A method and system for analysing environmental data

The present invention relates to methods and systems for analysing

environmental data.

Pollution is the introduction of contaminants into an environment that causes instability, disorder, harm or discomfort to the ecosystem i.e. physical systems or living organisms. Pollution can take the form of chemical substances or energy, such as noise, heat, or light. Pollutants, the basic elements of pollution, can be foreign substances or energies, or naturally occurring. When naturally occurring, they are considered contaminants when they exceed natural levels. Pollution events may be defined as the exceeding of pollutant parameter values relative to normal or desired ambient conditions. Characteristic of many pollution events is their variation in time and location. This time dependency originates from many factors and it includes both variation in the scale of minutes up to hours and even of longer period variation such as seasonal variations. For example, pollution events related to transportation (road vehicles, aeroplanes, trains etc.) may include high air and noise pollution levels subject to high transportation activity, or to the geographical vicinity to the transportation activity. These events are characterised by short term, high pollution levels commonly called pollution peaks.

Environmental data can include data regarding the level of various environmental characteristics. This can include the level of various substances in the

environment. These substances can include particulate substances, such as dust, soot and smoke. These substances can also include gaseous substances, for example nitrogen dioxide, carbon monoxide and ozone. Environmental data can also include data on other environmental characteristics, such as noise.

At least some of these environmental characteristics can be considered to be pollutants. These pollutants can come from a variety of sources, for example industrial emissions and vehicle emissions. Pollution in the environment can have several adverse effects. These include harmful effects on human health. For example, inhalation of particulate matter, especially small particles, has been linked to various respiratory diseases.

Inhalation of gaseous pollutants may also cause or exacerbate disease or illness in humans. Other types of pollution, for example noise pollution, can also have adverse effects on human health.

Various types of pollution may be associated with various risks to health. The various types of pollution can be generated by different sources of pollution.

It is desirable to minimise exposure to such pollutants to minimise the risk to health, it is also desirable to know about the possible sources of pollution, so that the risk to health can be controlled. A conventional sensing device may sense one pollutant. Such a sensing device can provide readings of levels of that pollutant, but may be unable to distinguish the source of pollution.

The sensing device may be inaccurate and/or the data from the sensing device may be noisy. Overall data quality can therefore be poor. This may reduce the reliability of any analysis conducted with the data.

It is an aim of the present invention to address at least one of these drawbacks. According to a first aspect of the present invention there is provided an environmental monitoring system comprising at least one sensor for sensing an environmental parameter.

According to a second aspect of the present invention there is provided an environmental monitoring system comprising a plurality of sensors each for sensing at least one respective environmental parameter, and a central unit arranged to receive readings from the plurality of sensors. The central unit may be arranged to receive the readings over a communications network, such as the internet. According to a third aspect of the present invention there is provided a pollution source identification method comprising the steps of obtaining readings from a plurality of sensors each for sensing at least one respective environmental parameter; analyzing the readings using rules which link sensor readings to pollution characteristics; and determining characteristics of an indicated pollution event in dependence on said analysis.

A pollution event may, for example, be (or it may be due to) a source of pollution, or it may be {or it may be due to) pollution that has moved from its source. As an example, a pollution source may be a combustion engine; a pollution event may be due to the pollution in the vicinity of the combustion engine, or it may be due to pollution from the combustion engine that has drifted downwind.

According to a fourth aspect of the present invention there is provided a method of assessing the validity of environmental data comprising the steps of obtaining readings from a plurality of sensors each for sensing at least one respective environmental parameter; analyzing the readings using rules which link sensor readings to validity of those readings; and determining the validity of the obtained readings in dependence on said analysis.

The environmental monitoring system may be arranged to carry out a method as defined above, the system comprising a plurality of sensors each for sensing at least one respective environmental parameter, and a central unit arranged to receive readings from the plurality of sensors.

The plurality of sensors may comprise at least one sensor for sensing a first parameter and at least one sensor for sensing a second, different parameter. The plurality of sensors may comprise at least one sensor for sensing a parameter at a first location and at least one sensor for sensing a parameter at a second, different location.

The central unit may comprise a server and may comprise a database. The central unit may be arranged to determine a recommended action to be taken in response to the readings received from the plurality of sensors. The method may comprise determining a recommended action to be taken in response to the readings received from the plurality of sensors. The system may be arranged to determine a recommended action to be taken in response to the readings received from the plurality of sensors, The central unit may be arranged to output data to end users determined on the basis of the readings received from the plurality of sensors, The method may comprise outputting data to end users determined on the basis of the readings received from the plurality of sensors. According to a fifth aspect of the present invention, there is provided a pollution source identification method comprising the steps of

analysing data obtained from a sensing means for sensing an

environmental parameter, the analysis using at least one rule which links the obtained data to at least one pollution characteristic, and

determining characteristics of an indicated pollution event in dependence on said analysis.

According to a sixth aspect of the present invention, there is provided a pollution source identification method comprising the steps of

analysing data obtained from a sensing apparatus for sensing an environmental parameter, the analysis using at least one rule which links the obtained data to at least one pollution characteristic, and

determining characteristics of an indicated pollution event in dependence on said analysis.

The determination of characteristics of an indicated pollution event in this way may allow risk minimisation or avoidance action to be taken. For example, if a pollution event or source is identified in a particular area, that area could be avoided to minimise exposure to pollution from the identified pollution event or source.

The data may comprise one or more reading.

The method may comprise the step of obtaining the data from the sensing means and/or apparatus. The method may include the step of sensing an environmental parameter using the sensing means and/or apparatus to obtain the data. The method may include the step of sensing a plurality of environmental parameters using the sensing means and/or apparatus to obtain the data. Basing the pollution source identification on a plurality of environmental parameters may increase the accuracy of the identification. This can help in narrowing down the possible source of pollution. Basing the pollution source identification on a plurality of environmental parameters may also allow the proximity of the pollution source to be determined, or to estimate the distance to the pollution source more accurately. This can be done, as will be explained in more detail below, by analysing the changes and/or rates of change in the sensed parameters and/or the timings of these changes and/or rates of change.

Basing the pollution source identification on a plurality of environmental parameters may also allow the determination of the intensity of a pollution source. Again, this can be done, as will be explained in more detail below, by analysing the changes and/or rates of change in the sensed parameters and/or the timings of these changes and/or rates of change. The sensing means and/or sensing apparatus may comprise a sensor. The sensor may sense at least one environmental parameter. The sensing means and/or apparatus may comprise a plurality of sensors. Each of the plurality of sensors may sense at least one respective environmental parameter. The plurality of sensors may comprise at least one sensor for sensing a first parameter and at least one sensor for sensing a second, different parameter. The plurality of sensors may comprise at least one sensor for sensing a parameter at a first location and at least one sensor for sensing a parameter at a second, different location.

The step of determining characteristics of an indicated pollution event may comprise identifying the pollution source based on the analysis.

The step of analysing the data may comprise comparing the data to reference data. The reference data may be stored data. The reference data may be representative of at least one pollution characteristic. The pollution source may be identified based on the comparison.

The comparison of the data to reference data may comprise using the at least one rule. In other words, the comparison may comprise determining whether the at least one rule is satisfied.

At least one of the at least one rule may be user-definable and/or user modifiable. The method may permit identification of a pollution source on the basis of one or more pre-set or pre-determined pollution characteristic and/or on the basis of a user-defined or user-modified pollution characteristic. The identification of the pollution source can be achieved by comparison with differing reference data. The identification can take into account pollution characteristics in a particular local environment.

The obtained data may comprise a sensed parameter value. The sensed parameter value may be representative of a sensed environmental parameter. The obtained data may comprise a plurality of sensed parameter values. The plurality of sensed parameter values may be representative of different sensed environmental parameters and/or an environmental parameter sensed at different times.

The reference data may comprise a first reference parameter value. The first reference parameter value may comprise a value representative of a pollution source and/or a pollution event. In other words, the first reference parameter value may be a value of an environmental parameter that is indicative of a pollution source and/or a pollution event. The step of comparing the obtained data to the first reference parameter value may include determining whether the sensed parameter value is greater or less than the first reference parameter value.

A finding that the sensed parameter value is, for example, greater than the first reference parameter value may lead to a determination that a pollution event has occurred. The first reference parameter value can be set at a level such that exceeding this value results in a high confidence level that the pollution event has occurred.

Of course, it may also be the case that a finding that the sensed parameter value is, for example, less than the first reference parameter value may lead to a determination that a pollution event has occurred. In this case, the first reference parameter value can be set at a level such that falling below this value results in a high confidence level that the pollution event has occurred.

The reference data may comprise a second reference parameter value. The second reference parameter value may comprise a value representative of a threshold difference value indicating a pollution event. The step of comparing the obtained data to the second reference parameter value may include calculating a sensed difference value from the difference between the sensed parameter value and an ambient parameter value representative of the value of the respective environmental parameter under ambient conditions. The ambient parameter value may be a predetermined value. The ambient parameter value may be calculated as an average of the respective environmental parameter values over a given time period.

The method may comprise determining whether the sensed difference value is greater or less than the threshold difference value.

Thus the comparison might indicate that the sensed environmental parameter has changed by an amount such that its value is above (or below) a given threshold difference from an ambient value. Such a change can be representative of a pollution source, for example if the environmental parameter value increases such that it is above a value for that parameter that might occur naturally in the environment. The pollution source can be identified based on the environmental parameter that is sensed, and/or on the amount by which the value differs from the ambient value.

The step of sensing the environmental parameter may include sensing the environmental parameter at a first time to obtain a first sensed parameter value, and sensing the environmental parameter at a second, later, time to obtain a second sensed parameter value. The first sensed parameter value and/or the second sensed parameter value may be stored. The method might include storing at least one of the first sensed parameter value and the second sensed parameter value as part of the stored data. The step of comparing the sensed parameter value may include comparing the second sensed parameter value with the first sensed parameter value.

Thus the comparison might indicate that the sensed environmental parameter has changed by a predetermined amount in a given time period. This may be indicative of a particular source of pollution or of a pollution event.

The step of sensing the environmental parameter may include obtaining a change value representative of the change from the first to the second sensed parameter value, in other words, obtaining a value representing the amount by which the sensed environmental parameter has changed between the first and second times. The stored data may comprise a reference change value representative of a threshold change in the environmental parameter.

Thus the comparison might indicate that the sensed environmental parameter has changed by an amount in a given time period which is greater than a threshold change value. In other words, if the change value is greater than the reference change value, this may be indicative of a pollution event.

The step of sensing the environmental parameter might include calculating the rate of change of the environmental parameter. The step of calculating the rate of change of the environmental parameter may include calculating the rate of change based on the change value, i.e. the change from the first to the second sensed parameter values, and the time difference between the first and second times.

The stored data may comprise a reference rate of change value representative of a threshold rate of change in the environmental parameter.

Thus the step of comparing the data to reference data may indicate that the sensed environmental parameter has changed at a rate that is greater than the reference rate of change value. This can be indicative of a source of pollution, and can allow identification of the pollution source.

The method may include the step of determining, based on the rate of change of the environmental parameter, the proximity of the pollution source. The method may include the step of estimating the distance from the pollution source. The estimation of the distance from the pollution source may be based on the rate of change of the environmental parameter. The step of estimating the distance from the pollution source may comprise analysing local environmental conditions such as wind speed and/or wind direction. The method may comprise the step of sensing the local environmental conditions. The method may comprise the step of obtaining the local environmental conditions from an external data source, for example a local meteorological station. The method may comprise the step of notifying a user of the identified pollution source. The step of notifying a user may include outputting data to end users. The data output to end users may comprise a signal. The signal may be arranged to cause display of the notification for viewing by a user. The step of notifying a user may include causing display of the notification for viewing by a user. Such display may be on a display such as a screen. The screen may be on a portable electronic device such as a mobile telephone. The display may be one of an image representative of the pollution source, text, and a combination of an image and text. The step of notifying a user may include causing an audio notification, such as an audio alert.

The method may include the step of analysing the results of the comparison of the sensed data to the reference data. The method may include the step of comparing at least one sensed environmental parameter to at least one other sensed environmental parameter. That is to say, the correlation between different sensed environmental parameters can be investigated, in other words, the method may comprise determining whether one sensed parameter value rises or falls in response to or together with a change in another sensed parameter value, or whether the one sensed parameter value is independent of the other sensed parameter value. For each different environmental parameter, the method may include comparing obtained data to respective reference parameter values and determining whether the sensed parameter value is greater or less than the reference parameter value. This comparison may allow determination of whether a change in one

environmental parameter value is likely to be a 'real' change, i.e. a real physical effect, or whether this change is likely to be a 'false' change, i.e. an artefact of the sensing, and not a real physical effect. As an example, if an increase in one environmental parameter is associated with a corresponding decrease in another environmental parameter, then if the sensed one environmental parameter is seen to increase, and the sensed other environmental parameter decreases, these sensed changes may be determined to be real. On the other hand, if the sensed other environmental parameter does not change when the sensed one environmental parameter increases, the sensed behaviour of at least one of the one environmental parameter and the other environmental parameter may be determined to be false.

Such a determination may allow inaccurate data to be discarded. The method may include the step of determining whether a change in the sensed parameter value is a real change. The method may include, where it is determined that the change in the sensed parameter value is not a real change, the step of discarding the sensed parameter value. The method may include, where it is determined that the change in the sensed parameter value is a real change, the step of storing the sensed parameter value. This can help in creating a data set with a greater confidence in the data. This can lead to a more accurate identification of a pollution source.

The method may comprise the step of determining whether the difference between a sensed parameter value and a reference parameter value for one environmental parameter is a positive or negative multiple of the difference between a sensed parameter value and a reference parameter value for another environmental parameter. The method may include the step of determining whether there is a time difference in the behaviour of one sensed parameter value compared to another sensed parameter value. In other words, whether there is at least one of a time lag between changes in the values and a difference in the time period over which the changes occur.

The investigation of the correlation between two or more sensed parameter values can allow a greater confidence in the use of these sensed parameter values to identify the pollution source, and/or the proximity of the pollution source. As an example, if the behaviour of one sensed parameter is consistent with the behaviour of another sensed parameter, then it may be more likely that the behaviour is real.

According to a seventh aspect of the present invention, there is provided a method of assessing the validity of environmental data comprising the steps of analysing data obtained from a sensing means for sensing an

environmental parameter, the analysis using at least one rule which links the obtained data to validity of the obtained data, and

determining the validity of the obtained data in dependence on said analysis.

According to an eighth aspect of the present invention, there is provided a method of assessing the validity of environmental data comprising the steps of analysing data obtained from a sensing apparatus for sensing an environmental parameter, the analysis using at least one rule which links the obtained data to validity of the obtained data, and

determining the validity of the obtained data in dependence on said analysis.

Determining the validity of the data may allow the determination of whether a change in a sensed parameter value is likely to be a 'real' change, i.e. a real physical effect, or whether this change is likely to be a 'false' change, for example an artefact of the sensing, and not a real physical effect. Such a determination may allow accurate data to be stored, inaccurate data may be discarded. This can help in creating a data set with a greater confidence in the data. Subsequent analysis of this data set may therefore lead to more accurate results.

The data may comprise one or more re. The method may comprise the step of obtaining the data from the sensing means and/or apparatus. The method may include the step of sensing an environmental parameter using the sensing means and/or apparatus to obtain the data. The method may include the step of sensing a plurality of environmental parameters using the sensing means and/or apparatus to obtain the data.

Basing the analysis of the data validity on a plurality of sensed environmental parameters may increase the accuracy of the analysis.

The sensing means and/or sensing apparatus may comprise a sensor. The sensor may sense at least one environmental parameter. The sensing means and/or apparatus may comprise a plurality of sensors. Each of the plurality of sensors may sense at least one respective environmental parameter.

The plurality of sensors may comprise at least one sensor for sensing a first parameter and at least one sensor for sensing a second, different parameter. The plurality of sensors may comprise at least one sensor for sensing a parameter at a first location and at least one sensor for sensing a parameter at a second, different location.

The step of analysing the data may comprise comparing the data to reference validity data. The reference validity data may be stored data. The reference validity data may be representative of at least one characteristic of the sensed parameter.

The reference validity data may be representative of at least one characteristic of an environmental parameter other than the sensed parameter. In this way, the validity of one parameter value can be related to the behaviour of another environmental parameter. Thus, if, for example, the behaviour of two

environmental parameters is always correlated, then the validity of a change in one of these parameters can be analysed by determining whether there has been a correlated change in the other of these parameters. If there has been such a correlated change, the data may be valid, or real. If there has not been such a correlated change, the data may be invalid, or false. The method may comprise the step of obtaining the reference validity data from the sensing means. The method may comprise the step of obtaining the reference validity data from the sensing apparatus.

The method may comprise comparing a sensed parameter value with an ambient parameter value representative of the value of the sensed parameter under ambient conditions to determine a sensed difference value. The method may comprise comparing a reference validity parameter value with a reference ambient parameter value representative of the value of the reference parameter under ambient conditions to determine a reference validity difference value. The method may comprise comparing the sensed difference value and the reference validity difference value. The method may comprise determining on the basis of the comparison whether the sensed data is valid.

The method may comprise determining, where the sensed difference value is, for example, greater than a given fraction (including fractions greater than one) of the reference validity difference value, that the sensed data is not valid. The method may comprise determining, where the sensed difference value is, for example, less than or equal to a given fraction (including fractions greater than one) of the reference validity difference value, that the sensed data is valid.

As an example, consider the situation in which two environmental parameters are sensed, one being considered the 'sensed' parameter, the validity of which is analysed, and the other being considered the 'reference' parameter which is used to assess the validity of the 'sensed' parameter. Both values may be compared to ambient values for those parameters to determine difference values. The difference values may then be compared.

In the case where both environmental parameters actually increase or decrease at a similar rate to one another, from a determination that the 'sensed' parameter has increased by, say, twice the amount of the 'reference' parameter, it may be concluded that the increase in the 'sensed' parameter is not real, i.e. is not valid. In another situation, there may be cross-sensitivity between the sensing of different parameters in a sensor. This means that if one sensed parameter value increases (for example a 'reference' parameter), the sensor may be more likely to show an increase in another parameter value {for example a 'sensed' parameter) that is not real. In this case, where it is determined that the 'sensed' parameter has increased together with an increase in the 'reference' parameter, it may be concluded that the increase in the 'sensed' parameter is not real, i.e. is not valid.

The method may comprise comparing a change in one sensed parameter value with a change in another sensed parameter value (for example, a 'reference' parameter value). The comparison may include determining whether a change in one of a positive and negative direction in the one sensed parameter value is accompanied by a change in the positive or negative direction in the other sensed parameter value. The comparison may include determining whether the change in the one sensed parameter value is above or below a given multiple of the change in the other sensed parameter value.

Where the change in the one sensed parameter value does not meet a validity threshold, it may be concluded that the one sensed parameter value is not real, i.e. is not valid. The validity threshold may be that a positive change in the one sensed parameter value should be accompanied by a positive change in the other sensed parameter value. The validity threshold may be that a positive change in the one sensed parameter value should be accompanied by a negative change in the other sensed parameter value. The validity threshold may be that a change in the one sensed parameter value should be a given multiple of the other sensed parameter value, for example a multiple of at least two. Where the comparison indicates that the validity threshold is not met, the sensed data may be determined to be not valid.

The method may comprise comparing a rate of change of one sensed parameter value with a rate of change of another sensed parameter value. The comparison may include determining whether a rate of change in one of a positive and negative direction of the one sensed parameter value is accompanied by a rate of change in the positive or negative direction of the other sensed parameter value. The comparison may include determining whether the rate of change of the one sensed parameter value is above or below a given multiple of the rate of change of the other sensed parameter value.

Where the rate of change of the one sensed parameter value does not meet a validity threshold, it may be concluded that the one sensed parameter value is not real, i.e. is not valid. The validity threshold may be that a positive rate of change of the one sensed parameter value should be accompanied by a positive rate of change of the other sensed parameter value. The validity threshold may be that a positive rate of change of the one sensed parameter value should be

accompanied by a negative rate of change of the other sensed parameter value. The validity threshold may be that a rate of change of the one sensed parameter value should be a given multiple of the rate of change of the other sensed parameter value, for example a multiple of at least two. Where the comparison indicates that the validity threshold is not met, the data may be determined to be not valid.

Where it is determined that the data is not valid, for example from a comparison of the changes in different sensed parameter values, the method may comprise making a further comparison, for example between the rates of change in different sensed parameter values. This may be a further check of the validity of the data. This may allow a double-check to be made, increasing the accuracy of the validity determination, in other words, a first check may be made by comparing, for example, the change of the one sensed parameter value with the change of the other sensed parameter value, and a second check may be made by comparing, for example, the rate of change of the one sensed parameter value with the rate of change of the other sensed parameter value.

Where it is determined that the sensed parameter value is not valid, the method may comprise the step of discarding the sensed parameter value. Where, from both the first check and the second check, it is determined that the sensed parameter value is not valid, the method may comprise the step of discarding the sensed parameter value. Where it is determined that the sensed parameter value is valid, the method may comprise the step of storing the sensed parameter value This may allow false data, such as a data spike caused by, for example, sensor error, to be excluded from the data. This may lead to a more accurate set of data.

Where the sensed parameter value is discarded, the method may comprise the step of saving or storing another value instead of the sensed parameter value. The method may comprise storing a value which is a given multiple of the discarded sensed parameter value. The method may comprise the step of storing a value which is representative of an average of previous sensed parameter values for that sensed parameter.

Storing a value in place of a discarded value in this way may avoid gaps in the data, or inappropriate zero values. Such gaps or inappropriate zero values might otherwise adversely affect subsequent data analysis. According to a ninth aspect of the present invention, there is provided an environmental monitoring system comprising

means to analyse data obtained from a sensing means for sensing an environmental parameter, the analysis using at least one rule which links the obtained data to at least one pollution characteristic, and

means to determine characteristics of an indicated pollution event in dependence on said analysis.

According to a tenth aspect of the present invention, there is provided an environmental monitoring system comprising

an apparatus to analyse data obtained from a sensing apparatus for sensing an environmental parameter, the analysis using at least one rule which links the obtained data to at least one pollution characteristic, and

an apparatus to determine characteristics of an indicated pollution event in dependence on said analysis.

The system may comprise a sensing means and/or apparatus for sensing at least one environmental parameter. The system may be arranged to obtain the data from the sensing means and/or apparatus. The sensing means and/or apparatus may comprise a sensor. The sensing means and/or apparatus may comprise a plurality of sensors. Each of the plurality of sensors may sense at least one respective environmental parameter.

The data may comprise one or more reading.

The environmental monitoring system may comprise a central unit arranged to receive data from the plurality of sensors. The central unit may be arranged to receive the data over a communications network, such as the internet. The system may comprise a data store for storing data. The data store may be local to the sensing means and/or apparatus. The data store may be a database. The central unit may comprise the database. The system may comprise a server. The central unit may comprise the server. The server may be arranged to communicate with the sensing means and/or apparatus over a network. The server may be arranged to communicate with a plurality of sensing means and/or apparatuses over a network. The network may comprise at least one of a wired network and a wireless network. For example, the network may comprise at least one of a local area network, a wide area network and a connection over the internet.

The server may be remote from the sensing means and/or apparatus. For example, the server may be located geographically distant from the sensing means and/or apparatus, but connected thereto over the network. The system may comprise a processor. The server may comprise the processor. The system may comprise an analysis module. The server may comprise the analysis module.

The analysis module may be arranged to compare at least one sensed parameter value with at least one other sensed parameter value and/or with at least one stored value. The analysis module may be arranged to compare at least one sensed parameter value with at least one reference parameter value.

The analysis module may be arranged to compare the sensed parameter value with reference data, such as data representative of at least one pollution characteristic, and to determine characteristics of an indicated pollution event in dependence on said analysis. The analysis module may be arranged to determine the validity of the sensed parameter value in dependence on said analysis. The system may be arranged to store the sensed parameter value where it is determined that the sensed parameter value is valid. The system may be arranged to discard the sensed parameter value where it is determined that the sensed parameter value is not valid.

The system may comprise a timer. The timer may comprise a clock.

The system may comprise a calculation module. The calculation module may be arranged to calculate at least one of a difference between a sensed parameter value and a first reference parameter value, and a sensed parameter value and an ambient parameter value. The calculation module may be arranged to calculate a difference between a first sensed parameter value and a second sensed parameter value. The calculation module may be arranged to calculate a rate of change of the sensed parameter value, for example by using timing signals from the timer.

The central unit may be arranged to determine a recommended action to be taken in response to the data, such as readings, received from the plurality of sensors. The method may comprise determining a recommended action to be taken in response to the data, such as readings, received from the plurality of sensors. The system may be arranged to determine a recommended action to be taken in response to the data, such as readings, received from the plurality of sensors.

The method or system may be arranged to output data to end users determined on the basis of the data, such as readings, received from the plurality of sensors. The system may comprise an output means and/or an output apparatus arranged to output data to end users. The central unit may be arranged to output data to end users determined on the basis of the readings received from the plurality of sensors. The method may comprise outputting data to end users determined on the basis of the readings received from the plurality of sensors The data output to end users may comprise a signal. The data may be arranged to cause display of a notification for viewing by a user. The data may be arranged to cause display of the notification on a display such as a screen. The data may be arranged to cause an audio notification such as an audio alert.

According to an eleventh aspect of the present invention, there is provided an environmental monitoring system comprising

means for analysing data obtained from a sensing means for sensing an environmental parameter, the analysis using at least one rule which links the obtained data to validity of the obtained data, and

means for determining the validity of the obtained data in dependence on said analysis.

According to a twelfth aspect of the present invention, there is provided an environmental monitoring system comprising

an apparatus for analysing data obtained from a sensing apparatus for sensing an environmental parameter, the analysis using at least one rule which links the obtained data to validity of the obtained data, and

an apparatus for determining the validity of the obtained data in dependence on said analysis.

The environmental monitoring system according to any aspect or combination of aspects above may be arranged for carrying out any method or combination of methods above.

Features of any aspect above may be provided with any other aspect. Apparatus features may be provided with the method aspect and vice versa. These features have not been written out in full merely for the sake of brevity. Preferred features of the present invention will now be described, by way of example only, with reference to the accompanying drawings in which:

Figure 1 is a block diagram of a sensing unit;

Figure 2 is a schematic illustration of a system for analysing environmental data; Figure 3 is another schematic illustration of a system for analysing environmental data;

Figure 4 is a flow chart representing a method for analysing environmental data; Figure 5 shows the time variation of levels of VOCs and N0 ₂ for (a) an outdoor combustion event, and (b) an indoor solvent event;

Figure 6 is a flow chart representing another method for analysing environmental data; and

Figure 7 is a graphical representation of the interrelated behaviour of

environmental parameters.

A system for analysing environmental data comprises a plurality of sensors for sensing and measuring a plurality of different environmental parameters, for example pollution of a variety of pollution types. Referring to Figure 1 , in one embodiment a sensing unit 100 comprises a plurality of sensors 1 1 A within a sensing section 1 1 . in the description that follows, certain outdoors environmental parameters are discussed. It is noted however that the sensing unit 100 can equally be used to sense indoors as well as outdoors environmental parameters. Additionally the sensing unit 100 can in some embodiments comprise sensors 1 1A which include both indoor and outdoor sensors.

The sensors 1 1 A comprise air pollution sensors, for sensing gasses, such as:

nitrogen dioxide (N0 ₂ ), nitrogen oxide (NO), carbon monoxide (CO), sulfur dioxide (S0 ₂ ), ozone (0 ₃ ), and Volatile Organic Compounds (VOC), and for sensing particles by size, such as: 10 micron, 2.5 micron and nanometre scale particles. It would be appreciated that other air pollutants may be measured in addition or alternatively. In some embodiments, the sensors 1 1 A further comprise at least one of noise sensors, radiation sensors, water pollution sensors and atmospheric condition sensors for sensing ambient temperature, barometric pressure, relative humidity, UV radiation (of preferably all types) and other or additional atmospheric conditions. It would be appreciated by those skilled in the art that other or additional pollution sensors may be used, it would be appreciated by those skilled in the art that when the sensing unit 100 is adapted for indoor use, the sensors 1 1A are designed to operate and monitor pollution in indoor environments. The sensing capabilities may target indoor pollutants such as tobacco smoke, formaldehyde, volatile organic compounds, radiation etc. The sensors 1 1 A may be designed to work in common indoor spaces such as homes, offices, public indoor spaces etc. In such a configuration, the communication of the sensing unit 100 over a communication network such as the Internet may be based on commonly available Internet accesses points, both wire-based (copper or optical or both) and/or wireless. It would be further appreciated that when the sensing unit 100 is adapted for operation as an outdoor sensing unit, the sensors 1 1 A are designed to work in and monitor the outdoor environment. The sensing capabilities may target common outdoor pollution factors such as outdoor air pollution, noise, radiation etc. The sensing unit 100 further comprises a data acquisition and processing unit 12. The acquisition and processing unit 12 performs the sensors' sampling and device general management. However, at least some of the sensors include signal sampling and communication capabilities. The acquisition and processing unit 12 comprises an analog to digital converter 18, to translate analog readings of the sensors 1 1A to digital format, a real time clock 17 for time tagging, for example of signal sampling of the sensors 1 1A, a memory unit 19 to store the device operating software, system and / or user changeable parameters, and sensor readings, a microprocessor 16 for managing the device and other components needed for computing functionality, as known in the art.

As illustrated in Fig. 1 , the sensing unit 100 further comprises a communication unit 13. The communication unit 13 is adapted to communicate with external devices and networks, such as USB channels, telephone channels (land line and / or cellular) and network channels. The communication unit 13 further provides connectivity of the sensing unit 100 to a communication network such as the

Internet, and any other software and hardware known in the art which is required in order to allow communication between the sensing unit 100 and a communication network. The sensing unit 100 further comprises a power unit 14 for delivering electrical power to each of the other units of the sensing unit 100. in this configuration, the power unit 14 comprises an energy storage component such as a battery or a capacitor. The energy storage component is rechargeable. The power unit is adapted to provide the required power to the units of the sensing unit 100.

According to some embodiments, the power unit 14 is further adapted to be powered by an external power source 15, such as grid power or an energy scavenging device adapted to charge the energy storage component. According to some embodiments, the external power source 15 is a solar panel, wind or vibration electricity generation device etc.

Reference is now made to Fig. 2 which is a schematic illustration of a personal / local pollution monitoring system 200. The monitoring system 200 comprises at least one outdoor sensing unit 21 , at least one indoor sensing unit 24, a

communication network, such as the Internet 21 1 , and interfacing means thereto, such as an Internet hub 27.

The monitoring system 200 further comprises a server 214 and a database 217 to allow data accumulation from the at least one outdoors sensing unit 21 and the at least one indoors sensing unit 24 and to store the data, analyse the data and circulate information to other end users.

As may be seen in Fig. 2, the at least one outdoor sensing unit 21 , such as the sensing unit 10 described in Fig. 1 above, operates at the geographical close vicinity of an indoor space 218. The outdoor sensing unit 21 is in active

communication with the indoor sensing unit 24 and/or in active communication with the server 214 in bidirectional communication channels 22, 23. The outdoor sensing unit 21 is arranged to receive information originating from the server 214, such as time synchronization messages and software updates or other special inquiries from the server 214. The sensing unit 21 is arranged to send data such as a sensor unit ID, pollution readings (e.g. levels of Ozone, Nitrogen Dioxide, noise, RF radiation and the like), real time / reading time clock, location and meteorological conditions (e.g. temperature, humidity and the like). The data from the outdoor sensing unit 21 is sent directly to the server 214 via a hub 27 and a communication network 21 1 or through a mediator device that is, according to some embodiments, the indoor sensing unit 24. According to some embodiments, the mediator device, such as the indoor sensing unit 24, is connected to a dedicated or non-dedicated

communication device 28, which in turn may be connected to the communication network 21 1 via the hub 27. The communication device 28 is, in some

embodiments, one of a Personal Computer (PC), a Personal Digital Assistant (PDA), a telephone, a mobile phone such as a smart phone, or any other communication device capable of communicating with another communication device via a communication network such as the internet.

The indoor sensing unit 24, placed in the indoor space 218, is arranged to communicate bidirectionally (25, 26) with the server 214 similarly to the outdoor sensing unit 21 (receiving time synchronization messages and sending sensor readings and other information to the server 214). In addition, the indoor sensing unit 24 serves as communication mediator to the outdoor sensing unit 21 as explained above. It would be appreciated by those skilled in the art that the communication between ail or some of the above elements (i.e. the outdoor sensing unit 21 , the indoor sensing unit 24 and the server 214) may be by wired and/or wireless communication.

The hub 27 is, in this configuration, a conventional gateway to the communication network 21 1 . It can connect either through a cable connection and a modem or wireless internet hub or via any other communication channel known in the art. The data from and to the sensing units 21 and 24 is routed through the hub 27.

The communication device 28 is embodied in a computer, tablet computer or the like, in which case it serves as a user interface (Ul) to the system, presenting the retrieved and processed data from the sensors and enabling operation of the system features. The user interface may be dedicated software installed on the communication device 28 or may be a website accessible from the communication device 28.

According to some embodiments, the communication device 28 takes part in the communication routing between the indoor sensing unit 24 and the server 214 and/or between the outdoor sensing unit 21 and the server 214. The sensing units 21 and 24 are connected to the communication device 28 via any connection means known in the art, such as by a Universal Serial Bus (USB) connection. It would be appreciated that other means of connection may be used. As may be seen in Fig. 2, the server 214 is arranged to receive data from the sensing units 21 and / or 24 and send data messages to the sensing units 21 and 24. The data sent to the sensing units 21 and 24 includes time synchronization messages, software updates, instructions and other software-based functionally targeting elements (e.g. sensor sampling scheduling).

According to some embodiments the server 214 is arranged to receive raw data from the sensing units 21 and 24, atmospheric conditions data and pollution data from sources that are not received via the system's sensing units, such as municipal monitoring units, university research units, etc. The server 214 is arranged to use calibration tables and mathematical functions (group theorem etc.) to modify raw data received from the sensors to obtain more accurate data or data modified for purposes such as trending, long-term data cumulating and others. For example, accuracy of readings from some sensors may depend on the relative humidity and ambient temperature at the vicinity of those sensors (e.g. metal oxide gas sensors).

The server 214 is arranged to adjust the sensor readings based on the sensor ID (that may contain the sensor's part number or any other unique ID data) and calibration data that may be provided by the sensor manufacturer. Performing this function at the server 214 (and not on the sensor itself) is possible because of the overall system design; data from the sensors is obtainable by the users after it has been processed at the server 214. The importance of this feature is in reducing the complexity of the sensing units 21 and 24, hardware and software, therefore reducing their cost of manufacturing. The last is important to support large scale deployment of the sensing units 21 and 24, and especially deployment in areas and suburbs where low-income populations reside.

According to some embodiments, when a new sensing unit 21 or 24 is added to the system and new entry to the database is created, the server software performs a search to learn what other sensors exist in the geographical vicinity of the new sensing unit 21 or 24, based on range definitions. The server 214 may use all available geospatial information retrieved from both neighbouring sensing units 21 and 24 and from information in the public domain, such as municipal air monitoring stations, pollution sources, pollution next to roads and transportation lines.

According to one embodiment, the server 214 registers all sensing units within a distance Z (i.e. neighbouring sensing units), thus creating a cluster of sensors. The server 214 may further register nearby air quality monitoring stations that are not part of the system's sensing units. These stations may include, beside pollution sensors, also meteorological sensors like wind speed and direction. For example, in the USA it would be the Environmental Protection Agency (EPA) network called "Airnow".

The server 214 is further arranged to register pollution sources inventory like: power plants, chemical production plants or refineries and nearby transportation routes such as highways, ports and the like. The server 214 is further arranged to register the distance and relative direction of these known pollution sources from the new sensing unit. In many countries and jurisdictions this information is in the public domain.

It would be appreciated that the information received from neighbouring sensing units, from other quality monitoring stations and from the public domain may provide information for adjusting the data received from a new sensing unit and for calibrating the new sensing unit.

According to some embodiments, the server 214 is in active communication with a storage unit storing a database 217. The data received from the sensing units 21 and 24 is stored in the database 217. According to some embodiments, the database 217 includes data and information received from other sources. According to some embodiments, the database 217 saves a lookup table including sensed pollutants rates and recommended reactions to reduce hazard.

Computations performed on the server 214 may extract pollution information from the sensed data and search for the recommended countermeasure to the pollution as reflected in the extracted information and may send a recommendation of these countermeasures to end users. FIG. 3 is high level overall data aggregation and distribution scheme of the system 300 for personal pollution monitoring based on multitude-points ambient quality data. As may be seen in Fig. 3, a plurality of pollution sensing units 31 are located in a plurality of geographical locations. The geographical locations may be indoor and/or outdoor locations. The plurality of sensing units 31 are arranged to send pollution data 32, over a communication network 33 such as the Internet, to a centralised server 34. in one configuration, data sent from the sensing units 31 is based on sampling periods and includes: unit ID, pollution sensors readings, real time clock value, location of the sensing unit and meteorological conditions (e.g. temperature, humidity and the like) in the vicinity of the sensing unit. The sensing units 31 transmit the data to the server 34 via one or more communication networks 33, such as the Internet. The data may reach the communication network 33 through an indoor internet hub (not shown), either wirelessly or by wire. The sensing units 31 are installed by users at their living places (homes, offices, kindergartens, schools, vehicles etc.) and therefore are designed to connect to the Internet by commonly available Internet connection devices. The sensing units 31 are adapted to allow automatic initialization and connection of each sensing unit to the network 33 (i.e. plug and play capabilities), thus allowing each sensing unit 31 to serve as an independent entity capable of communicating with other sensing units and with the server 34 to send and receive data therebetween. Furthermore, the connection of each sensing unit 31 to the communication network 33 allows each sensing unit 31 to be accessed via the network 33.

It would be appreciated by those skilled in the art that each sensing unit's plug and play capabilities allow easy installation by end users and facilitate wide deployment of the sensing units 31 .

The location of each sensing unit 31 may be determined by one or more of the following methods: GPS receiver (which may be part of, or together with, the sensing unit 31 ); wireless communication trianguiation methods (cellular or Wi-Fi); manually by the user or at the assembly line based on the user delivery address. The last enables cost reduction of the sensing devices and supports large scale deployment. According to one ernbodiment, the sensing unit 31 has a Wi-Fi module that is pre-set to peer to peer mode, to allow it to communicate directly with a Smart phone or any other wirelessly connectable device, via the Wi-Fi channel. Once the peer to peer connection with the Smart phone is established, the Smart phone may transmit data to the sensing unit 31 , such as the security information needed to connect to the home wireless router and get out to the Internet (SSID) and the location data of the Smart phone.

The server 34 comprises a processor (not shown) and storage means (not shown) for processing and storing and retrieving data received from the sensing units 31. The processed data may then be distributed to end-users 35, 38 and 37. The processed data may be distributed to the system in the form of various services via the Internet and cellular networks.

The processed data may be sent to users' mobile devices, personal computers, PDA's, laptops, tablet computers, mobile phones and the like and/or to business servers. For each end user specialised information formats / message may be used because of the differentiation both in need and communication method. For instance, when the processed data is sent to an end user's cellular devices the information may be sent over the cellular network or other communication methods to mobile devices (e.g. Wi-Fi, Wi-Max etc.). The information sent may include the information needed to support the system's mobile device features. Alternatively or additionally, information may be sent over the internet and presented on the system webpage to support the system internet features. The information may be viewed and used by any conventional browsing technique both from stationary devices and mobile devices. The information may be further sent to other businesses clients over the internet, for example pollution mapping facilities, or sent to a weather news Company. Other interface options may include installing custom-made software at the user end device to receive the system information, such as via a specialised application on a smart phone.

A pollution source may be identified by linking the characteristics of a sensed environmental parameter to a signature or profile for the pollution source. The characteristics of an environmental parameter which may lead to an identification of a pollution source include the sensed value of the environmental parameter itself, a change in the sensed value of the environmental parameter and a rate of change of the sensed value of the environmental parameter. Reference is now made to FIG. 4 which is a flowchart of a rule based pollution source identification method, in block 410 a set of rules linking interrelations between readings of a plurality of pollution sensors and pollution characteristics of a pollution event are generated. For example, the rules may link interrelation between readings of a plurality of pollution sensors to a possible pollution source of the pollution event.

The variables for the rules may include measured levels of pollutants, and parameters derived by various mathematical operations and signal analysis tools applied to the measured level of pollutants. These parameters may include measured values of pollutants (such as a sensed parameter value), variations of measured values of pollutants over a time period (such as changes in a sensed parameter value), time derivatives of measured values of pollutants, for example rate of change, integral of measured values of pollutants over a time period, etc (such as rate of change of a sensed parameter value). For example, for a measured pollutant X, the variables may include the level of pollutant X over time t denoted as LJJ), or more simply L _x , the difference between the pollutant level at two different times t ₂ and t _h referred to as level of change and denoted L _x (t), or more simply AL _X . where AL _x (t = L _x (t ₂ - and the time derivative of the pollutant level which is the rate of level change, denoted (t)). In addition, the variables may include relations between measured levels of two pollutants. For example, for pollutants X and Y, the variables may include the measured level of pollutant X divided by the measured level of pollutant Y, denoted L _x / L _y , Additionally, signal analysis methods may be used to extract more information from the measured pollutant levels. For example, sound processing algorithms may be utilized to extract more information regarding the pollution event from the measured sound levels. It is possible to include a sound classification capability, whereby a determination can be made to distinguish between, for example, the sound of a lorry/truck and a car. This determination can provide additional information on the nature of the pollution source. In other words, where the type of vehicle can be determined, the engine associated with that vehicle type can be taken into account when analysing the data and predicting future pollution profiles and/or warning users. Each rule may relate variables of a plurality of pollution sensors to a source of the pollution event. Typically, the plurality of different sensors may include different types of sensors for measuring different types of pollutants. Therefore, the rules may relate measured levels of different types of pollutants, or interrelations between these measured levels of pollutants to a source or sources of the pollution event.

The rules may be generated based on prior knowledge of typical relations between levels of various pollutants and pollution sources

The linking of various characteristics of a sensed environmental parameter to pollution sources can require knowledge of ambient conditions in the location of the sensor, for example an ambient value of a particular environmental parameter, and/or knowledge of threshold values indicating pollution events.

If it is known that a nitrogen dioxide concentration above a certain value is indicative of a pollution source or a pollution event then that value can be set as a threshold value (denoted as L _x - _th ) against which a sensed nitrogen dioxide concentration can be compared. Therefore, if the sensed nitrogen dioxide concentration is below this threshold value, it may be determined that no pollution event has occurred. If the sensed nitrogen dioxide concentration is above the threshold value (i.e. if L _{N02 >} L _N0 ^ _th ) it may be determined that a pollution event has occurred. Similar comparisons and determinations can be made for other sensed

environmental parameters.

This combination of determinations for different environmental parameters allows a more detailed profile of the pollution event to be constructed. This profile allows a more accurate identification of the pollution source.

An example of a partial set of rules linking excess levels of various pollutants to possible sources of pollution is presented in Table 1 below. The measured pollutants presented in Table 1 are N0 ₂ , VOC, sound, Fine Dust / smoke and S0 ₂ . Each row represents a rule, and each column represents a variable related to a pollutant, as indicated in the first row. A + (plus) sign represents an excess level of the corresponding pollutant during a pollution event, for example a measured level that crosses a threshold (note that this could be a sensed value that is greater than a threshold for an environmental parameter that is normally low, or it could be a sensed value that is less than a threshold for an environmental parameter that is normally high) and a - (minus) sign represents a normal level (i.e. a level that has not crossed a threshold) of the corresponding pollutant during the pollution event.

For example, the rule presented in the first row of Table 1 may be expressed as: if L _NQ2 > L _NQ2 _ _T , L _VOC > L _VQC _ _TH , L _SOUND > L _SOUND _ _T , L _DUST ≤ L _DUST _ _T and o-, ≤ ^L so ₂ -tn _> hen a pollution event has occurred, and that event can be identified as combustion from a nearby motor vehicle, in other words, the pollution source can be identified or classified based on a comparison of sensor readings. It will be seen that where different sensed parameter values cross respective threshold values, corresponding to different conditions in the vicinity of a sensor, this allows the identification of different pollution sources.

Thus, an indication of the pollution source may be given based on interrelation between data gathered by the plurality of different pollution sensors.

Table 1 : Examples for source identifications based on readings of a plurality of pollution sensors

The rule may base the comparison on an ambient value rather than on a threshold value. Once an ambient value is known, a reference value for that environmental parameter can be determined. This can be denoted as L _x _ _am . This may be set at a given multiple or fraction of the ambient environmental parameter value, or it may be set at an average value of the ambient environmental parameter, for example the average value over a given preceding time period.

It may be that rather than simply crossing a threshold value, the value of an environmental parameter may need to change by a certain amount from an ambient value to be indicative of a pollution source or pollution event. For example, in various different environments the ambient value for a particular environmental parameter might differ significantly. Thus a single threshold value for this parameter would not provide a useful indication of the presence or absence of a pollution event.

However, it may be that a change in the sensed environmental parameter by a certain amount from the ambient value is indicative of a pollution event. As an example, irrespective of whether the ambient value of an environmental parameter is 4 or 8 (arbitrary units), an increase of at least 5 (again, arbitrary units) could indicate a pollution event. Thus, where the difference between a sensed environmental parameter value and the ambient value (denoted as AL _X ) is above a threshold change (denoted as AL _x _ _th ), this would then be indicative of a pollution event. As an example, if L _SOz > AL _SOz _ _th , then a pollution event has occurred, and that event can be identified by a comparison with pollution profiles in a similar manner to that mentioned above with reference to Table 1 .

Alternatively or additionally, the rate of change of a sensed environmental parameter value may be used to identify a pollution event, in some situations it may be that a particular environmental parameter value can change, within a range of values, without this being indicative of a pollution event. The particular upper and lower boundaries of the variation may not be precisely defined.

Therefore it may not in this case be possible to define either a reference value for this environmental parameter beyond which a pollution event can be determined, nor may it be possible to specify a change from an ambient value which would also indicate a pollution event.

It may be possible, for example in the case mentioned in the preceding paragraph, to use a rate of change of the sensed environmental parameter value to

determine the occurrence of a pollution event, and identify the pollution source. The rate of change of the sensed environmental parameter value may be calculated instantaneously in a known way. The rate of change of the sensed environmental parameter value may be determined by sensing the environmental parameter at a first time, and sensing the environmental parameter at a later, second time. The rate of change may then be calculated using the difference between the sensed values and the difference between the first and second times.

Determining the rate of change of the environmental parameter value in this way allows the rate of change to be determined over a given time period. The selection of the time period may be chosen such that noisy fluctuations in the sensed signal may be smoothed out. The selection of the time period may be chosen such that real short term changes are not lost in a longer time period. The appropriate time period over which the rate of change is calculated is likely to differ between different environmental parameters, and in different environments.

As an example, the concentration of ozone is likely to fluctuate naturally over a given time period and the rate of change of these fluctuations may be considered to lie within what is expected in a normal environment. However, if the rate of change of the ozone concentration (denoted as ~ (i ₀ J) s calculated to be above a threshold rate of change (denoted as ~ (L ₀ „_ _tft )), then this may be indicative of a pollution event. This may be the case if the ozone concentration suddenly fails to zero. In other words, if there is a large negative rate of change in the ozone concentration. This may be indicative of, for example, a fire in the sensed environment which might consume the ozone.

The pollution source may be identified on one of or a combination of more than one of the sensed parameter value exceeding a threshold, the sensed parameter value differing from an ambient by more than a threshold change or the sensed parameter value changing by more than a threshold rate of change. The combination of these may lead to a more accurate identification of the pollution source, For instance, if the ozone concentration drops below a particular threshold value, this may be indicative of a combustion engine near the sensor (this is because the combustion engine will emit nitrogen oxide which will react in the atmosphere so as to reduce the concentration of ozone). However, if this is combined with the rate of change of the ozone concentration and this is seen to experience a high negative rate of change then this may be indicative of a more serious event such as a fire in the sensor surroundings. it is also possible to further increase the accuracy of the pollution source identification by combining data from more than one environmental parameter. For example, an increase above a threshold or by above a threshold change in the concentration of nitrogen dioxide may indicate that a combustion engine is a pollution source, if there is, at the same time as an increase in the nitrogen dioxide concentration, an increase in the noise level then this may increase the confidence in the identification of the combustion engine as the source of pollution.

Different attributes of the sensed environmental parameters can also be compared as part of the analysis to increase the accuracy of the pollution source identification. For example, the comparison can include determining whether the sensed values have crossed respective thresholds, and whether the rates of change of the sensed values have also crossed respective thresholds.

An example of a rule linking the interrelation between the readings of two sensors, one measuring the level of pollutant X and the other measuring

the level of pollutant Y, to pollution sources is given below.

If AL _X > &L

and M _y > My- _th

over a given time period t ₂ - t then "Value A"; else if AL _X > AL _x _ _th

and AL _y « AL _y _ _th and (L _¾ )≠£ (L _y )

over a given time period t ₂ ~ ,

then "Value B". where "Value A" and "Value B" may represent a value of a pollution

characteristic, such as a pollution source, in the above, ≠ f~(A _y )

may be considered to be— (L _x ) »— (L _y ) and/or— (L _x ) «— (L _v ).

A combination of different sensed environmental parameters may also be used to determine the proximity of a pollution source to the sensor, As an example, if the nitrogen dioxide concentration sensed by a sensor increases and the noise level sensed by the sensor does not increase this may be indicative of a combustion engine being the source of the pollution but that the combustion engine is not close to the sensor, in other words, nitrogen dioxide produced by the combustion engine may have moved from the source to the sensor such that it is detected but the source may be sufficiently far from the sensor that there is no increase in the noise sensed.

The rate of change of an environmental parameter may also be used to determine the proximity of the source to the sensor. A spike in the data, in other words a high positive rate of change in the sensed environmental parameter, optionally followed by a high negative rate of change in the same parameter, may indicate that the source is close to the sensor. On the other hand, a much slower increase, i.e. a lower positive rate of change of the sensed parameter, optionally followed by a slower decrease, i.e. a lower negative rate of change of the parameter, may indicate that the pollution source is spaced from the sensor. The calculated rate of change of the sensed environmental parameter can also give a quantitative indication of the proximity of the pollution source. The further away from the source the sensor is located the greater the spreading of the spike in the data, in other words the lower the rate of change of the sensed environmental parameter. The proximity of a source may be empirically determined for various rates of change of that sensed parameter and this may be used to calibrate the proximity calculations,

The sensed environmental parameter can also be used to determine or predict the intensity or severity of a pollution event. As mentioned above, if it is determined that a sensed environmental parameter has an extreme value {such as a very high value or zero) then this can be indicative of an event in that area.

Thus if, for example, the sensed ozone concentration in the vicinity of a particular sensor drops to zero, this might be indicative of an event such as a fire in that area. For a network of connected sensors, if a geographically neighbouring sensor also senses that the concentration of ozone has dropped to zero then this might indicate a more severe event, i.e. one that is geographically spread. The geographical scope of the sensors that detect a similar drop to zero in the ozone concentration will give an indication of the geographical spread and severity of the event.

Also, for a single sensor, it may be that the ozone concentration only momentarily drops to zero or that the ozone concentration drops to zero for a much longer period. Where the ozone concentration drops to zero for a much longer period this may be indicative of a more intense event. In this context, intense might mean a much more serious event over a short time period which causes the environmental parameter to take longer to recover, or it may be a relatively less serious event but over a longer time period. It is possible, based on empirical data or otherwise, to set a threshold time such that if the ozone concentration is sensed to drop to zero, or below a particular threshold concentration, for longer than a threshold time then this would be indicative of an intense event.

The sensing of the environmental parameters in this way can also be used to predict the presence of other pollutants in the area of the sensor or in the proximity of the sensor, or in the area or proximity of the pollution source (which as described above may be spaced from the sensor). This prediction can be made whether or not such other pollutant itself can be sensed by the sensor. The ability to do this thus increases the usefulness of the sensor and the data obtained. As an example of this, if it is determined that a pollution event has occurred and that the source of the pollution is a combustion engine, i.e. that the sensor has sensed vehicle emissions, then it may be predicted that benzene will also be present in that area since it is known that benzene is typically emitted together with other vehicle emissions. Using the values of, for example, the sensed nitrogen dioxide and noise, the concentration of benzene can be quantitatively estimated.

Thus, even for a sensor not calibrated or able to directly detect an environmental parameter such as benzene, the sensor can still be used to make quantitative determinations about the value of that parameter at a given time and/or in a given c! GcL

In another way of predicting the severity of a pollution event, consider two geographically spaced sensors, if a person adjacent one of the sensors starts smoking then that sensor might sense a change in an environmental parameter that is both well above a threshold and occurs very quickly, i.e. with a high rate of change. This may lead to the conclusion that the pollution event is serious in the vicinity of that sensor. However, for the geographically spaced sensor, the same environmental parameter will behave very differently. Depending on the distance from the first sensor the second sensor may sense a slight increase in the value of the environmental parameter at a much lower rate of change. Thus, if what appears to be a serious event is only sensed by a single sensor in a network of sensors, and relatively geographically close sensors to that sensor do not detect such a severe event then it can be concluded that the event is very localised and/or not severe.

If, on the other hand, a number of spaced sensors ail individually determine the presence of a severe event, then this might be indicative of a more serious event such as a fire. In the event that the determination is made that a serious event, such as a fire, has been detected, the present system may cause or allow a notification of the event to be made to, for example, the local authorities. Such an early notification may reduce the time and cost of dealing with the event. The present system may also issue a public alert in the event of predetermined serious events occurring. Thus people can avoid the area subject to the alert, avoiding or reducing the risk of harm from the event.

The geographically spread network of sensors can also provide advance warning of a pollution event in a particular area. If a particular environmental parameter is sensed as increasing by one sensor at a particular time, and as increasing by another sensor at a later time, then the system can determine the speed of movement of the pollution (or pollution source) within the area over which the sensors are spread.

As an example, if a static pollution source such as a factory emits pollution, this may be first detected by a sensor near the factory. The plume of pollution may then spread over a larger area and the presence of this plume may be detected by subsequent geographically spaced sensors. Knowledge of the movement of the plume and of local wind speed and direction can enable an advance warning to be issued for areas outside the extent of the pollution plume at that time. Thus, warnings can be issued in respect of areas in which sensors will not yet have recorded that a pollution event has occurred. This advance warning may provide advance notice such that preventative action can be taken to avoid exposure to the pollution. Such action may include closing windows, and remaining indoors.

It is also useful to determine whether changes in two or more different

environmental parameters are related to the same pollution source. This may help both in determining the pollution source itself and in identifying the movement of the pollution and determining potential avoidance action. Two or more environmental parameters can be either positively or negatively correlated. In other words, an increase in one parameter may be linked to an increase or decrease in the other parameter. The increase or decrease in the other parameter may be spaced in time from the change in the first parameter or it may be concurrent with the change in the first parameter, in other words, there may be a time lag between the changes in the different environmental parameters.

For a pollution source, P, in which nitrogen dioxide formation takes more time relative to the production of volatile organic compounds, i.e. the local level of VOCs will increase before the local level of nitrogen dioxide, the following rule may be used.

If (ljyQ ₂ (£ ₂ ) > ½o ₂ -£ft) and (Lyocih) > L _V oc-th) , * ^hen * ^he pollution source is P.

Here, t ₂ is a later time than ti. Furthermore, the distance of the pollution source may be estimated, inter alia, based on the difference between t ₂ and t

Hence, in this example, if the level of VOCs exceeds the relevant threshold, followed some time thereafter by the level of N0 ₂ exceeding the relevant threshold, then it can be determined that the source of pollution is P.

In another example, nitrogen dioxide formation may occur at a similar time to the production of VOCs, This may be due to a combustion event - see Figure 5a, showing the time variation of levels of VOC and N0 ₂ for an outdoor combustion event. Thus, if the levels of VOCs and N0 ₂ exceed their respective thresholds, then it can be determined that the pollution source is a combustion engine.

On the other hand, Figure 5b shows the time variation of levels of VOCs and N0 ₂ which are due to an indoor solvent event. In this example, if the level of VOCs is determined to exceed its threshold, and the level of N0 ₂ does not change (i.e. it does not exceed its threshold, and/or the rate of change of NO? stays below a threshold rate of change), then it can be determined that the pollution source is a solvent event.

As mentioned above, the ozone concentration may be linked to the emission of nitrogen oxide such as by a combustion engine, or it may be linked to another event such as a fire. In predicting the nature and severity of the event, it is useful to know which of these situations is the case. Thus, the change in the nitrogen dioxide concentration (which is related to the change in the nitrogen oxide concentration) can be compared to the change in the ozone concentration, and the rate of change of the nitrogen dioxide concentration can be compared to the rate of change of the ozone concentration. This may lead to the following relationships. and (L _w¾ ) = - fc ₂ (L ₀₈ )

In the above relationships, the calculated constants k ₁ and k ₂ can be compared with empirically derived values to identify whether any changes in the nitrogen dioxide concentration and ozone concentration are related to the same event.

For the combustion engine example, the emission of nitrogen oxide will cause chemical reactions in the atmosphere which produce an increase in the concentration of nitrogen dioxide and a decrease in the concentration of ozone. The rate of increase and the rate of decrease of these respective concentrations will also be linked. It has been found that there is a time lag between the change in these concentrations such that the sensed decrease in the ozone

concentration lags behind the sensed increase in the nitrogen dioxide

concentration. Thus, if it is found that the sensed ozone concentration does indeed decrease shortly after an increase in the sensed nitrogen dioxide concentration and the rate of change of each is within a given range then it can be concluded that a combustion engine is the source of the pollution that has been sensed. it is also noted that the correlation of the sensed environmental parameters can be used to validate the data from the sensor. Thus in the above example relating to a combustion engine if the nitrogen dioxide concentration increases, and there is no corresponding decrease in the ozone concentration, then it can be concluded that either the nitrogen dioxide reading or the ozone concentration reading (or both) are incorrect, in this case further validation steps can be taken to identify which reading is incorrect.

The sensors might experience cross-sensitivity. This means that the readings for one environmental parameter produced by the sensor may depend on the behaviour of another environmental parameter. As an example, a sensor may be sensitive to humidity such that when humidity increases this can cause an increase in the reading for another parameter, where the other parameter does not actually increase in value. In other words, the change in humidity can produce false readings of other parameters. For example, an increase in humidity can lead to a false increase in the reading for nitrogen dioxide, Thus the nitrogen dioxide reading might be seen to increase at the same time as rain starts falling. Comparing rainfall timing with the nitrogen dioxide readings can help identify false readings, and lead to an increase in the accuracy of the data obtained.

Also, the readings produced can be checked against expected patterns or profiles. As an example, it is noted that in cities nitrogen dioxide levels are expected to increase in the morning and evening rush hours, and relax after these peak times. Thus if the readings follow such general trends, the data can be treated with a greater confidence. On the other hand, if the readings deviate from these trends without apparent reason, the data might be false and further checks might need to be made before these data can be relied upon. This serves to further increase the accuracy of the data, and its reliability.

This approach of assessing the validity of the data, and of identifying false readings, enables the use of sensors that might otherwise be considered to be too inaccurate or inconsistent. Thus cheaper sensors, which might be considered to be more error-prone, can be effectively used as part of a system as described herein whilst maintaining data quality. This can allow a greater density of sensors to be provided across a given area. Such an increase in sensor density can lead to more accurate data on a local scale - since model-based interpolation may not need to be relied upon. This can help avoid the situation where a sparse network is used to gather data, where model-based interpolation is needed, in which an event close to one sensor can have a disproportionate effect on the model, and hence on the analysis of the data.

Referring again to Fig. 4, in block 420 readings of the plurality of pollution sensors are obtained; the readings relate to a pollution event. The readings may be obtained as a time series of measured values. In block 430 the readings are analyzed based on the set of predetermined rules. Variables for the rules may be calculated and the predefined rules may be applied. In block 440 characteristics of the pollution event may be determined based on the analysis. For example, the source of the pollution event may be identified based on the rule based analysis of the readings of the pollution sensors. In block 450 a user may be notified of the characteristics, for example the source, of the pollution event. Additionally, an alert may be sent, for example using a communication system, to a central location, such as centralized server 34, where data may be gathered from a plurality of such rule based systems or other sources of environmental information.

Additionally or alternatively, other actions may be taken based on the pollution source. For example, as mentioned above, an indication may be given of the possible presence of pollutants that are not measured by the system, but result from the indicated source of pollution. The level of these pollutants may be estimated and indicated to a user. The pollution source may be correlated for further verification and for a more accurate identification of the pollution source, with the inventory of pollution sources that may be registered, for example, at server 214 or centralized server 34.

Some of the pollution sensors used such as in the sensing unit 100 may produce noisy signals that may include spikes that do not result from real increase in the quantity of the measured pollutant. These spikes may therefore be considered as noise. The present techniques provide a method, as introduced above, to assess the validity of the signal, for example to indicate whether a measured increase in a level of a pollutant is true or false. The discussion below will treat the more general case.

Reference is now made to FIG. 6 which is a flowchart of a rule based method for assessing validity of environmental data. This method may be seen as a rule based smart filter for filtering out false spikes in readings of pollution sensors. In block 510 a set of rules linking the interrelation between readings of a plurality of pollution sensors and validity of the readings is generated. The variables for the rules may include similar variables as described hereinabove with relation to block 410 of Fig. 4. Typically, the plurality of different sensors may include different types of sensors for measuring different types of pollutants. Therefore, the rules may relate measured levels of different types of pollutants, or the interrelations between these measured levels of pollutants to validity of the measurements. The rules may be based on prior knowledge of interrelation between pollutant levels. For example, if an increase in a level of pollutant X is known to be

necessarily followed by an increase in the level of pollutant Y then a rule for determining validity of a measured spike in pollutant X may be:

If (ti) >

and L _y (t ₂ ) >

then the spike is true,

else the spike is false.

Here, f _? is the time, or range of times, at which a spike of pollutant X is sensed, and t ₂ is the time, or range of times, at which there is an expected increase in the level of pollutant Y. In some situations t ₁ and t ₂ may be the same. For example, internal combustion engines of motor vehicles (which are the source of approximately 60% of global air pollution) emit NO mostly. Following emission, the NO is transformed into N0 ₂ by taking an oxygen atom from ozone, thus reducing the ozone concentration in the ambient, as mentioned above. As a result, there is typically a correlation between the level of change and the time derivative of nitrogen dioxide (N0 ₂ ) and ozone (0 ₃ ). A typical example of a decrease of 0 ₃ 610 and an opposite increase of N0 ₂ 620 is presented in Fig. 7, which presents simulation of levels of 0 ₃ and N0 ₂ following a wood fire.

The rule for assessing validity of spikes of N0 ₂ may therefore be formulated as follows:

and ALJVO., (tj = - / AL^ (ti)

then spike is real; is saved

else spike is false; AL _jVOi (t ₁ ) ₂ is saved.

Here, ki is a coefficient based on empiric and/or analytic data, and k ₂ is a coefficient having values in the range of, for example, 0 < k ₂ < \ . The value of k ₂ may be determined by the level of confidence in the operator, such that the saved value of AL _N0^ may, for example, be set at an average value of AL _N07 , and the spike filtered out. Alternatively, ½ ₀ , and related variables may take other appropriate values in case the spike is false, as long as the spike is filtered out. In block 520 readings of the plurality of pollution sensors are obtained; the readings relate to a possible pollution event, for example to a spike in the reading of at least one pollution sensor. The readings may be obtained as a time series of measured values. In block 530 the readings are analyzed based on the set of predetermined rules. Variables for the rules may be calculated and the predefined rules may be applied. In block 540, validity of the readings may be determined based on the analysis. If the readings are not valid, indicated as 'NO' on block 550, than the readings may be discarded, as indicated in block 560. If the readings are valid, indicated as 'YES' on block 550, the readings may be stored in the system memory for further analysis, as indicated in block 570. The rules may be applied to the reading in real time. Therefore, the method presented hereinabove may serve as real time filter. This may enable filtering of streaming data from environmental sensors.

In this way data from multiple units can increase the confidence in the data, in other words, simultaneous (or nearly simultaneous) spikes sensed by units geographically close to one another are more likely to be real than a single spike that is not repeated in other data obtained. Similarly, spikes that are sequentially sensed in time by a series of geographically spaced sensors are likely to be real, showing the progression or evolution of the pollution/pollution source with time. in the case where sequential spikes are sensed, this may be due to a source in one location, with the pollution moving, for example by diffusion or with the wind. Thus, for two sensors which are located 1 km apart, where a spike is sensed by one of the sensors an hour after the sensing of the spike by the other sensor, and the wind speed is measured at 1 km per hour, this shows that the pollution is moving with the wind. Correlation of the wind speed with speed of movement of the pollution can also increase confidence in the reliability of the measurements. Knowledge of local wind speed and direction can provide information regarding the predicted movement of the pollution, allowing advance warnings to be issued to downwind locations. These warnings can be time-sensitive. In other words, the warnings can link future pollution increases to a time frame within which the increase is expected.

In a related manner, the duration of the spike as sensed at at least one location can provide information regarding the expected duration of the pollution at downwind locations. This allows the duration of the pollution to also be predicted.

The use of many sensing units, as contemplated in the present techniques, allows relatively dense networks of sensing units to be used. This allows more accurate determination of the 'wavefronts ^' of the pollution moving across a location, for example a city. The sensing of the pollution 'wave' at a plurality of points allows predictions to be made regarding the movement of the pollution.

The techniques presented hereinabove with reference to Figs 4 and 7 may be performed in any sensing unit that may obtain the readings of the plurality of pollution sensors. For example, considering the sensing unit 100, these methods may be performed locally by the data acquisition and processing unit 12.

Additionally or alternatively, these methods may be performed in a remote location such as at the server 214 or at the centralized server 34.

The present techniques may be implemented in software for execution by a processor-based system, for example, the rule-based pollution source identification method and the rule-based method for assessing validity of environmental data. For example, the present techniques may be implemented in code and may be stored on a non-transitory storage medium having stored thereon instructions which can be used to program a system to perform the instructions. The non-transitory storage medium may include, but is not limited to, any type of disk including floppy disks, optical disks, compact disk read-only memories (CD-ROMs), rewritable compact disk (CD-RW), and magneto-optical disks, semiconductor devices such as read- only memories (ROMs), random access memories (RAMs), such as a dynamic RAM (DRAM), erasable programmable read-only memories (EPROMs), flash memories, electrically erasable programmable read-only memories (EEPROMs), magnetic or optical cards, or any type of media suitable for storing electronic instructions, including programmable storage devices. Other implementations may comprise dedicated, custom made or off-the-shelf hardware, firmware or a combination thereof.

The present techniques may be realized by a system that may include components such as, but not limited to, a plurality of central processing units (CPU) or any other suitable multi-purpose or specific processors or controllers, a plurality of input units, a plurality of output units, a plurality of memory units, and a plurality of storage units. Such a system may additionally include other suitable hardware components and/or software components.