| JP63208755 | MANUFACTURE OF ELECTRODE FOR POLAROGRAPH |
| JP06213857 | DISPOSABLE CARTRIDGE FOR ION SELECTIVE ELECTRODE SENSOR |
| JP01216257 | ELECTROCHEMICAL SENSOR |
AYMEN, M., Karim (Mimos Berhad, Technology Park Malaysia, Kuala Lumpur, 57000, MY)
MASURI, Bin Othman (Mimos Berhad, Technology Park Malaysia, Kuala Lumpur, 57000, MY)
WAN ADIL, Bin Wan Jamil (Mimos Berhad, Technology Park Malaysia, Kuala Lumpur, 57000, MY)
AIRUL AZHA, Bin Abd Rahman (Mimos Berhad, Technology Park Malaysia, Kuala Lumpur, 57000, MY)
AYMEN, M., Karim (Mimos Berhad, Technology Park Malaysia, Kuala Lumpur, 57000, MY)
MASURI, Bin Othman (Mimos Berhad, Technology Park Malaysia, Kuala Lumpur, 57000, MY)
WAN ADIL, Bin Wan Jamil (Mimos Berhad, Technology Park Malaysia, Kuala Lumpur, 57000, MY)
| WHAT IS CLAIMED IS: 1. A method of predicting steady state response of a slow sensor reaction in data acquisition comprising steps of, obtaining at least one batch of initial measured data in rising time , storing said measured data in a memory, determining the function augments based on at least one batch of initial measured data utilising a processor in the calibration component (29) of an apparatus for predicting steady state response; predicting steady state response in the steady state prediction component (30) of an apparatus for predicting steady state response. 2. A method of predicting steady state response of a slow sensor reaction in data acquisition as in Claim 2 further characterized in that the determining of the said function augments comprises steps of, transferring measured data into the calibration component (29) of an apparatus to predict steady state response, determining errors in the said calibration component (29) ; minirnising errors in the said calibration component (29) ; utilising an ascertained constant value (a) in the steady state prediction component (30) to predict steady-state response 3. A method of predicting steady state response of a slow sensor reaction in data acquisition as in Claim 2 which utilizes a pre-determined nonlinear curve fitting algorithm. 4. A method of predicting steady state response of a slow sensor reaction in data acquisition as in Claim 2 wherein a pre-determined constant value (a) is ascertained after the said measured data undergoes calibration processes in the calibration component (29) of the steady state predicting apparatus. 5. An apparatus for predicting steady state response of a slow sensor reaction in data acquisition wherein its calibration component (29) comprises at least a summation circuitry (12), at least a division circuitry (14), at least a multiplication circuitry (16), at least an addition circuitry (18) and at least an exponential circuitry (20) to manipulate at least one measured data (input data) to generate a plurality of estimated data at different times whereby each estimated data (22) (24) (26) (28) is multiplied in the multiplication circuitry (16) by an exponential data (22A) (24A) (26A)(28A) generated by the exponential circuitry (20) to obtain estimated values (22B) (24B) (26B) (28B) where the estimated values (22B) (24B) (26B) (28B) are then summed up in the summation circuitry (12); said summed-up value will then be added to an exponential function in an addition circuitry (18) and then divided by another exponential function in a division circuitry (14) to arrive at a constant value (a) for calculating estimated steady state value when time is set at a predetermined value or infinity in the steady state predicting component (30) of the said apparatus for predicting steady state response. |
1. TECHNICAL FIELD OF THE INVENTION
The present invention relates generally to a method and apparatus for predicting steady state response of a slow sensor reaction for efficient data acquisition thereby saving time, energy and cost.
2. BACKGROUND OF THE INVENTION
As is known chemical reaction takes time to stabilize before any parameters for example the value of pH or concentration can be measurable. Some chemical sensors require a long time for the chemical effects to stabilize. Consequently taking readings usually lasts more than five minutes. Also the time interval between data reading is too long. This is not only time consuming but a boring task that as one has to wait patiently to be able to acquire data from the chemical reaction. In many applications several minutes waiting time is just too long for efficient data acquisition or collection. This is illustrated in the FIG. 1 which shows the slow response curve taking approximately 10 minutes to reach its steady state after the pH sensor is dipped into a solution. If a sensor hardware requires 10 minutes per reading then if the data acquisition has to perform 12 times per day, then the total time is 120 minutes per day which is a lengthy two hours. Hence the data acquisition time is too long for effective power saving. Consequently, this can lead to high power consumption required by the system. For field applications, where the system is normally powered by batteries that have limited power, such high power consumption would need regular replacement of batteries which would be impractical. In this new era where speed is the order of the day there is an increasing requirement of performing tasks in a faster and more efficient fashion in order to excel. The reduction of the time required by the sensor hardware to take each reading to less than 10 minutes can ultimately shorten the actual reading time quite considerably, extend the battery life and multiple data reading is made possible.
It would hence be extremely advantageous if the above shortcoming is alleviated by having a method of predicting steady state response of a slow sensor reaction in data acquisition that is capable of reducing the reading time thereby promoting efficiency in acquisition of data. 3. SUMMARY OF THE INVENTION
Accordingly, it is the primary aim of the present invention to provide a method of predicting steady state response of a slow sensor reaction that is capable of reducing the reading time during data acquisition. Yet another object of the present invention is to provide a method of predicting steady state response of a slow sensor reaction that is capable of reducing power consumption.
It is a further object of the present invention to provide a method of predicting steady state response of a slow sensor reaction that is capable of prolonging battery life.
It is a further object of the present invention to provide a method of predicting steady state response of a slow sensor reaction that promotes efficiency in data acquisition.
Yet another object of the present invention is to provide a method of predicting steady state response of a slow sensor reaction that is cost effective.
It is a further object of the present invention to provide a method of predicting steady state response of a slow sensor reaction that is capable of being employed on all sensors which have slow response characteristics. Yet a further object of the present invention is to provide a method of predicting steady state response of a slow sensor reaction which exploits the nonlinear curve fitting technique by taking only the real read data from the slow sensor within a short time (initial few seconds) and stores these acquired data in a memory to be applied later on a processor that predicts the steady state response of the sensor.
Other and further objects of the invention will become apparent with an understanding of the following detailed description of the invention or upon employment of the invention in practice. According to the one embodiment of the present invention there is provided,
An apparatus for predicting steady state response of a slow sensor reaction in data acquisition wherein its calibration component comprises at least a summation circuitry (12), at least a division circuitry (14), at least a multiplication circuitry (16), at least an addition circuitry (18) and at least an exponential circuitry (20) to manipulate at least one measured data (input data) to generate a plurality of estimated data at different times whereby each estimated data (22) (24) (26) (28) is multiplied in the multiplication circuitry (16) by an exponential data (22A) (24A) (26A)(28A) generated by the exponential circuitry (20) to obtain estimated values (22B) (24B) (26B) (28B) where the estimated values (22B) (24B) (26B) (28B) are then summed up in the summation circuitry (12); said summed-up value will then be added to an exponential function in an addition circuitry (18) and then divided by another exponential function in a division circuitry (14) to arrive at a constant value (a) for calculating estimated steady state value when time is set at a predetermined value or infinity in the steady state predicting module (30) of the said calibration component.
In another aspect of the said invention there is provided,
A method of predicting steady state response of a slow sensor reaction in data acquisition comprising steps of, obtaining at least one batch of initial measured data in rising time J storing said measured data in a memory, determining the function augments based on at least one batch of initial measured data utilising a processor in the calibration component (29) of an apparatus for predicting steady state response; predicting steady state response in the steady state prediction component
(30) of an apparatus for predicting steady state response. 4. BRIEF DESCRIPTION OF THE DRAWINGS
Other aspect of the present invention and their advantages will be discerned after studying the Detailed Description in conjunction with the accompanying drawings in which: FIG. 1 is a graph illustration of concentration against time after the pH sensor is dipped into a solution.
FIG. 2 is the process flow for predicting steady state value.
FIG. 3 shows a firmware architecture of a calibration component and a steady state prediction component of a steady state predicting apparatus in the form of block diagram.
FIG. 4- A shows the comparison between the overall curve responses plotted from real read data and curve fitting function estimation.
FIG. 4-B shows the curve response based on actual real read data shown in FIG. 4-A in isolation. FIG. 4-C shows the curve response based on curve fitting function estimation shown in FIG. 4-A in isolation. 5. DETAILED DESCRIPTION OF THE DRAWINGS
In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be understood by those with ordinary skill in the art that the invention may be practised without these specific details. In other instances, well known methods, procedures and/ or components have not been described in detail so as not to obscure the invention.
The invention will be more clearly understood from the following description of the embodiments thereof, given by way of example only with reference to the accompanying drawings, which are not drawn to scale.
Referring to the drawings in which like numerals indicate like parts throughout the views shown, FIG. 1 is a graph illustration of the function in concentration against time; with the function in concentration domain designated as the x-axis and function in time domain designated as the y-axis. This graph is plotted using actual data read during a chemical sensor reaction wherein there is a region of initial contact (2) where the graph illustration continuously increases in time and a region for capturing data (4) where data is acquired. The said graph shows the actual slow response curve which takes approximately 10 minutes to reach its steady state after the pH sensor is dipped into a solution. This time consuming problem in data acquisition is inevitably faced by researchers or chemists using the conventional models.
To overcome the above problem the present invention which offers a technique or method that attempts to find a function in the time domain that best fits the curve illustrated in FIG. 1 is disclosed.
It has been found that the trajectory of the curve in FIG. 1 could be approximated by a certain function. Below are four functions that are capable of producing an estimated curve that best fitted the curve in FIG. 1,
The symbols are mathematical notation, where:- g of t is mathematical symbol for representing system or function of system. t is mathematical notation for time a constant coefficient that influences the curve response of the output,; b is a variable representing the gradient of the curve; e is an exponential function; n represents the number of times; and
N represents the total number of reading required. The present method utilizes an algorithm preferably a non-linear curve fitting algorithm to determine the function augments in time domain that best fits the curve by using only several initial read-samples in rising time. This proposed method shortens the data acquisition process of a slow response sensor. It exploits the non-linear curve fitting technique by taking only some of the data read from the slow sensor in a short time (initial few seconds) and stores these data in a memory to be applied later. A processor is subsequently utilized to perform non-linear curve fitting using the data stored in the memory. The processor determines the augments (for the last function « * <½), and then computes the function value when time is at infinity (t→∞ ) which is the steady- state value.
The said algorithm is useful in overcoming delay in data acquisition for slow response sensor employed in soil engineering applications and for any type of slow response sensors such as gas sensors and fruit ripening application. FIG. 2 is the process flow for predicting steady state value consisting of four major steps namely a data collection step (6), an error determination step (8) wherein R 2 is used as a measure of error, an error mmimising step (10) and a steady-state prediction step (12). All these steps are carried out in the calibration component of the steady state predicting apparatus. At least one batch of initial measured data will be collected in the data sample collection step (6). Then the said measured data will be used to compute the error (R 2 ) which is the summation of [(y -g(ti)] 2 in the error (Redetermination step (8). Upon obtaining the value of error (R 2 ), the next step is to find a constant coefficient (a) that influences the curve response of the output, which mmimises the error (R 2 ) and this is carried out in the error minimising step (10). Finally, when the value of the constant coefficient (a) that influences the curve response of the output is determined, it is used in the function of graph of time \g(t)] to predict the steady- state value by setting time to infinity. Alternatively time may be set to a pre- determined value to predict steady state response.
FIG. 3 shows a firmware architecture of a calibration component (29) and a steady state prediction component (30) of a steady state predicting apparatus in the form of block diagram comprising various circuitry namely at least a summation circuitry (12), at least a division circuitry (14), at least a multiplication circuitry (16), at least an addition circuitry (18) and at least an exponential circuitry (20) wherein at least one measured data is used to generate a plurality of estimated data at different times which are thereafter being processed by the various circuitry to arrive at the steady state value in the steady state prediction component (30) when time is set at a pre-determined value or at infinity. The first estimated data (22) taken when time is at 1 is multiplied by the first exponential data [eh A] (22 A) by a multiplication circuitry (16) to obtain a first estimated value (22B). The second estimated data (24) taken when time is at 2 is multiplied by the second exponential data [eh A] (24A) by a multiplication circuitry (16) to obtain a second estimated value (24B). The third estimated data (26) taken when time is at 3 is multiplied by the third exponential data [eh A] (26A) by a multiplication circuitry (16) to obtain a third estimated value (26B). This is repeated by having n number of estimated data (28) taken when time is at n being multiplied by the n number of exponential data [e f « A] (28A) by a multiplication circuitry (16) to obtain n estimated value (28B). The exponential data is created by the exponential circuitry (20) and is a value obtained from off- line calibration. When the above is done all the data acquired by multiplication that is the first, second, third and n number of estimated values is summed-up using the summation circuitry (12). The summed-up value will be further added with Ee n /t, an exponential function of tl over t of magnitude E (where t and tl are the values of time) by the addition circuitry (18) and divided by Ee 2n /t, an exponential function of double tl over t of magnitude E by the division circuitry (14) to obtain a constant coefficient ( ) that influences the curve response of the output. Once the constant coefficient (a) is obtained it is utilised to calculate the function graph of time [g(t)] by multiplying it with an exponential function e- J /f and then added to the numeral 1 to predict steady-state value in the steady state prediction component (30) when time is set at mfinity. Alternatively time may be set to a pre-determined value to predict steady state response. Referring now to FIG. 4-A, 4-B and 4-C, there are respectively shown the overall curve responses between the one plotted using measured data and the one plotted using fitting function estimation, the actual curve response based on measured data is shown in FIG. 4-B in isolation and the curve response based on fitting function estimation is shown in FIG. 4-C in isolation. In FIG. 4-A the actual curve is indicated by dash lines and the best fitted function estimated curve is indicated by dotted lines. The actual curve response shown in FIG. 4-A and 4-B is based on actual measured data over a span of time of more than five minutes. The estimated curve response shown in FIG. 4-A and 4-C is based on fitting function estimation using algebraic sums of building variables. It can be seen that both the curves, that is the actual curve response derived from actual measured data and the one derived from fitting function estimation are very similar. Hence the curve derived from fitting function estimation can reliably be used to approximate or estimate the actual curve response.
While the preferred embodiment of the present invention and its advantages has been disclosed in the above Detailed Description, the invention is not limited thereto but only by the spirit and scope of the appended claim.