view of a vacuum floor cleaner; Figure 2 shows a perspective view of a wet floor cleaner; Figure 3 shows a schematic plan view of the underside of the cleaner head of the floor cleaner of Figure 2; Figures 4a and 4b show schematic side views of the cleaner head of Figure 3; Figure 5 shows a schematic view of a floor type detection system; Figure 6a shows a schematic graphical representation of reflected light intensity variation with wavelength for different ceramic floors in the visible and near infrared spectral range; Figure 6b shows a schematic graphical representation of reflected light intensity variation with wavelength for different vinyl floors in the visible and near infrared spectral range; Figure 6c shows a schematic graphical representation of reflected light intensity variation with wavelength for different wood floors in the visible and near infrared spectral range; Figure 6d shows a schematic graphical representation of reflected light intensity variation with wavelength for a laminate floor in the visible and near infrared spectral range; Figure 7 shows a schematic view of a light sensor; Figures 8a to 8d show the schematic graphical representations of reflected light intensity variation with floor type and wavelength as shown in Figures 6a to 6d; Figure 9 shows a schematic view of another floor type detection system; Figures 10a to lOd show schematic graphical representations of reflected light intensity variation with wavelength for four different materials; Figures Ila to 11c show schematic representations of reflected light intensity variation with wavelength for surface water; Figure 12 shows test data comparisons for several different floor type sensor configurations. Figure 13 shows a schematic representation of a control system; and Figure 14 shows a flowchart describing a process of controlling a floor cleaner. DETAILED DESCRIPTION Figure 1 shows a vacuum floor cleaner 10 according to an embodiment of the disclosure comprising a main body 12, a wand 14 and a cleaner head 16. The main body 12 comprises a separating system 13 in the form of a cyclonic separator, a motor and impeller (not visible) arranged to draw air through the separating system 13, and a power supply, in the form of a battery 11, for powering the motor. The wand 14 is attached at one end to the main body 12 and at the other end to the cleaner head 16. The wand 14 provides fluid communication between the cleaner head 16 and the separating system 13 within the main body 12 and supports the cleaner head 16 during use. The cleaner head comprises a brush bar roller 17 for sweeping the floor surface to be cleaned during use. The brush bar roller 17 is configured to rotate in use to agitate dust and debris on the floor surface to be cleaned so that it can be drawn into the separation system 13 for separation and disposal. In this example, the cleaner head 16 comprises one brush bar roller 17 located within a housing 18 of the cleaner head 16. However, in another example, the cleaner head 16 may comprise two or more brush bar rollers arranged one in front of the other and / or side by side. A floor type sensor 50 (as will be described in greater detail below) is located in the housing 18. Figure 2 shows a wet floor cleaner 20 according to an embodiment of the disclosure comprising a main body 22, a handle 24 and a cleaner head 26. The main body 22 comprises a water distribution system having a clean water tank 21, a waste water tank 23 and corresponding clean and waste water pumps (not shown). The water distribution system is arranged to deliver clean water to a floor surface to be cleaned and / or to one or more mopping rollers 30, and to remove soiled water from a debris collection tray 32 (Figure 3) and / or from mopping rollers 30. The handle 24 is attached at one end to the main body 22. An electronic user interface 25 in the form of a touch screen is located at an upper end of the handle 24. The cleaner head 26 of the wet floor cleaner 20 comprises a pair of mopping rollers 30 for cleaning the floor surface during use. Figure 3 shows a schematic plan view of the underside of the cleaner head 26 of the wet floor cleaner 20, and Figures 4a and 4b show schematic side views of the cleaner head 26. The cleaner head 26 comprises a housing 28 within which are located two mopping rollers 30. The mopping rollers 30 are supported for rotation within the housing 28 and spaced apart such that each roller 30 is located proximate the front 31 and rear 32 of the housing 28 (where the front 31 of the housing 28 corresponds to a leading side of the housing 28 when a user pushes the wet floor cleaner 20 forward in use, and where the rear 32 of the housing 28 corresponds to a leading side of the housing 28 when a user pulls the wet floor cleaner 20 back in use). As best shown in Figures 4a and 4b, resilient scrapers 36 are located proximate each mopping roller 30. The scrapers 36 are located inboard of the mopping rollers 30 and are configured to scrape debris and water from the floor surface 29 during cleaning. A removeable debris tray 33 for collecting water and debris is located between the mopping rollers 30. The debris tray 33 extends from a second side 35 of the housing 28 towards a first side 34 of the housing 28. A floor type sensor 50 (as will be described in greater detail below) is located proximate the first side 35 of the housing 28 between the rollers 30 and the scrapers 36. It will be understood that the scrapers 36 are not essential and may be omitted. In this example, a lower surface 37 of the floor type sensor 50 is co-planar with the lowermost surface 38 of the debris tray 33 which is arranged to be located directly above the floor surface 29 during use. In this example, the distance 39 (see Figure 4b) between the lowermost surface 38 of the debris tray 33 (and therefore also the lowermost surface 37 of the floor type sensor 50) and the floor surface 29 in use is 15mm. However, the distance 39 between the lowermost surface 38 of the sensor 50 and the floor surface 29 in use may be any suitable size in the range 3mm to 100mm, alternatively in the range 5mm to 30mm, or alternatively again in the range 10mm to 20mm. In use, the distance between the lower surface of the sensor 50 and the floor surface 29 is 15mm in this example. A distance between the lower surface of the sensor 50 and the floor surface 29 of between around 10mm and 20mm has been found to provide improved performance. Reducing the distance to less than 10mm may result in a drop in performance due to decreased field of view of the sensor 50, which may result in a smaller area of the floor being sampled and less accurate results. Increasing the distance to above 20mm may be possible, for example up to around 100mm or even up to around 350mm, but this may require additional optical features to prevent a drop in performance. Preferably the distance between the lower surface of the sensor 50 and the floor surface 29 is less than around 150mm. Figure 5 shows a schematic representation of a floor type detection system 70 comprising a floor type sensor 50. In this example, the floor type sensor 50 comprises a pair of light emitters 52, a first light detector 54 and a second light detector 56. In this example, the light emitters 52 comprise light emitting diodes (LED) configured to emit light 55 having wavelengths spanning the visible and infrared spectrums from 440nm to 950mn. However, any suitable LED or other light source capable of emitting light spanning the visible and infrared light spectrums may be used. In use, a portion of the light 55 emitted by the light emitters 52 is reflected from the floor surface 29 to be cleaned. The first light detector 54 is configured to detect reflected visible light 57 of a specific wavelength in the visible light spectrum, and the second light detector 56 is configured to detect reflected light 58 of a specific wavelength in the near infrared (NIR) spectrum. In this example, the first light detector 54 is configured to detect visible light 57 having a wavelength of 555nm, and the second light detector 56 is configured to detect NIR light 58 having a wavelength of 910nm. It will be understood that the above examples are illustrative only, and that the first light detector 54 may be configured to detect visible light 57 of a specific wavelength in the range 400 to 750nm, and the second light detector 56 may be configured to detect light 58 of a specific wavelength in the range 751 to 1 lOOnm. The first and second light detectors 54, 56 are configured to output first and second signals 64, 66 respectively, where the first signal 64 is indicative of the intensity of the reflected visible light 57, and the second signal 66 is indicative of the intensity of the reflected NIR light 58. The first and second signals are received by a processor 110 of a control system 100. The processor 110 is configured to determine a floor type or floor identifier in dependence on the first and second signals 64, 66, and to determine one or more operating parameters of the wet floor cleaner 20 and to output at least one control signal 102 comprising one or more instructions to operate the wet floor cleaner 20 in accordance with the one or more operating parameters as will be described in greater detail below. Figures 6a to 6d show schematic graphical representations of characteristic reflected light intensity variation with wavelength for different ceramic, vinyl, wood and laminate floors respectively in the visible and near infrared spectral range. As shown in Figures 6a to 6d, the intensity of the reflected light varies across the spectrum and varies depending on floor material type, colour and surface finish. The variation of the reflected light intensity across the spectrum is a characteristic of the floor which may be used to identify the floor type or a floor identifier associated with a particular floor such as colour or surface finish. Referring to Figure 6a, traces 150a, 150b show the characteristic reflected light variation traces for two different ceramic floors, referring to Figure 6b, traces 160a, 160b show the characteristic reflected light variation traces for two different vinyl floors, referring to Figure 6c, traces 170a, 170b show the characteristic reflected light variation traces for two different wooden floors, and referring to Figure 6d, trace 180a shows the characteristic reflected light variation traces for a laminate floor. The floor type sensor 50 described above is able to detect light in the visible spectrum having a wavelength of 555nm - indicated by line A in Figures 6a and 6b, and to detect light in the NIR spectrum having a wavelength of 910nm - indicated by line B in Figures 6a and 6b. If the first detector 54 detects reflected visible light 57 of intensity xl, and the second detector detects reflected NIR light 58 of intensity yl, this corresponds to the ceramic floor having the characteristic trace 150b. Alternatively, if the first detector 54 detects reflected visible light 57 of intensity x2, and the second detector detects reflected NIR light 58 of intensity y2, this corresponds to the vinyl floor having the characteristic trace 160a. Consequently, the reflected light intensity signals 64, 66 received by the processor 110 from the detectors 54, 56 can be used by the processor 110 to identify the floor type, or a floor identifier such as colour, by comparing the reflected light intensities with known reference characteristic reflected light intensity traces for different types / colours of floor. Alternatively, the floor type or floor identifier may be identified by implementing machine learning algorithms which are able to identify a floor type or floor identifier in dependence on the reflected light intensity signals 64, 66. Figure 7 shows a schematic view of an alternative floor type sensor 250 that may be used in the floor type detection system 70. In this example, the floor type sensor 250 comprises a light emitter 252, an array 253 of first light detectors 254, and a second light detector 256. In this example, the light emitter 252 comprises a LED configured to emit light having wavelengths spanning the visible and infrared spectrums from 400nm to HOOmn. However, any suitable LED or other light source capable of emitting light spanning the visible and infrared light spectrums may be used. As above, in use, a portion of the light emitted by the light emitter 252 is reflected from the floor surface to be cleaned. The array 253 of first light detectors 254 is configured to detect reflected visible light of a plurality of different specific wavelengths in the visible light spectrum, and the second light detector 256 is configured to detect reflected light of a specific wavelength in the NIR spectrum. In this example, the floor type sensor 250 comprises an array 253 of sixteen first light detectors 254 configured to detect visible light having wavelengths of: 415nm, 445nm, 480nm, 515nm, 555nm, 590nm, 630nm, 680nm, and a second light detector 256 is configured to detect NIR light having a wavelength of 910nm. Each of the first light detectors 254 in the array 253 is configured to detect light of a specific wavelength, with two first light detectors 254 being provided per specific wavelength. However, this is not essential, and in a different example, only one first light detector 254 in the array is provided per specific wavelength. In another example, more than two first light detectors 254 may be provided in the array 253 per specific wavelength. Each light detector 254 in the array 253 is configured to output a signal 264 indicative of the intensity of light at one of the plurality of visible wavelengths, and the second light detector 256 is configured to output a second signal 266 indicative of the intensity of the reflected NIR light. It will be understood that for clarity only one of the signals 264 is illustrated in Figure 7. It will be understood that the above examples are illustrative only, and that the first light detectors 254 in the array 253 may be configured to detect visible light of any suitable number of specific wavelengths in the range 400 to 750nm, and the second light detector 256 may be configured to detect light of a specific wavelength in the range 751 to 1 lOOnm. In another example, more than one second light detector 256 may be provided, each configured to detect a specific wavelength in the range 751nm to 1 lOOnm, and each configured to output a signal 266 indicative of the intensity of the specific wavelength of reflected light in the range 751nm to 1 lOOnm. The plurality of signals 264, and the second signal 266, are received by the processor 110 (Figure 5) of the control system 100. The processor 110 is configured to determine a floor type or floor identifier in dependence on the plurality of signals 264, and the second signal 266, and to determine one or more operating parameters of the wet floor cleaner 20 and to output at least one control signal 102 comprising one or more instruction to operate the wet floor cleaner 20 in accordance with the one or more operating parameters as will be described in greater detail below. Figures 8a to 8d show the same schematic graphical representation of characteristic reflected light intensity variation with floor type and wavelength as in Figures 6a to 6d. The floor type sensor 250 described above is able to detect light in the visible spectrum having wavelengths of 415nm, 445nm, 480nm, 515nm, 555nm, 590nm, 630nm, 680nm -indicated by lines Al to A8, and to detect light in the NIR spectrum having a wavelength of 910nm - indicated by line B. Consequently, there are additional data points available in the visible light portion of the spectrum than for sensor 50 described above allowing increased accuracy of floor type or floor characteristic identification. The data points to identify the vinyl floor corresponding to characteristic trace 160a are shown on Figure 8b for illustrative purposes. The reflected light intensity signals 264, 266 received by the processor 110 can be used by the processor to identify the floor type or floor identifier by comparing the reflected light intensity data points with reference characteristic reflected light traces for different types of floor. Alternatively, the floor type or floor identifier may be identified by implementing machine learning algorithms which are able to identify floor type or floor identifier in dependence on the reflected light intensity signals 264, 266. It will be understood that although Figures 6a to 6d, and Figures 8a to 8d, show scaled reflected light intensity traces, this is illustrative only and absolute or normalised (or otherwise pre-processed) reflected light intensity data may equally well be used. The reflected light intensity data need not be a continuous trace and may rather be an array of data points. Figure 9 shows a schematic representation of another floor type detection system 370 comprising a floor type sensor 350. In this example, the floor type sensor 350 comprises a pair of light emitters 352 and light detectors 354a, 354b, 354c, 354d. In this example, the light emitters 352 comprise incandescent light bulbs. However, any suitable light source capable of emitting a spectrum of short wave infrared light (SWIR) may be used, where SWIR light includes light having wavelengths in the range 1100 to 2500nm. In some embodiments, only a single light emitter 352 is provided. In use, a portion of the light 355 emitted by the light emitters 352 is reflected from the floor surface 29 to be cleaned. Each light detector 354a, 354b, 354c, 354d is configured to detect reflected SWIR light 357 of a specific wavelength. In this example, the light detector 354a is configured to detect SWIR light 357 having a wavelength of 1450nm, the light detector 354b is configured to detect SWIR light 357 having a wavelength of 1700nm, the light detector 354c is configured to detect SWIR light 357 having a wavelength of 1920nm, and the light detector 354d is configured to detect SWIR light 357 having a wavelength of 2150nm. It will be understood that the above examples are illustrative only, and that the light detectors 354a, 354b, 354c, 354d may be configured to detect SWIR light 357 of specific wavelengths in the range 1100 to 2500nm. The light detectors 354a, 354b, 354c, 354d are each configured to output signals 364a, 364b, 364c, 364d respectively, where the signal 364a is indicative of the intensity of the reflected SWIR light 357 at 1450nm, the signal 364b is indicative of the intensity of the reflected SWIR light 357 at 1700nm, the signal 364c is indicative of the intensity of the reflected SWIR light 357 at 1920nm, and the signal 364d is indicative of the intensity of the reflected SWIR light 357 at 2150nm. The signals 364a, 364b, 364c, 364d are received by the processor 110 of the control system 100. The processor 110 is configured to determine a floor type or floor identifier in dependence on the signals 364a, 364b, 364c, 364d, and to determine one or more operating parameters of the wet floor cleaner 20 and to output at least one control signal 102 comprising one or more instructions to operate the wet floor cleaner 20 in accordance with one or more operating parameters as will be described in greater detail below. Figures 10a to lOd show schematic graphical representations of characteristic reflected SWIR light intensity variation with wavelength for four different materials. As shown in Figures 10a to lOd, the intensity of the reflected SWIR light varies across the spectrum and varies depending on floor material type. Trace 450a in Figure 10a shows the characteristic normalised reflected SWIR light variation for a wood floor, trace 450b in Figure 10b shows the characteristic normalised reflected light variation for a ceramic floor, trace 450c in Figure 10c shows the characteristic normalised reflected light variation for a vinyl floor, and trace 450d in Figure 10c shows the characteristic normalised reflected light variation for a laminate floor. The floor type sensor 350 is able to detect reflected SWIR light having wavelengths of 1450nm, 1700nm, 1920nm, and 2150nm - indicated by data points in Figures 10a to lOd respectively. As shown in Figures 10a to lOd, the characteristic traces 450a, 450b, 450c, 450d are distinct for each material type. Consequently, the reflected SWIR light intensity signals 364a, 364b, 364c, 364d received by the processor 110 from the detectors 354a, 354b, 354c, 354d can be used by the processor to identify the floor type or floor identifier by comparing the normalised reflected SWIR light intensity data points with known reference characteristic reflected SWIR light traces for different floor materials. Alternatively, the floor type or floor identifier may be identified by implementing machine learning algorithms which are able to identify floor type or floor identifier in dependence on the reflected SWIR light intensity signals 364a, 364b, 364c, 364d. Although normalised reflected SWIR light intensities are shown in Figures 10a to lOd and described above, it will be understood that this is illustrative only and that absolute, scaled (or otherwise pre-processed) reflected light intensity data may equally well be used. It will be understood that the use of four detectors 354a, 354b, 354c, 354d in the example above is illustrative only, and that in other examples fewer light detectors, or a greater number of light detectors may be used. In addition, it will be understood that the specific SWIR wavelengths described above are illustrative only and that any suitable number of wavelengths in the range 1 lOOnm to 2500nm may be used. To improve accuracy of floor type or floor identifier identification, it may be beneficial to use at least three reference wavelengths above 1400nm.The sensor 350 described above with respect to Figure 9 differs from the sensors 50, 250 described above with respect to Figures 5 and 7 respectively in that the light detectors 354a, 354b, 354c, 354d of the sensor 350 are configured to detect reflected SWIR light, whereas the detectors 54, 56, 254, 256 of the sensors 50, 250 are configured to detect visible and NIR reflected light. SWIR light 355 that is incident on a floor surface 29 is able to penetrate into the material of the floor to a degree where it reveals the molecular structure of the floor material. By contrast, visible and NIR light 55 that is incident on a floor surface 29 is reflected by the surface or surface coating of the floor. As a consequence of this, the characteristic traces 450a, 450b, 450c, 450d corresponding to the different materials are predictable and repeatable as they are dependent on the material properties of the floor and not on its colour or finish. Certain wavelengths of SWIR light are absorbed by certain floor materials. Such absorption wavelengths can be used to identify the floor material by targeting known absorption frequencies for particular materials. For example, laminate absorbs light having wavelengths of 1450nm and 1850nm, vinyl absorbs light having a wavelength of 1650nm, and wood absorbs light having a wavelength of 1950nm. By incorporating light detectors configured to detect SWIR light of these particular wavelengths, the sensor 350 may be able to more effectively detect particular materials by the presence of absorption bands in the characteristic reflected light traces associated with these particular wavelengths. The sensor 350 may comprise light detectors for all such wavelengths, or alternatively may comprise light detectors for multiple wavelengths, some of which are associated with detection of one or more particular materials, and some of which are not associated with detection of a particular material. Figures 1 la to 11c show schematic representations of characteristic reflected light intensity variation with wavelength for surface water. Figure Ila represents a characteristic light pattern for a drop of water detected by a visible light sensor, Figure 11b represents a characteristic light pattern for a drop of water detected by a SWIR light sensor able to detect light having a wavelength of 1450nm, and Figure 11c represents a characteristic light pattern for a drop of water detected by a SWIR light sensor able to detect light having a wavelength of 1920nm. By using SWIR light detectors able to detect light having a wavelength of 1450nm and / or 1920nm, it is possible to detect the presence of surface water. Conversely, the inventors have found that due to the absorption of these wavelengths by water molecules, it may not be possible to infer further information (e.g. floor type) about a wet / damp floor based on detecting these wavelengths alone. In such cases it may be useful to also detect wavelengths from other parts of the visible, NIR or SWIR spectrum. For example, since water is transparent to visible light, the reflected light detected in the visible spectrum would be largely unchanged for a wet / damp floor compared to a dry floor. As will be described in greater detail below, this information may be used by the controller 100 to issue a control signal 102 comprising an instruction to increase the rotational speed of the mopping rollers 30 of the wet floor cleaner 20 until the surface water has been removed. It will be understood that it is not essential to use SWIR light detectors able to detect light having a wavelength of 1450nm and / or 1920nm to detect surface water, and that SWIR light detectors able to detect light having a wavelength in the range 1440nm to 1460nm and / or 1900nm to 2000nm may be used. As mentioned above, it may be useful to detect wavelengths from parts of the visible (400nm to 750nm), NIR (751nm to 1 lOOnm) or SWIR (1 lOOnm to 2500nm) spectrum. The sensors 50, 250 described above detect wavelengths from parts of the visible and NIR spectrums, and the sensor 350 described above detect wavelengths from parts of the SWIR spectrum. A sensor comprising light detectors able to detect wavelengths from parts of at least two of the visible, NIR and / or SWIR spectrums is also contemplated. Figure 12 is a bar chart showing test results comparing floor detection accuracy for different configurations of floor type sensor. The first bar 380 shows that in a test of a floor type sensor having a plurality of light detectors configured to detect reflected light in the visible spectrum, and one light detector configured to detect reflected light in the NIR spectrum (i.e. the floor type sensor 250 described above), the average accuracy of floor detection is 54%. The second bar 381 shows an average accuracy of 70% for a floor type detector comprising only two light detectors in the SWIR spectrum. The third bar 382 shows an average accuracy of 72% for a floor type detector comprising a combination of the floor type sensor 250 and two light detectors in the SWIR spectrum. These results show that a floor type detector comprising only two light detectors in the SWIR spectrum is more accurate, on average, than the floor type detector 250, but that a floor type detector comprising a combination of the floor type sensor 250 and two light detectors in the SWIR spectrum is more accurate than each individually. The fourth bar 383 shows an average accuracy of 78% for a floor type detector comprising only four light detectors in the SWIR spectrum. The fifth bar 384 shows an average accuracy of 80% for a floor type detector comprising a combination of the floor type sensor 250 and four light detectors in the SWIR spectrum. The sixth bar 385 shows an average accuracy of 81% for a floor type sensor having an array of eight light detectors configured to detect reflected light in the SWIR spectrum. These results show that floor type sensors having more detectors in the SWIR range are more accurate than those with less, and that, even so, a floor type sensor comprising a combination of the floor type sensor 250 and four light detectors in the SWIR spectrum is more accurate than a floor type sensor comprising only four light detectors in the SWIR spectrum. It will be understood that any combination of light detectors as best suits intended end use and floor cleaner type may be selected. Although not explicitly described above, it will be understood that a floor type sensor comprising a plurality of light detectors configured to detect reflected light in the visible spectrum, and one or more light detectors configured to detect reflected light in the SWIR spectrum is contemplated. As mentioned above, the reflected light intensity signals (absolute, scaled or normalised) received by the processor 110 from the detectors can be used by the processor 110 to identify the floor type or floor identifier. In one example, this is achieved by comparing the reflected light intensity data points with known reference characteristic reflected light traces for different types and / or colour of floor. For floor type sensors which use visible and NIR light detectors, such as sensors 50 and 250 described above, the reflected light intensity signals can be compared to reference characteristic trace data stored in a look-up table which contains a plurality of characteristic traces for different types, colours and finishes of floor surface. The floor characteristics in the look-up table may be pre-loaded, or may be obtained from floors in a user’s home using the sensor 50, 250 as described in greater detail below. Where a floor identifier (such as colour) is identified by the processor 110, the floor identifier is used to identify floor type by referencing correlation data (also stored in the look-up table) which links specific floor identifiers to specific floor types. The same method of determining floor type or floor identifier can be used for floor type sensors which use SWIR light detectors, such as sensor 350 described above. In this case, the reflected light intensity data points can be compared to reference characteristic traces stored in a look-up table which contains a plurality of characteristic traces for different floor materials (and optionally also different floor colours and finishes). In either case, the reference characteristic reflected light traces may contain absolute intensity values at each of the wavelengths detected, and / or they may contain relative intensity values for each of the wavelengths detected (e.g. expressed in the form of a fraction or percentage of intensity values for one or more of the other detected wavelengths). The data points obtained by the floor type sensor 50, 250, 350 enable a floor type or floor identifier to be determined by comparison to known reference characteristic traces. The determined floor type or floor identifier may then be used to determine an operating parameter of the wet floor cleaner 20. In another example, the data points obtained by the floor type sensor 50, 250, 350 can be provided as input data to a machine learning algorithm which is configured to use the data points to determine floor type or floor identifier. Example machine learning methods include statistical learning methods such as Linear Discriminant Analysis. Once determined, the floor type or floor identifier can then be used to determine an operating parameter of the wet floor cleaner 20 which is dependent on the determined floor type. Figure 13 depicts the control system 100. The controller 100 comprises the processor 110 and a memory device 120. The processor 110 may be one or more electronic processing device which operably executes computer-readable instructions. The memory device 120 may be one or more memory device. The memory device 120 is electrically coupled to the processor 110 and configured to store instructions and / or data (for example a look-up table containing characteristic floor type trace data and / or floor identifier correlation data). The processor 110 is configured to access the memory device 120 to execute the instructions stored thereon, to access data stored thereon and / or to store data thereon. The controller 100 comprises at least one input 140 and at least one output 145. The at least one input 140 may comprise an electrical input of the controller 10. The at least one output 145 may comprise an electrical output of the controller 100. The at least one input 140 is arranged to receive the reflected light intensity signals 64, 66, 264, 266, 364a-d from the sensors 50, 250, 350, and the one or more output 145 is arranged to output one or more control signals 102 comprising one or more instruction to operate the wet floor cleaner 20 in accordance with the one or more operating parameters determined by the processor 110. The operating parameters of the wet floor cleaner 20 include speed of rotation of the mopping rollers 30 and hydration level to be applied to the mopping rollers 30 and / or floor 29. Once the floor type or floor identifier has been determined, the processor 110 determines appropriate operating parameters and outputs one or more control signals 102 comprising one or more instructions to operate the wet floor cleaner 20 in accordance with the determined operating parameters. In the case that the wet floor cleaner 20 comprises a floor type sensor 350 having detectors configured to detect reflected SWIR light having a wavelength in the range 1440nm to 1460nm and / or 1900nm to 2000nm, the processor 110 may be configured to determine if there is surface water on the floor If it is determined that surface water is present, the processor 110 may be configured to determine a speed of rotation for the mopping rollers 30 that is greater than their standard speed of rotation for the floor surface in question and / or may be configured to reduce the hydration level to be applied to the mopping rollers 30 and / or floor 29. In the case of a dry floor cleaner such as vacuum cleaner 10, if it is determined that surface water is present, the processor 110 may be configured to disable the vacuum cleaner. The floor type determined by the processor 110 may be provided as a suggested floor type to a user via a user interface. Tn an example process 400 (depicted schematically in Figure 14) the control system 100 is configured to receive the detector signals 64, 66, 264, 266, 364a-d from the floor type detection system 70, 270, 370 at step 401, and to determine a detected floor type in dependence on the one or more detector signals at step 402. At step 403 the control system outputs a detected floor type signal which may be received, and displayed to a user, by the user interface 25 at step 404. Upon receipt of the detected floor type information, at step 405, the user can input an instruction to accept or reject the detected floor type. At step 406a, 406b, the acceptance or rejection of the detected floor type is received as an input signal by the control system 100. In the case that the user accepts the detected floor type, at step 407 the control system 100 is configured to determine one or more operating parameters of the floor cleaner in dependence on the detected floor type, and at step 408 output at least one control signal 102 comprising one or more instructions to operate the floor cleaner 20 in accordance with the one or more operating parameters. In the case that the user rejects the detected floor type, the user interface 25 is configured to allow the user to input the floor type at step 409. The user interface 25 then issues a signal indicative of the user identified floor type to the control system 100 at step 410. At step 411 the control system 100 determines one or more operating parameters of the floor cleaner 20 in dependence on the user-identified floor type, and, as above, outputs at least one control signal comprising one or more instructions to operate the floor cleaner in accordance with the one or more operating parameters at step 408. In the case that the user accepts the detected floor type (step 406a), the control system 100 may be configured to associate the detected data points with the detected floor type at step 412, and issue an instruction to store the association between the detected data points and the detected floor type (for example in a look-up table stored in memory device 120) at step 413. Alternatively, in the case that the user rejects the detected floor type (step 406b), the control system may be configured to associate the detected data points with the user-identified floor type at step 414, and, as above, issue an instruction to store the association between the detected data points and the user-identified floor type (for example in a lookup table stored in memory device 120) at step 413. It will be understood that many variations to the above examples are possible. In particular, the light emitters 52, 252, 352 may be separate to the sensors 70, 270, 370. In addition, the user interface 25 need not be an integral part of the floor cleaner but may instead be a remote device such as a mobile telephone, tablet or computer. Although the examples above have in the main been given in the context of a wet floor cleaner 20 having a pair of mopping rollers 30 with a floor type sensor 50, 250, 350 located between them, it will be understood that all of the examples may also be used with only one, or more than two, mopping rollers. The floor type detection system may be used with a dry floor cleaner such as a vacuum cleaner 10 having only one, or a plurality, of agitators. The floor cleaner 10, 20 may be user operated or may be a self-propelled robotic floor cleaner.

Claims

1. A floor type detection system for a floor cleaner, wherein the floor type detection system comprises:one or more light emitters;a light sensor configured to detect light having wavelengths in a reference range of 1 lOOnm to 3000nm and output one or more sensor signals indicative of the intensity of the light at a plurality of reference wavelengths within the reference range; andone or more processors collectively configured to:receive the one or more sensor signals; anddetermine a floor type in dependence on the one or more sensor signals.

2. The floor type detection system of claim 1, wherein the light sensor is configured to detect light having wavelengths in a reference range of 1 lOOnm to 2500nm, or 1400nm to 2500nm.

3. The floor type detection system of claim 1 or 2, wherein at least one of the reference wavelengths is in the range:1440nm to 1460nm; or1900nm to 2000nm.

4. The floor type detection system of any preceding claim, wherein at least one of the reference wavelengths is in the range: ...1440nm to 1470nm;1640nm to 1670nm;1840nmto 1870nm;1940nm to 1970nm.

5. The floor type detection system of any preceding claim, wherein the plurality of reference wavelengths comprise at least two, at least three or at least four different wavelengths.

6. The floor type detection system of any preceding claim, wherein at least three of the reference wavelengths are above 1400nm.

7. The floor type detection system of any preceding claim, comprising:a visible light detector configured to detect light of a visible wavelength and output a visible light signal indicative of the intensity of the light of the visible wavelength, wherein the visible wavelength is in the range 400 to 750nm; and / ora near infra-red (NIR) light detector configured to detect light of an NIR wavelength and output an NIR signal indicative of the intensity of the light of the NIR wavelength, wherein the NIR wavelength is in the range 751 to 1099nm,wherein the one or more processors are configured to:receive the visible and / or NIR signals; anddetermine a floor type in dependence on the visible and / or NIR signals.

8. The floor type detection system of any preceding claim, wherein determining the floor type comprises:using the sensor signals indicative of the intensity of the light at a plurality of reference wavelengths to determine a corresponding plurality of data points;comparing the plurality of data points with reference characteristic floor data; anddetermining the floor type in dependence on a correlation between the plurality of data points and the reference characteristic floor data.

9. The floor type detection system of claim 8, wherein the reference characteristic floor data comprises a continuous trace, or an array of discrete values, indicative of:absolute reflected light intensity across a spectrum of wavelengths;normalised reflected light intensity across a spectrum of wavelengths; or scaled reflected light intensity across a spectrum of wavelengths.

10. The floor type detection system of claim 8 or 9, comprising:determining a floor identifier in dependence on the correlation between the plurality of data points and the reference characteristic floor data; anddetermining a floor type in dependence on the floor identifier.

11. The floor type detection system of claim 10, wherein the floor identifier comprises a colour.

12. A control system for controlling a floor cleaner, the control system comprising the floor type detection system of any preceding claim, wherein the one or more processors are configured to:determine one or more operating parameters of the floor cleaner in dependence on the floor type or floor identifier; andoutput at least one control signal comprising one or more instructions to operate the floor cleaner in accordance with the one or more operating parameters.

13. The control system of claim 12, wherein the one or more processors are configured to:output a detected floor type signal; andreceive a user input signal indicative of:acceptance of the detected floor type; or rejection of the detected floor type;wherein if the user input signal is indicative of acceptance of the detected floor type, the control system is configured to:determine one or more operating parameters of the floor cleaner in dependence on the detected floor type.

14. The control system of claim 13, wherein if the user input signal is indicative of rejection of the detected floor type, the control system is configured to:receive a user-identified floor type input signal indicative of a user-identified floor type;determine one or more operating parameters of the floor cleaner in dependence on the user-identified floor type input signal.

15. The floor type detection system of claim 14, wherein the one or more processors are configured to:associate the one or more sensor signals with the floor type identified by the user; andissue an instruction to store the association between the one or more sensor signals and the floor type identified by the user.

16. The control system of any one of claims 12 to 15, wherein the one or more operating parameters comprise:a first speed of rotation for a roller or agitator element of the floor cleaner; and / ora level of hydration to be applied to the floor to be cleaned and / or to a roller or agitator of the floor cleaner.

17. The control system of any one of claims 12 to 16, comprising the floor type detection system of claim 3, wherein the one or more processors are configured to:determine that surface water is present if the one or more sensor signals indicate that the intensity of the light at the least one of reference wavelength in the range 1440nm to 1460nm or 1900nm to 2000nm is below a predetermined threshold; andif it is determined that surface water is present:determine a second speed of rotation that is greater than first speed of rotation; orreduce the level of hydration applied to the mopping rollers 30; or reduce the level of hydration applied to the floor.

18. A floor cleaner comprising the floor type detection system of any one of claims 1 to 11, or the control system of any one of claims 12 to 17.

Images