SENSOR SYSTEM AND METHOD

Abstract

A system for monitoring devices and appliances includes a series of sensors attached to the devices, the sensors generating data indicative of some action on or in relation to the devices or appliances. The data generated by the sensors is processed by suitable processor, the data providing information relating to the device and enabling servicing of the device to be scheduled, consumables required by the device to be ordered or further actions required by the device or appliance to be undertaken.

Description

The present invention relates to a sensor system and method. More specifically but not exclusively it relates to a sensor system and method for monitoring appliances, detecting events occurring in relation to said appliances and processing data generated in relation to said events.

This application claims priority from UK Patent Application Nos. GB1520522.2 filed on 20 Nov. 2015, GB1601294.0 filed on 25 Jan. 2016 and GB1603419.1 and GB1603420.9 filed on 26 Feb. 2016, the entire contents of which are hereby incorporated by reference.

In the retail sector, there is a demand to know what is happening in users' premises, for example kitchens (including fridges, cupboards, larders, surfaces, waste bins, ovens, coffee machines, kettles and washing machines), bathrooms (including storage areas, WCs, baths, showers, cabinets, sinks, toilet roll holders and WC bleach holders), cars (including screen washes), offices, industrial plants and any other areas in which consumables are stored and consumed. The requirement is driven by the desire to predict better what customers or users might wish to order before they have a clue themselves. This information is also needed in relation to our customer's businesses (e.g. offices, retail stores, factories) so that the opportunities for replenishing inventory levels may be understood before the customer (or the customer's customer) is aware of the need. In both consumer and industrial premises, there is also a desire to gain greater insights into the use of equipment and appliances, to understand patterns of behaviour and how different events are correlated.

At some point some in the future, appliances, kitchen furniture, office fixtures, industrial storage areas and other such equipment units will become smart enough to provide these data by using technology directly built into the unit. However, at present there is a need to gather such consumption data in relation to appliances, devices, unitsand storage areas that are not currently provided with such smart technology.

So the key shorter term objective is to find a way to collect these insights without having to build the smart technology into an appliance or a piece of furniture, both so that it can be retrofitted to existing equipment but also to establish a standard that might be adopted by smart appliance manufacturers.

According to the invention there is provided a system for monitoring apparatus used by a user, the system comprising at least one sensor communicably attached to the apparatus, the or each sensor comprising detecting means for detecting events indicating use of or in relation to the apparatus, the or each sensor further comprising transmitting means for transmitting output data generated by the sensor to signal processing means, said signal processing means comprising a determining utility, the determining utility comprising means capable of determining the use pattern of the apparatus such that the user may be informed of actions required in relation to the apparatus.

According to the invention there is further provided A method of monitoring apparatus comprising the steps of:

(a) detecting events relating to the use of the apparatus using at least one sensor means;

(b) generating a signal indicative of the event detected;

(c) outputting the signal to transmitting means;

(d) transmitting the signal to signal processing means, the signal processing means being located within a cloud computing platform having one or more servers and a database coupled to the one or more servers, the signal processing means being operable on the cloud computing platform;

(e) processing said transmitted signal, said processed signal comprising data indicative of the use of the apparatus; and

(f) replenishing, servicing or maintaining the apparatus based on the indications of the use of the apparatus.

In this way, a system is provided that gathers data and transmits said data to the cloud where event detection and signal processing occurs to enable estimations of usage and consumption of consumables to be made with a view to automatically (or otherwise) replenishing a user's stock or identifying a need for stock to be replenished.

The invention will now be described, with reference to the accompanying diagrammatic drawings in which:

FIG. 1 is a schematic diagrammatic drawing of a system in accordance with a first form of the invention, showing an appliance with a sensor mounted thereon, the sensor capable of transmitting data in relation to the appliance via suitable monitoring and communication means to data processing means in cloud storage means;

FIG. 2 is a schematic flow diagram of the system of FIG. 1 showing the flow of data around the various sub-systems; and

FIG. 3 is a schematic flow diagram of a second form of data flow around the system, the individual devices being provided with integrated sensors communicating directly with cloud providers.

FIG. 1 shows a schematic diagram of one form of the invention. An appliance such as a refrigerator 1 is provided with a sensor mountable 3 on the appliance. The sensor 3 may comprise a small encapsulated device that can simply be stuck on the side of an appliance, for example in this case on the fridge and/or freezer door. The encapsulated device may be powered by battery, solar cell, or power supply that fits in-line with the power supply of the appliance, such as in-line with a 13 amp plug. In this way the power adapter plugs into a power outlet and the appliance is plugged in to the back of the adaptor. It will be appreciated that any other suitable form of power supply may be used.

The sensor 3 is capable of communication with other devices within a network of devices on appliances within a user's environment, and may communicate with other devices within the network via Bluetooth, Zigbee, wifi with other devices so as to, for example form a mesh network or may communicate with internet router.

The sensor 3 may detect an action indicative of an incident associated with the device, in this case the fridge. For example, the sensor 3 may detect motion, sound, vibration, power consumption, permittivity, permeability, conductivity, light levels, infra red levels, temperature, spectral absorption at various frequencies or any other detectable incident. These detections may relate to incidents, the data on which may be used to predict the type of use of the device and whether any consumable related to the device or item in the device may have been depleted and need replenishing.

The sensor 3 may use a combination of analogue and digital signal processing on these data streams (including detecting correlation between multi-sensory inputs) to detect events such as door opening, in the example above of a fridge, a spin cycle of a washing machine, kettle or toaster activation, depletion of the contents of a container or an item (e.g. a kitchen roll) or depositing of packaging in a dustbin or any other incident where such an appliance is used or an item depleted.

In use, a user may initially train the sensor 3 by performing a form of calibration where the cycle is starting and/or finishing such that the sensor may learn the signature of a detectable incident. However, it will be appreciated that, over time, a library of signatures may be developed (e.g. the sound signature of different brands of washing machine; the permeability, permittivity signature of strawberry jam).

Raw data generated by the sensor 3 may be stored locally on the device for later analysis, transmitted via a router to a cloud processing system where the signal processing and event detection would be done.

It will be appreciated that when not sensing, the sensor 3 may be in a stand-by mode and may be reactivated by, for example, Bluetooth communication being used to wake up a computer such as a tablet in the vicinity of the sensor and appliance to send summarised consumption data or re-ordering suggestions/requests from a machine.

The signal processing comprises analysis of the raw data transmitted by the or each device in the network either directly or via a router at the user's premises. The data analysis relating to, for example power consumption, sound vibration, movement etc of the individual devices may be undertaken to detect the signature of events from which consumption of items or consumables may be inferred. For example, coffee capsules would be used one per cycle of a coffee machine, dishwasher capsules would be used every N cycles of a dishwasher, or even every N times a fridge door is opened being used to remind a user about things that may need to be reordered, for example milk.

In the general description above, the concept relating to the invention is described. Specific examples of parts of the system will now be described in an exemplary manner only.

Exemplary Types of Sensor

In order to generate data in relation to incidents described above numerous types of sensor or detector may be used. For example:

Motion

A 3-axis accelerometer may be used to detect movement such as a cupboard or fridge door opening, a jar or box being lifted or a drawer being moved.

Sound

A sound signature of an event such as a washing machine spin cycle may be recorded using acoustic sensors such as microphones.

Vibration

A vibration or shock signature of an event such as a washing machine spin cycle may be monitored using accelerometers.

Power consumption

Power consumption may be monitored in many ways, for example, coiling clips around power cable of an appliance may measure the signature of the current (and hence power) usage of the appliance. This may be used to provide the power signature of events such as a kettle or toaster being operated.

Fluid Level

Fluid levels of dispensers associated with appliances such as washing machines may be monitored using ultrasonic sensors, capacitance or any other suitable means.

Proximity

A proximity detector may be used to detect the proximity of a user to an appliance to infer use of the appliance. This may be by using infrared detectors or by other means.

Fridge and Shelf Occupancy

Occupancy of positions in appliances such as fridges may be inferred using visual techniques or other techniques based on the absorption, reflection or transmission of electromagnetic radiation. Alternatively, the permittivity or permeability of the contents may be measured using a field emitting and receiving device.

Vision

Camera means may be used to visually assess events or incidents occurring or even to monitor levels of consumables may be considered. Whilst somewhat controversial from a security perspective, users may be willing to allow a small camera to record detectable events. Another example that may be envisaged is the use of some form of visual reader such as a bar code reader associated with a bin such that items deposited in the bin are captured.

Environmental data

Although not necessarily related to the appliance events it may be useful for the device to record environmental data such as light level, temperature and time of each event. In this way additional information may be inferred from the data collected in relation to other sensors, for example more milk may be used at breakfast for a given fridge opening event being detected than at another time of the day.

Infrared and other frequency spectroscopy

IR and other frequency spectroscopy may be used to monitor a signature of an item when irradiated with low energy electromagnetic radiation of various frequencies (electrical, optical, infra red, ultra violet, radio frequencies, lower frequencies). Using an array or multiple arrays of transmitters/receivers may enable measurements to be localised to a position in 3D space and hence for inventory levels to be monitored.

It will be appreciated that the sensors and devices above are selected from a large number of sensing devices that may be used to record incidents and are given as examples only. Many other sensors may be used in the same manner to generate data indicative of events that may be interpreted in the following manner.

Data Generation

Event Processing

Raw data is generated by the or each of the sensors described above detecting events indicative of some action in relation to the device or fixture. The raw data may be streamed to the cloud where it may be stored and processed (using suitable digital signal processing techniques) to extract event data and to infer information as to the level of items in a cupboard or the need to order consumables for given devices and appliances or any other such event capable of detection and monitoring.

It will be appreciated that the sensor must be linked to a given device, unit or appliance in order that the data generated may be attributed to a given appliance, unit, or fixture and the event information correctly attributed in the data processing utility. For example, the user may use a mobile app to configure each of the above sensors or devices using information such as “Name of appliance” “Type of appliance” and “Details of the products consumed”—potentially by scanning barcode

Replenishment

Once a potentially required product is linked to a given appliance, the user's historical shopping data may be used to determine the rate at which the product is consumed. Once such consumption events are being recorded, a good prediction may be made of the number of consumption events that relate to a purchase of that product. In some cases it will be obvious, for example, one coffee capsule per use of the coffee machine.

Based on the type of appliance an event processing engine would make an assumption about the number of events required to trigger a replenishment event. A replenishment event would result in the consumed product (or products) being added to the user's online order basket with a marker to indicate that it has been automatically added by the relevant appliance. For example “The dishwasher just added washing tablets”; “The bin just added chicken”; “The bathroom just added toilet rolls”; “the larder just added olive oil”.

The user then accepts or rejects the basket addition. One reason for rejecting is that it may have added the product after too few events, in which case the replenishment engine would adjust the number of appliance events before the next replenishment event; in other words it would learn the rate at which products are consumed. It would also be possible for customers to manually edit the number of appliance events in a replenishment event using the mobile application. If customers manually add products to their basket that are the subject of consumption events, then the system would take account of this in terms of future replenishment events.

For some appliances, such as a vacuum cleaner, the length of the cycle and how hard the appliance is working might be relevant for predicting when the dust bag would fill up.

General Applicability

The concept extends to any situation in which a consumable is being depleted or in which a maintenance intervention is required after a certain period of time or usage (e.g. descale the kettle every six months or after every 200 uses) in any setting (home, shop, office, factory, business. . . ).

Machine Learning and Crowd-Sourcing

The relationship between the sensory measurements and the consumption of the consumable has the potential to be learned automatically; and for the accuracy to be improved by “crowd sourcing” through aggregating the acceptance/rejection of the corresponding orders by many people who use the same measuring device (e.g. kitchen roll emptiness detector+kitchen rolls in storage detector).

Sale of Timely Marketing Opportunities

The system would know the stock levels and rate of consumptions of consumables (and which brand of consumables had been bought) in a large number of customers. It would be possible to use this information to assist FMCG companies with marketing of existing products and launching new products, by targeting those potential customers who were about to run out of the product. It would also be possible to use this information to carry out product comparison for example, on average customers who bought Fairy Liquid rather than a less premium brand had to re-buy it less often e.g. because it lasted longer.

By combining bin scanning with stock level information with user ordering information, it would be possible to estimate the percentage of household grocery spend that is with a given retailer; and start to understand what sort of products this relates to, and hence take steps to investigate how more of the percentage not spent with that given retailer could be captured.

It will be appreciated that in order to achieve this level of data processing a number of mathematical models and algorithms may be implemented. These will now be described in more detail by example.

Device Events

A device event is an event that takes place within a device, appliance or unit such as dishwasher cycle starting, a fridge door opening, a vacuum cleaner being activated etc. Device events will be detected by analysing data returned from one or more sensors attached to the event monitor (EM). For example when a coffee machine is activated it produces audio, vibration and power consumption data that an attached EM would record and then stream to an Event Monitoring Service (EMS) running in the cloud.

When setting up an EM, the user would enter information about the device using a Configuration Application (CA) on a mobile phone or on a website. This would include picking from a list of generic device types (e.g. coffee machine) and then optionally from a list of specific device brands and models (e.g. Nespresso Pixie). The device type (and possibly the brand/model) would enable the EM to know what sensors it should activate to record that device's events.

The EMS would use the device information to configure the digital signal processing (DSP) algorithms that would be used to detect the device event or series of device events. For example a coffee machine might generate several different types of event such as pre-heating the water, grinding the coffee, frothing the milk, making the coffee and so on.

For a type, brand or model of device that is not known to the CA, the user would activate the device and then tell the CA when a significant event is starting and finishing. This would enable the device (or possibly the EMS) to learn what sensors are required to detect the event data and how the DSP algorithms should be configured to isolate the event. The user might be required to activate the device more than once to enable multiple readings to be taken. The CA might allow the user to take a picture of the device (or its identifying barcode, model information etc) and then the EMS could use image recognition to identify the device.

The information relating to device events and potentially their correlation (e.g. boiling a kettle and then using the toaster) might be useful information in their own right to suppliers etc.

Consumption Events

A consumption event is a type of device event that relates to the consumption of a resource that the user wishes to record. For example, the water pre-heating event for a coffee machine is a device event that would probably not be considered a consumption event (other than for water) whereas the coffee grinding event would be a consumption event for coffee.

Selecting a type of device (or a specific device) via the CA would also select a set of possible configuration events for that device. The user would select which configuration events they are interested in and for each, they would enter information about the specific product (or products) being consumed. For devices that are not known to the CA, the user would have to manually select the device events that are consumption events.

There are different categories of configuration event, including:

Discrete

These events result in the consumption of a discrete quantity of the resource. For example, a capsule coffee machine consumes one capsule per activation

Measured

These events result in the consumption of a measured quantity of the resource. For example, a dishwasher consuming rinse-aid, a washing machine consuming detergent and so on

Variable

These events result in the consumption of a variable quantity of the resource depending on factors such as the length of time the event lasts, how hard the device is working etc. For example, the rate at which a vacuum cleaner bag fills up has some relationship to how long it is used for and how hard it is working. The factors can be measured but information about what material it is picking up is clearly much harder to measure.

Replenishment Events

Replenishment events occur when a resource is fully consumed and must be replenished.

In order to predict when a replenishment event is going to occur, it is necessary for the EMS to learn the relationship between the replenishment event and its underlying consumption event(s). A replenishment event may be linked to more than one consumption event, for example if a household has more than one vacuum cleaner of the same type.

One variable in this relationship is the Consumption Quantum (CQ)—the amount of the resource consumed by each consumption event. The CQ depends on the type of consumption event. For example for a capsule coffee machine it would be one capsule per event or for a coffee machine that grinds coffee (or uses ground coffee) it would be an amount of coffee measured by weight or volume.

The other variable would be the number of CQs required to fully consume the resource and trigger a replenishment event. This would obviously depend on when the resource was last replenished and how much was sourced at that time.

The replenishment information could be obtained by different means, including:

Using information on the frequency and quantity of the consumer resource that the user purchases. This could be extracted from their online shopping account, scanned till receipts or any other suitable means.

The EMS would use machine learning techniques to map the CQ for each consumption event onto the corresponding replenishment events. However when a new device is configured, the EMS needs to know its CQ value(s) before the first few replenishment events have occurred and the machine learning has enough data to base its predictions on. If the device is of a known type, the EMS could you an average CQ value derived from other devices of the same type.

The user can then either start the first replenishment cycle immediately by entering the amount of the product they have in-stock or wait until the product runs out and then start the first replenishment cycle at that point.

Commercial Events

A commercial event relates to commercial actions taken in response to a replenishment event occurring or the probability of a replenishment event occurring within a given period.

For example, a typical commercial event might be to trigger targeted advertising, promotions, voucher allocations etc. The correlation of replenishment events might also trigger multi-buy offers. For example,

• • Customers enter details (or scan the barcode) of the product that relates to a consumption event, using the mobile app. The anonymised data could be valuable to suppliers and appliance manufacturers in terms of customers' use of their appliances and consumption of their products. The correlation of events could also be valuable, such as when a customer has a cup of coffee they also use the toaster
  • A retailer would be able to provide valuable data to consumables manufacturers. For example we might provide access to 1,000 customers who are about to run out of kitchen roll so that a brand owner can suggest trying out a new one they are launching
  • “vendor managed inventory” services may be offered in which the retailer automatically resupplies consumers or businesses to maintain an agreed level of inventory. For example, a consumer might request that they always have one spare bottle of tomato ketchup in their cupboard.

It will be appreciated that there are a large number of events that could be monitored both in domestic and commercial settings. The following is a list of exemplary situations, which should not be regarded as exhaustive, there are many similar situations in which such information may be generated and monitored by the equipment detailed above.

• • Office equipment requiring supplies (e.g. printers, copiers, soap dispensers). For example, the number of pages printed by a printer may be indicative of ink cartridges requiring replenishment. Equally, the level of ink in the relevant cartridge may be monitored directly and ink ordered by the printer direct from the supplier as necessary;
  • Equipment requiring servicing based on number of operations (e.g. lifts, air conditioning). For example, the number of times a door is opened and closed may be indicative of the use level of equipment such as lifts or elevators, environmental monitoring of locations may be indicative of the servicing requirement of air conditioning units;
  • Rooms requiring cleaning based on the number of people who have gone in/out (e.g. toilets, kitchens). Again this may be monitored by detecting the number of times a door has opened or closed. Alternatively, water usage may be monitored.
  • Containers requiring emptying (e.g. confidential waste, paper shredders). The levels of waste in bins or shredders may be monitored by the number of times the shredder has been operated, alternatively optical methods may be employed to establish if bins are full.
  • Retail or industrial shelves requiring monitoring for the number of items and the correct items. This may be achieved by optical monitoring or in the case of establishing whether shelves needed refilling, weighing means may be used.
  • Machines requiring replacement parts (e.g. cutting heads; spark plugs; rubber washers; insulation; belts; chains; filters; batteries; springs; fan blades; media; plastic pellets; thread; glue; tape; plastic wrapping; oil; cleaning; gas; OLEDs; coolant amount and purity, other fluids). There are numerous ways that these may be measured including optical, power consumption and fluid level monitoring.
  • Warehouse stock level monitoring, for example by optical or camera means or weighing means.
  • Human fatigue level monitoring (e.g. drivers of vehicles; operators of machinery). Events such as these may be monitored via optical means detecting eye movement, for example. Alternatively, efficiency of operation may be monitored by the rate at which human interactive events happen, for example number of actions a minute may be monitored.

It will be appreciated that all the events that may be monitored above in relation to a domestic situation may equally be applied to a commercial situation, for example industrial kitchens and the like.

Claims

1. A system for monitoring apparatus used by a user, the system comprising: at least one sensor configured to be communicably attached to an apparatus, the sensor having detecting means for detecting events indicating use of or in relation to the apparatus, and transmitting means for transmitting output data generated by the sensor anda signal processing means for receiving the output data, and having a determining utility, the determining utility having means for determining a use pattern of the apparatus such that a user may be informed of actions required in relation to the apparatus.

2. A system for monitoring apparatus according to claim 1, the signal processing means comprising; a data utility, the data utility having means for establishing relevant information in relation to the user or the apparatus such that said user may be provided with information or incentives or consumables for use in association with said apparatus.

3. A system according to claim 1, the signal processing means comprising: a commercial utility, the commercial utility having means establishing relevant information in relation to the user, the apparatus, or a location of the apparatus such that providers of services may be provided with information in relation to the user, the apparatus or the location of the apparatus.

4. A system according to claim 1, comprising: a cloud computing platform having one or more servers and a database coupled to the one or more servers, the signal processing means and the utility or utilities being operable on the cloud computing platform.

5. A system according to claim 1, in which the transmitting means comprises: router means, the router means being configured in transmitting the output data from each sensor to the signal processing means.

6. A method of monitoring an apparatus, the method comprising: (a) detecting events relating to the use of the apparatus using at least one sensor means;(b) generating a signal indicative of the event detected;(c) outputting the signal to a transmitting means;(d) transmitting the signal to a signal processing means, the signal processing means being located within a cloud computing platform having one or more servers and a database coupled to the one or more servers, the signal processing means being operable on the cloud computing platform;(e) processing said transmitted signal, said processed signal including data indicative of the use of the apparatus; and(f) replenishing, servicing or maintaining the apparatus based on indications of the use of the apparatus.

7. A method according to claim 6, comprising: performing data analysis of the data generated in (e), the data analysis providing information relating to the location, use pattern and replenishment requirements of the apparatus.

8. A method according to claim 6, comprising: informing the user of requirements of the apparatus.

9. A system according to claim 1, in combination with the apparatus, the apparatus being at least one of: a domestic or commercial appliance, peripheral, fixture, fitting, or unit that requires servicing, maintenance or replenishment with consumables.

10. A system according to claim 1, in which the sensor is one or more of the following sensor types: a motion sensor, sound sensor, vibration sensor, power consumption monitoring device, fluid level sensor, proximity detector, sensor monitoring absorption, reflection or transmission of EM radiation, vision sensor and IR or spectroscopy equipment.