METHOD FOR TRANSMITTING DATA FROM A SENSOR

Abstract

A method for transmitting data collected by at least one sensor to a monitoring device. The method includes, upon acquisition of a new piece of data by the at least one sensor, acts of calculating a deviation indicator indicating a deviation between the value of the new piece of data and a value predicted for this piece of data by a prediction model representative of previously acquired data, and transmitting the new piece of data to the monitoring device when the deviation indicator is higher than a threshold. Also provided are a monitoring method on a monitoring device, a terminal implementing the transmission method and a server implementing the monitoring method.

Claims

1. A method for transmitting, to a monitoring device, data collected by at least one sensor, wherein the method comprises the following act on acquisition of a new datum by the at least one sensor: computing an indicator of deviation between the value of the new datum and the value predicted for this datum by a prediction model representative of data previously acquired, and transmitting the new datum to a monitoring device when the deviation indicator is above a threshold.

2. The method as claimed in claim 1, which further comprises the following prior acts: transmitting to the monitoring device data acquired by the at least one sensor over a predetermined time period, and receiving, from the monitoring device, a prediction model representative of the data transmitted.

3. A method for monitoring, by a monitoring server using data from at least one sensor, wherein the method comprises: receiving data from the at least one sensor; storing the data in a non-transitory computer-readable medium; replacing in the medium a sensor value not received with a value predicted by a prediction model representative of data previously received as long as a new datum, for which an indicator of deviation between its value and the value predicted for this datum is above a threshold, is not received.

4. The monitoring method as claimed in claim 3, which further comprises the following acts: receiving the data from the at least one sensor over a predetermined time period, computing the prediction model, which is representative of the data received over the period, and transmitting the prediction model to the at least one sensor.

5. The method as claimed in claim 3, which further comprises, on reception of a new datum from the at least one sensor, an act of updating the prediction model.

6. The method as claimed in claim 5, which further comprises act of transmitting the updated model to the at least one sensor.

7. The method as claimed in claim 5, wherein the act of updating of the model is performed when the frequency of reception of new data is above a threshold.

8. A device for transmitting, to a monitoring device, data collected by at least one sensor, wherein the device comprises: an acquisition module for a datum measured by the least one sensor, a computer configured for computing an indicator of deviation between the value of the new datum and a value predicted for this datum by a prediction model representative of data previously acquired, a comparator, which is configured to compare the deviation indicator to a threshold, and a communication module configured to transmit the new datum to the monitoring device when the deviation indicator is above the threshold.

9. The transmission device as claimed in claim 7, wherein the communication module is also configured to receive a prediction model representative of the data transmitted from the monitoring device.

10. A monitoring device comprising: a communication module configured to receive data from at least one sensor of a monitoring system, a module configured to read a value predicted by a prediction module representative of sensor data previously received, a monitoring module configured to replace a sensor value not received, with the predicted value as long as a new datum, for which an indicator of deviation between its value and the value predicted for this datum is above a threshold, is not received by the communication module.

11. The monitoring device as claimed in claim 10, which further comprises a computer configured to compute a predictive model representative of data received over a predetermined period, and wherein the communication module is also configured to transmit the predictive model to at least one transmission device.

12. A terminal, which comprises a transmission device as claimed in claim 8.

13. A server, which comprises a monitoring device as claimed in claim 10.

14. (canceled)

15. (canceled)

16. (canceled)

Description

LIST OF FIGURES

[0049] Other features and advantages of the invention will become more clearly apparent on reading the following description of a particular embodiment, given as a simple illustrative and nonlimiting example, and the attached drawings, in which:

[0050] FIG. 1 shows a simplified illustration of an architecture suitable for implementing the transmission and monitoring methods according to a particular embodiment of the invention.

[0051] FIG. 2 illustrates the main steps of the transmission method according to a particular embodiment of the invention.

[0052] FIGS. 3a and 3b illustrate the main steps of the monitoring method according to a particular embodiment of the invention.

[0053] FIGS. 4a and 4b represent two series of measurements acquired by a sensor over a time period on which are overlaid two curves illustrating a predictive model for said periods.

[0054] FIG. 5 illustrates a simplified architecture of a transmission device according to a particular embodiment of the invention.

[0055] FIG. 6 illustrates a simplified architecture of a monitoring device according to a particular embodiment of the invention.

[0056] FIG. 7 is a diagram illustrating the implementation of the transmission method over 3 consecutive periods.

DETAILED DESCRIPTION

[0057] FIG. 1 shows a simplified illustration of the architecture of a monitoring system suitable for implementing the transmission and supervision methods according to a particular embodiment of the invention. It represents in particular a transmission device 100 comprising a thermal sensor 105 and a network interface 103 suitable for transmitting, for example, temperature readings measured by the sensor 105 to a server 102 via a network 104. The transmission device can, in another exemplary embodiment, comprise several measurement sensors of different types. This transmission device can for example be a communication terminal of “smartphone” type in which sensors are incorporated. It can also be a simple temperature transmission device as illustrated in FIG. 1. The network interface 103 is also capable of receiving data transmitted by the server 102, such as, for example, predictive models computed by the server 102 from data transmitted by the sensor 105. The server 102 can be hosted, for example, on a domestic gateway of a local area network and communicate with the transmission device 100 via a local area network. Various network technologies can be used, such as, for example, a Wifi, Ethernet or even Bluetooth network. It can also be hosted in a communication network of Internet type and communicate with the transmission device 100 via an Internet or 3G network for example. Although the invention is described here using the example of a temperature sensor, it is applicable to different types of sensors and is particularly advantageously applicable with sensors with a high measurement frequency, such as, for example, accelerometers or gyroscopes. The monitoring system represented here comprises only a single transmission device comprising a single sensor 105 but, according to other examples, the transmission device can comprise a plurality of sensors and the monitoring system can comprise a plurality of transmission devices 100.

[0058] FIG. 2 illustrates the main steps of the transmission method according to a particular embodiment of the invention.

[0059] In a step 200, the transmission device 100 obtains data measured by the temperature sensor 105. In other embodiments, the data can originate, for example and in a non-exhaustive manner, from sensors suitable for measuring accelerations, angular speeds or even magnetic fields. The data can also be obtained from several sensors or from several instances of a same type of sensor. For example, the data can originate from an accelerometer suitable for measuring accelerations on 3 axes.

[0060] In a step 203, the data obtained from the thermal sensor 105 are compared to a predictive model stored for example in a memory of the device. The model can also be stored in a database of the network 104 and can be consulted by the device or of the sensors.

[0061] The predictive model used is representative of the trend of the data measured by the sensor. This model can for example be a numeric function of affine type which, for a given instant, makes it possible to predict the value of a measurement. According to other embodiments, the data are modeled for example by a linear or polynomial regression or any other mathematical or statistical function suitable for describing the trend of the series of data measured. FIG. 4a represents, for example, a vertical line chart illustrating 12 temperature readings 400 over a period of 24 hours, at the rate of one measurement every two hours. In this example, the data are modeled by a normal law whose mean and variance are determined, for example, by successive tests. This representative model of the data is represented on the figure by the curve 401.

[0062] A deviation indicator is computed from a datum obtained from the sensor 105 and from its predictive value according to the predictive model so as to validate or invalidate the fit of the measured value with the value predicted by the model. The fit can be verified for example by measuring the deviation between the measured value and the predicted value, or, according to a particular embodiment, from a value obtained from a statistical study taking into account several measurements, or even for example from a test of χ.sup.2 (Khi-2) making it possible to validate the fit of a series of data with a model.

[0063] In the step 204, the data which do not fit with the model are transmitted to the server 102 via the network. The data which do fit with the model are, for their part, disregarded, so as to reduce the quantity of data transmitted over the network.

[0064] FIG. 4b represents a vertical line chart illustrating 12 temperature readings 402 over a period of 24 hours, at the rate of one measurement every two hours. In this example, the data are modeled by the same normal law as that represented in FIG. 4a, represented by the curve 401. In this example, the readings 403, 404 and 405 no longer correspond to the predictive model and the comparison of the deviation indicator with a predetermined tolerance threshold designates these readings as having to be transmitted to the server. Thus, only the readings 403, 404 and 405 are transmitted over the period.

[0065] According to a particular embodiment, the transmission method comprises an initialization phase during which no predictive model is available for the transmission device 100. During the step 201, all the measurements obtained during an initialization period are transmitted to the server 102 because, if there is no predictive model available, it is not possible to compute a deviation indicator. At the end of this first period, the server 102 transmits a predictive model computed from the data transmitted by the transmission device 100 during the initialization period. Thus, a predictive model representative of the data measured over the initialization period is received in the step 202. This model can then be used to verify the fit of the data from the sensor 100 in subsequent periods.

[0066] FIG. 3a represents steps that can be performed to implement the monitoring method according to a particular embodiment of the invention. The method is for example implemented on the server 102 described with reference to FIG. 1.

[0067] In the step 300, the server 102 initializes a monitoring task for the data obtained from the transmission device 100 with the aim, for example, of storing, in a database, the temperatures read by the temperature sensor 105 during the day.

[0068] For each time band, the server verifies, in the step 301, whether a datum from the transmission device has been received. For that, the server stores, for example in a random access memory, the data received and the time band to which they correspond. If a datum is found in the memory for a time band, this datum stored in the database in the step 302 and the next time band can be processed.

[0069] When, in the step 301, a datum is not found in the random access memory for a given time band, the server 102 assesses, in the step 303, a predictive model representative of data previously measured by the sensor. The assessment of this model allows the server 102 to obtain a predictive value for the time band concerned when a datum is not received.

[0070] FIG. 3b illustrates steps of the monitoring method according to a particular embodiment of the invention.

[0071] In an initial step 304, the server 102 receives measurements relating to a given period from the transmission device 100. These data are for example transmitted by the transmission device 100 in the initialization step 201 described with reference to FIG. 2 and corresponding to the readings performed by the sensor over a first initialization period. According to a particular embodiment, these data are stored by the server in a database or a random access memory.

[0072] From these data, the server 102 computes, in the step 305, a predictive model representative of the data received over the period concerned. For that, the server can determine parameters of a numeric function, such as, for example, parameters of a function based on an affine or normal law or even a polynomial function. The number and the value of the parameters are chosen so as to obtain a function for which the values approximate measurements transmitted by the transmission device. The choice of the parameters can be made according to different optimization techniques known to those skilled in the art, such as, for example, a least squares optimization method or a splines-type technique.

[0073] According to another particular embodiment, the data to be modeled are segmented into a plurality of time bands, each of the bands being modeled independently by a numeric function and parameters, said parameters being determined for example by an optimization method of least squares or splines type.

[0074] In the step 306, the model is transmitted to the transmission device 100, for example in the form of a numeric function and parameters computed in the preceding step.

[0075] In this way, the monitoring method relieves the transmission device of the step of computation of the model which is particularly costly in terms of computation time.

[0076] At the end of the step 306, a copy of the predictive model is retained on the server 102 such that, subsequently, according to the steps described with reference to FIG. 3a, if an item of equipment interrogates the server to consult a measurement transmitted by the sensor, or for the requirements of a monitoring task, the server can assess the predictive model to learn the value of a measurement which has not been received. Thus, according to a particular embodiment, there is no need to retain the data from which the model was computed and the data received in the step 304 can be deleted in order not to unnecessarily occupy storage space on the server.

[0077] According to a particular embodiment, the data set initially used to compute the model is stored in a database. On reception of a new datum from the transmission device, the corresponding datum is replaced by the new datum in the database. The server uses this modified data set to compute a new predictive model. For that, the server once again executes the step 305 from the modified data set stored in the database.

[0078] According to a particular embodiment, when the model is updated by the server 102 following the reception of a new datum, this updated model is transmitted to the transmission device 100. The method thus makes it possible to improve the fit of the data from the sensor 105 with the predictive model so as to further reduce the quantity of the data exchanged between the transmission device 100 and the server 102.

[0079] According to a particular embodiment, the server 102 measures the frequency at which measurements are transmitted by the transmission device 100. For that, the server computes, for example, an indicator that takes into account the frequency of reception of the measurements over a period and the number of measurements transmitted initially by the transmission device 100 in the step 304. Since the measurements are transmitted only when they do not fit with the model, the more measurements the server receives, the less the model fits with the measured data. Thus, when the indicator is above a predetermined threshold, the server recomputes a predictive model on the basis of the latest data received for the period and transmits this new model to the transmission device.

[0080] According to a particular embodiment, the server 102 analyzes the temporal distribution of the received data. When, for example, received data are grouped together over a restricted time interval relative to the observation period, only the part of the model corresponding to this time interval is updated from the new data and transmitted to the transmission device. For example, when the temperature transmission device 100 uses a 24-hour predictive model to filter the sending of the temperature readings and the server 102 receives readings corresponding to the time interval [1200 hours-1400 hours], the server can deduce therefrom that the model used by the sensor is no longer suitable for this time period. The server 102 then computes a new predictive model representative of the data received over the interval [1200 hours-1400 hours] and transmits this new model to the transmission device 100. The method thus makes it possible to update the predictive model without the need to recompute it in its entirety. This embodiment thus preserves the computation resources of the server and the bandwidth for transmitting the model.

[0081] FIG. 7 gives a summary illustration in the form of a diagram of the temperature readings transferred between the transmission device 100 and the server 102. The figure shows three periods of operation P0, P1 and P2, each period having for example a duration of 24 hours. The period P0 is an initialization period. During this period, the transmission device 100 does not have a predictive model and all the measurements are transmitted to the server 102. At the end of the period P0, the server 102 computes and transmits to the transmission device a predictive model according to the steps 301 and 302 described previously. In the period P1, the transmission device implements a filtering of the measurements from the predictive model 700 by verifying the fit of the measurements with the model according to the step 203. Only the measurements which do not fit with the model, such as the measurements 701 for example, are transmitted to the server. At the end of the period P1, the server analyzes the distribution of the measurements received and decides to update the predictive model from these new measurements and to transmit it to the transmission device 100 so as to reduce the traffic. In the period P2, the updated model 702 is used by the transmission device 100 to filter the measurements. Since the model fits with the measured data, no measurement is transferred.

[0082] FIG. 5 illustrates a device 500 implementing the transmission method according to a particular embodiment of the invention. The device comprises a storage space 502, for example a memory MEM, a processing unit 501 equipped for example with a processor PROC. The processing unit can be driven by a program 503, for example a computer program PGR, implementing the transmission method as described in the invention with reference to FIG. 2, and notably the steps of computation of an indicator of deviation between the value of the new datum and a value predicted for this datum by a prediction model representative of data previously acquired, and of transmission of the new datum to the monitoring device when the deviation indicator is above a threshold. According to a particular embodiment, the device also implements the steps of transmission to the monitoring device of the data acquired by the at least one sensor over a predetermined time period, and of reception, from the monitoring device, of a prediction model representative of the transmitted data.

[0083] On initialization, the instructions of the computer program 503 are for example loaded into a RAM memory (Random Access Memory in English) before being executed by the processor of the processing unit 501. The processor of the processing unit 501 implements the steps of the transmission method according to the instructions of the computer program 503.

[0084] For that, the device comprises, in addition to the memory 502, communication means 504 (COM) enabling the device to connect to a telecommunication network and to exchange data with other devices via the telecommunication network, and in particular to transmit measurements to a server and to receive a predictive model. According to a particular embodiment, the device further comprises a module for acquiring a measurement 506 suitable for capturing, for example, a physical quantity linked to the environment or to movements. For example, the acquisition module 506 is a temperature sensor, an accelerometer, a gyroscope, a compass, an anemometer or even an interfacing module suitable for connecting a remote sensor. This interfacing unit may correspond for example to a USB (Universal Serial Bus), Bluetooth, Ethernet interface or even, for example, to a communication bus. The device also comprises a computer 507 (CALC) suitable for computing an indicator of deviation between the value of the new datum and a value predicted for this datum by a prediction model representative of the data previously acquired by the acquisition module 506, a comparator 505 (CMP) suitable for comparing the deviation indicator to a tolerance threshold and allowing the communication module 504 to transmit a new datum to the monitoring device when the deviation indicator is above the threshold.

[0085] According to a particular embodiment, the device can be incorporated in a terminal or a home gateway.

[0086] FIG. 6 illustrates a device 600 implementing the monitoring method according to a particular embodiment of the invention. The device comprises a storage space 602, for example a memory MEM, a processing unit 601 equipped for example with a processor PROC. The processing unit can be driven by a program 603, for example a computer program PGR, implementing the monitoring method as described in the invention with reference to FIG. 3, and notably the steps of reception of data from at least one sensor; of taking into account, to replace a sensor value not received, a value predicted by a prediction model representative of data previously received as long as a new datum, for which an indicator of deviation between its value and the value predicted for this datum is above a threshold, is not received; of computation of a predictive model representative of the data received over a predetermined period; and of transmission of the predictive model to at least one sensor. According to a particular embodiment, the device also implements the steps of updating of a predictive model from new data received.

[0087] On initialization, the instructions of the computer program 603 are for example loaded into a RAM memory (Random Access Memory in English) before being executed by the processor of the processing unit 601. The processor of the processing unit 601 implements the steps of the transmission method according to the instructions of the computer program 603.

[0088] For that, the device comprises, in addition to the memory 602, communication means 604 (COM) allowing the device to connect to a telecommunication network and to exchange data with other devices via the telecommunication network, and in particular to receive measurement data from a transmission device and to transmit a predictive model representative of data received. The device also comprises a computer 605 (PRED) suitable for computing a predictive model representative of the data received over a predetermined period and a monitoring module 608 (MON) suitable for taking into account, to replace a sensor value not received, a predicted value as long as a new datum for which an indicator of deviation between its value and the value predicted for this datum is above a threshold, is not received by the communication module.

[0089] According to a particular embodiment, the device comprises a module 606 for analyzing the frequency of reception of data by the communication module and a database 607 suitable for storing a set of measurement data received over an observation period.

[0090] According to a particular embodiment, the device can be incorporated in a server or a home gateway.