SwiftCounter: An Advanced Portable Peso Currency Counter PDF

SwiftCounter: An Advance Portable Device for Accurate Peso Paper Currency Counting Robotics and Intelligent Machines – Team Theme: Business Researchers: CHESKA NICOLE P. GELLANGALA LALAINE VIENNE D. BATOL MI...

SwiftCounter: An Advance Portable Device for Accurate Peso Paper Currency Counting Robotics and Intelligent Machines – Team Theme: Business Researchers: CHESKA NICOLE P. GELLANGALA LALAINE VIENNE D. BATOL MIGUEL LAWRENCE M. OSORIO Research Adviser: KATHY LYN G. DAGA-AS 1|Page  SwiftCounter: An Advance Portable Device for Accurate Peso Paper Currency Counting RESEARCH PAPER 2|Page SwiftCounter: An Advance Portable Device for Accurate Peso Paper Currency Counting Robotics and Intelligent Machines – Team Theme: Business Researchers: CHESKA NICOLE P. GELLANGALA LALAINE VIENNE D. BATOL MIGUEL LAWRENCE M. OSORIO Research Adviser: KATHY LYN G. DAGA-AS 3|Page  SwiftCounter: An Advanced Portable Device for Accurate Peso Paper Currency Counting By: Cheska Nicole P. Gellangala Lalaine Vienne D. Batol Miguel Lawrence M. Osorio Adviser: Koronadal Kathy National Lyn G. Daga-as Comprehensive High School The main goal of the study is to develop a Peso paper bill portable amount counter. The study aimed to create the circuit design and flowchart of the SwiftCounter counting system. Furthermore, it assessed the percent accuracy of the SwiftCounter in two setups: an image sensor with 2 infrared sensors and an image sensor with 4 infrared sensors. Percent accuracy was calculated by dividing the number of correct responses divided by the number of trials in all setups. The study evaluated the efficiency of the SwiftCounter in terms of response time in milliseconds which was measured by the oscilloscope. Additionally, it aimed to test if there was a significant difference in the efficiency of the SwiftCounter regarding response time among all setups. A combination of developmental research design, experimental research design, and post-test design was used, with two setups and fifteen trials. The percent accuracy of the SwiftCounter showed a full 100% accuracy in all setups. The highest average efficiency of 639.55 ms was exhibited by the SwiftCounter in setup 1 with 4 IR sensors and the lowest average response of 637.42 ms was from SwiftCounter in setup 2. There is a significant difference in the response time between 2 setups with a calculated p-value of 0.007 which is less than the 0.05 significance level. Setup 2 was found to be the most efficient in terms of response time signifying faster detection and response rate. Keywords: Response Time, Percent accuracy, Efficiency, Response rate, Detection 4|Page  Introduction In the past, banks and other financial industries made numerous counting errors in businesses and financial institutions that resulted in the invention of money-counting devices, designed to streamline the counting process and ensure precision. On the other hand, many money-counting devices have an error rate of less than 1%, which is regarded as quite accurate with a rate of 99.99% (Ribao Technology, 2022). Money counters provide a level of accuracy not possible when counting by hand (Ribao Technology, 2022). These devices employ advanced technologies, including infrared sensors, ultraviolet (UV) and magnetic (MG) detection, and magnetic thread (MT) and density detection (DD), to not only count bills but also to differentiate between various denominations and detect counterfeit currency (LE RAYON INTERNATIONAL, 2020). The main goal of the study was to develop a Peso paper bill portable amount counter that could count different types of Peso paper bills at the same time. The study intended to create the circuit design and the flowchart of the SwiftCounter counting system. Furthermore, it also wanted to assess the percent accuracy of the SwiftCounter in the following setups. Image sensor with 2 Infrared sensors and Image sensor with 4 infrared sensors. The study evaluated the efficiency of the device in terms of the response time in milliseconds among all trials in all setups. The study wanted to test the significant difference in the efficiency of the device in terms of the response time in milliseconds among all trials in all setups. The study tested the null hypothesis stating that there is no significant difference in the efficiency of the device in terms of the response time in milliseconds among all trials in all setups. This study would directly benefit agencies involved with the bank industry when handling peso bills. It can efficiently count the different types of peso bills at once and less time consumption in money counting procedures happening in the banking industry, finance 5|Page industry, and even the business industry at a faster rate than counting them one by one. The organizations must handle amounts of cash with the utmost caution and at a fast rate while maintaining the accuracy in money counting procedures. Large amounts of money can be reliably sorted and counted by the money counter. This study could also benefit the researchers themselves by gaining a customizable and automated solution, enabling precise Peso bill counting and enhancing efficiency in financial data collection for diverse research applications. This study was limited to developing a Peso paper bill portable amount counter that counts various Philippine peso bills simultaneously and displays the total value at the end. A combination of developmental research design, experimental research design, and post-test design was used in this study. The testing for accuracy and efficiency had fifteen trials in 2 setups using the t-test in all trials in all setups. The accuracy was measured in terms of the number of correct responses done in every setup. Efficiency was also measured through the response time in milliseconds using an oscilloscope. Materials that were used for this include a buzzer, infrared sensors, batteries, screws, wires, oscilloscope, etc. These supplies were bought online and delimited within the boundaries of Koronadal City, South Cotabato. The process was conducted at Koronadal National Comprehensive High School (KNCHS) Science Laboratory from January 2024 until October 2024. 6|Page  Material and Methods Research Locale The study was conducted in Koronadal City, specifically in KNCHS Science Laboratory in all phases of the study. The laboratory is well enough to cater the process. Particularly the assembly of the device and the testing for accuracy and efficiency. This was because the place had all the necessary materials needed to test the efficiency and accuracy of the device. Research Design Developmental research design, experimental research design, and post-test design were used in this study. Assembly of the prototype device was the development stage, testing the effect of using two different sensors on the efficiency of the SwiftCounter was the experimental stage, and post-test design was used for the testing of the SwiftCounter’s response time in detecting different types of bills. Two setups have 15 trials each, with a total of 30 trials to test the efficiency of the SwiftCounter The first phase was the conceptualization and identification of the problem. The second phase was the planning of the study. During this phase, a cost analysis of the materials was done. The third phase was the development of the device. In this phase, the gathering and assembling of materials was done. The testing and gathering of results were the fourth phase of the study. The efficiency of the device was tested in this phase. In the fifth phase, the analysis of data from the experimentation was done. The results and discussion of the study were done 7|Page in the sixth phase. During this phase, the conclusion was discussed, and the hypotheses was approved. Gathering of Materials Commonly used materials such as the Arduino Nano, infrared sensor, battery, and wires will be purchased from a local store within Koronadal City. Materials not available within Koronadal City, like the Huskylens, DC motor gear, LCD display, and 10K resistor, were bought online. Circuit Design There were two circuit designs for this study given that there were two setups. The first circuit design of the study includes an image sensor, DC motors, 5V relays, LCD display, DC- DC buck converter, and 2 IR sensors which were connected to the Arduino Nano according to the corresponding pins as shown in the figure below. Figure 1 Circuit Design of Setup 1 8|Page  The second circuit design of the study includes an image sensor, DC motors, 5V relays, LCD display, DC-DC buck converter, and 4 IR sensors which were connected to the Arduino Nano according to the corresponding pins as shown in the figure below. Figure 2 Circuit Design of Setup 2 Interfacing the Huskylens AI Image Sensor The VCC pin of the Huskylens AI Image Sensor was connected to the VIN of the board. Then, the ground (GND) pin was connected to the ground (GND) of the board. The SDA pin of the Sensor was connected to the Digital input pin D7 of the board and the SCL pin of the Sensor was connected to the Digital input pin D8 of the board. This would receive and transmit data from the Image Sensor to the Arduino. Interfacing the Four (4) IR Sensor The VCC pins of all four IR Sensors were connected to the 5v pin of the Arduino Nano board. The ground (GND) pin of all four IR Sensors was connected to the ground (GND) pin of the Arduino Nano Board. The OUT pins of all four IR Sensors were connected to the Analog pins A0, A1, A2, and A3 of the Arduino Nano board. 9|Page Interfacing the Two (2) DC Motors and Relays The negative terminal of the first DC Motor was connected to the IN+ of the Buck Converter and the negative terminal of the second DC Motor was connected to the OUT- of the Buck Converter. The GND pin was connected to any ground of the circuit. The positive terminal of the first DC Motor was connected to the Normally Closed (NC) Contact of the 5V Relay and the Common Contact of the 5V Relay was connected to the IN- of the Buck Converter. The positive terminal of the second DC Motor was connected to the Normally Closed (NC) Contact of the 5V Relay and the Common Contact of the 5V Relay was connected to the OUT+ of the Buck Converter. The IN of the first 5V Relay was connected to the D10 pin of the Arduino Nano, and the VCC of the 5V Relay was connected to the OUT+ of the Buck Converter. The IN of the second 5V Relay was connected to the D11 pin of the Arduino Nano, and the VCC of the 5V Relay was connected to the OUT+ of the Buck Converter. Interfacing the Charging Module and Buck Converter The positive side of the 12V Lithium-Ion Battery was connected to the on/off switch going to the VIN of the Arduino Nano board and the ground of the battery was connected to the ground of the Arduino Nano board. The positive side of the 12V Lithium-Ion Battery was connected to the IN+ of the Buck Converter, and the negative side of the 12V Lithium-Ion Battery was connected to the IN- of the Buck Converter. This powers the IR sensors, Huskylens, DC Motors, and LCD Screen. Interfacing other Components The GND of the LCD screen was connected to the ground of the Swift Counter, the VCC of the LCD screen was connected to the VIN of the Arduino Nano, The SDA of the LCD screen was connected to the Analog pin A4 of the Arduino Nano and the SCL of the LCD screen was connected to the Analog pin A5 of the Arduino Nano. This served as the display for the types of peso paper bills that were counted and the total amount of the peso paper bills. After 10 | P a g e connecting all the wires and pins, the Arduino Nano board with the Arduino IDE application was programmed and the code was uploaded. The Swift Counter was switched on and different types of peso paper bills were placed. The Swift Counter was tested based on the recognition of the peso paper bills and the function based on the program. Flowchart of the SwiftCounter System The device starts when the command is given. If the IR sensor detects any presence of a peso bill, then the peso bill then proceeds into the roller and passes the image sensor. The image sensor detects a 20 denomination it will count as 20php. If the device did not detect a 20 denomination it proceeds to detect another denomination. For example, the 50 denomination counts as 50php. The 100 denomination counts as 100php, the 200 denomination is the 200php, 500 denomination is the 500php bill. And lastly, the 1000 denomination counts as 1000php. If the system does not detect any Peso bill, the device will automatically stop and then show the total amount and number of bills in each denomination. Which was displayed on the LCD. After that, the process will repeat. 11 | P a g e Figure 3 Flowchart of the System Creating of Casing The casing of the device is 12 inches long, 16 inches wide, and 2.5 inches tall. The opening of the bin was 11.25 inches long, and 6 inches tall. The illustration of the casing is shown in the figure below: 12 | P a g e Figure 4 Cutting of the Acrylic Casing The device was installed inside the casing without the wires being removed. The physical design of the case consisted of 3 major parts. The first part was the opening made of an unused tray of a printer where the peso paper bills get placed. The second part is where the batteries, Arduino Nano, IR Sensors, Motors, and other components are covered with acrylic sheets that open left and right attached to hinges. The uncovered part is placed on an acrylic sheet where the peso paper bills get sensed by the Image Sensor. Assembling of the Device The setup 1 of the device consisted of an Arduino Nano, 2 IR sensors, and an image sensor. The IR sensor determined the presence of bills. The Image sensor distinguishes the different types of bills. An enclosure was put to the device. This enclosure is the main component that holds everything in it. The sensors were connected to the pins. The first IR sensor was connected to the A3 pin, the 5v pin, and the ground (GND) pin. The second IR 13 | P a g e sensor was connected to the A2 pin, the 5v pin, and the ground (GND) pin. The Image sensor was connected to the D7, D8, the ground (GND) pin, and the VIN input pin. The setup 2 of the device consisted of Arduino Nano, 4 IR sensors, and an image sensor. The IR sensor determined the presence of bills. The Image sensor distinguishes different types of bills. An enclosure was put to the device. This enclosure is the main component that holds everything in it. The sensors were connected to the pins. The first IR sensor was connected to the A3 pin, the 5v pin, and the ground (GND) pin. The second IR sensor was connected to the A2 pin, the 5v pin, and the ground (GND) pin. The third IR sensor was connected to the A1 pin, the 5v pin, and the ground (GND) pin. The fourth IR sensor was connected to the A0 pin, the 5v pin, and the ground (GND) pin. The Image sensor was connected to the D7, D8, the ground (GND) pin, and the VIN input pin. Testing for Accuracy The data taken in this test was about the ability of the device to correctly identify a peso paper bill introduced to it. The number of correct responses was counted for the calculation of the percent accuracy which was done by dividing the number of correct responses and the total number of trials for every Peso paper bill and multiplying it by 100 as follows: Figure 5. Percent Accuracy formula Testing for Efficiency The response time was measured using the oscilloscope. The ground (GND) pin of the Arduino Nano was connected to the ground (GND) terminal of the oscilloscope using the probe of the oscilloscope. The input channel of the oscilloscope was connected to the analog pin GPIO 33 of the Arduino Nano also using the probe of the oscilloscope. The efficiency was also 14 | P a g e measured by the response time in milliseconds. The process began when the command was given. The data was calculated by taking the average response time in milliseconds of the device. The oscilloscope started to read the response time if the device would correctly count the exact amount of the paper bill. This process would be repeated for 15 trials in 2 setups with 6 different peso paper bills per setup. The data was presented in a table showing how much time each setup took and was calculated by taking the average time of all trials in all setups. Data Analysis A normality test was conducted to know if the response time values are normally distributed. Using the Kolmogorov-Smirnov normality test, both Setup 1 and Setup 2 showed a normal distribution, as the p-values for each setup were greater than 0.05. A parametric test was conducted in the form of a t-test to measure if the data from both setups 1 and 2 has a significant difference and know what setup was the most efficient. 15 | P a g e  Results There is no misidentified Peso paper bill on testing the percent accuracy in all trials in all setups. The SwiftCounter showed a full 100% percent accuracy in all setups in every Peso paper bill (Table 1). Table 1 Percent Accuracy of the SwiftCounter System in terms of the number of correct responses in all trials in all setups Philippine Peso Setup 1 Setup 2 Paper Bill (Image sensor with 2 IR (Image sensor with 4 IR sensors) sensors) 20 – peso bill 100% 100% 50 – peso bill 100% 100% 100 – peso bill 100% 100% 500 – peso bill 100% 100% 1000 – peso bill 100% 100% The highest average response time in milliseconds in Setup 1 was recorded to be 641.14 and the lowest average is 638.15 ms in all trials in every Peso paper bill (Table 3). The response time measured were similar among all trials signifying the consistency detection rates. 16 | P a g e Table 2 Response time in milliseconds in detecting the Peso paper bills in all trials in all setups of the SwiftCounter system using one Image sensor and 2 IR sensors (Setup 1) Trials 20 – peso 50 – peso 100 – peso 200 – peso 500 – peso 1000– bill bill bill bill bill peso bill 1 639 640 641 635 641 638 2 640 640 640 640 640 640 3 635 635 639 640 641 638 4 637 640 640 640 641 638 5 637 640 640 640 640 640 6 638 635 641 635 639 638 7 640 639 640 637 640 640 8 635 640 641 640 641 639 9 640 638 641 640 639 641 10 640 640 640 638 641 639 11 640 639 641 640 640 638 12 637 640 640 638 638 638 13 640 638 640 640 640 639 14 639 640 640 639 641 638 15 638 640 641 640 641 638 Average 638.15 639.17 641.14 638.83 640.15 638.40 17 | P a g e  The highest average response time in Setup 2 is 638.87 ms and the lowest average is 636.13 ms in all trials in every Peso paper bill in terms of the efficiency of the response time (Table 3). The response time measured were similar among all trials signifying the consistency detection rates. Table 3 Response time in milliseconds in detecting the Peso paper bills in all trials in all setups of the SwiftCounter system using one Image sensor and 4 IR sensors (Setup 2) Trials 20 – peso 50 – peso 100 – peso 200 – peso 500 – peso 1000 – bill bill bill bill bill peso bill 1 635 636 630 635 638 638 2 630 635 635 639 638 634 3 635 639 635 650 635 639 4 630 635 635 638 634 645 5 635 635 645 639 635 630 6 635 635 636 635 630 639 7 636 650 636 640 634 640 8 634 645 638 645 638 639 9 639 639 635 638 638 635 10 634 645 652 639 650 639 11 645 638 639 635 638 630 12 635 638 634 635 635 638 13 645 639 639 635 634 639 14 635 635 635 639 638 638 15 639 639 635 635 634 635 Average 636.13 638.87 637.27 638.47 636.6 637.2 18 | P a g e  Kolmogorov-Smirnov and Shapiro-Wilk tests for Normality showed that there is no significant difference in both setups 1 and 2 since the calculated p-value of 0.200 for both setups is greater than the 0.05 level of significance. This result implies that there is a normal distribution of data for setups 1 and 2. Table 4 Kolmogorov-Smirnov and Shapiro-Wilk Normality Test Results of Efficiency of the System in Terms of Response Time Kolmogorov-Smirnov Shapiro-Wilk Statistic degrees of Significance Statistic Degrees of Significance freedom freedom Setup.214 6.200.922 6.519 1 Setup.224 6.200.937 6.633 2 19 | P a g e  There is a significant difference in the response time in identifying the correct denomination of every peso paper bill among all trials among the two setups (Table 5) since the p-value of 0.007 calculated is less than the 0.05 level of significance. The mean difference calculated is 2.12333. Table 5 t-test Results for Efficiency of the System in Terms of Response Time in Milliseconds Tvalue df Sig. (2-tailed) Mean Std. Error Difference Difference Equals 3.414 10.007 2.12333.62199 Variances Assumed Equals 3.414 9.989.007 2.12333.62199 Variances not Assumed Since it was found out that there is a significant difference in the response time in identifying the correct denomination of every peso paper bill among all trials between the two setups. Setup 1 has a higher response time average of 639.55 compared to Setup 2 with 637.42. Table 6 Mean of the Two Setups Setup N Mean Standard Standard Error Deviation Mean Setup 1 6 639.5467 1.09489.44699 Setup 2 6 637.4233 1.05947.43252 20 | P a g e  Discussion The percent accuracy of the SwiftCounter shows 100% accuracy in all trials in all setups. Achieving 100% accuracy in image detection can be attributed to a combination of high quantum efficiency, thorough calibration, advanced processing algorithms, and effective noise reduction techniques. Additionally, the design quality and testing conditions further contribute to ensuring that no mistakes were committed by the image sensor during the detection test. Such precision indicates that the sensor reliably processes input data without errors, which is critical for achieving effective results in practical applications (York, n.d.). All 4 IR sensors were tested for functionality, the peso bills were directly placed in front of the 4 IR sensors. Each LED of the 4 IR sensors illuminated and the peso bills were detected. As supported by the result of the study IoT-Enabled Covert Surveillance Camera Identification through IR Sensor conducted by Ruksana, et al. (2024). stating that the effectiveness of the IR sensor enables quick and reliable identification, highlighting the capability of IR technology in object detection. The combination of IR and image sensors facilitates a seamless and efficient counting process, ensuring that each paper bill is accounted for accurately and swiftly. Moreover, with the integration of advanced counting algorithms and real-time data processing, any potential errors in counting due to issues such as misalignment or bill jams can be rapidly resolved. The accuracy of your device is attributed to the strategic implementation of multiple IR and image sensors that collectively enhance detection capabilities, facilitate bill recognition, and ensure error-free counting through robust multi-point validation (Penny, 2024). There is a significant difference in the response time among all trials in all setups having Setup 2 with the lowest mean of response time of 637.42ms compared to the 639.6547ms response time of setup 1. This indicates that setup 2 is more efficient since it has a lower average response time implying a faster detection response. With this, the null hypothesis is not 21 | P a g e established. Utilizing multiple IR sensors can provide faster and more reliable detection of objects. This is primarily because the system has access to a broader range of spatial data, which allows it to make quicker decisions in response to incoming signals. The ability to collect and analyze data from various angles and distances leads to a reduction in the time taken to identify and respond to stimuli. This claim is supported by the result of the study conducted by Fiorino et al., (2022) additionally stating that having multiple IR sensors enables more reliable assessments by capturing a wider range of data simultaneously, thus leading to improved response times and more accurate quantification. The use of four IR sensors significantly increases the detection coverage area compared to just two sensors. Each IR sensor has a specific field of view. The configuration with four sensors expands the overall area being monitored, which allows for more opportunities to detect objects as they pass through the system. This is crucial in high-speed applications, such as those involving fast-moving paper bills. The deployment of four sensors minimizes potential blind spots that could be present with only two. In cases where bills overlap or are not perfectly aligned, having additional sensors ensures that at least one can detect the bill effectively. This leads to fewer missed detections, which enhances the speed of processing as bills are counted more reliably. The performance benefits of multiple sensors directly correlate with enhanced response times in the device. With four sensors, the device can detect multiple bills at once. This parallel detection method accelerates the overall counting process, as the system does not need to sequentially wait for one sensor to complete its task before the next sensors can activate. Each IR sensor generates a signal upon detection. Utilizing four sensors increases the number of signals simultaneously processed, facilitating quicker signal processing. The control system can employ parallel processing techniques, handling inputs from multiple sensors at once and decreasing response times effectively 22 | P a g e  Utilizing additional IR sensors improves the quality of the detection signals. Increased numbers of sensors provide a redundancy factor. If one sensor faces interference or malfunction, the others can continue detecting bills, ensuring consistent operation and minimizing disruptions in the counting process. With multiple sensors, the system can cross- verify detections in real time. If one sensor identifies a bill, the others can confirm the detection. This enhances confidence in the readings and can expedite decision-making regarding bill processing, thereby contributing to faster overall response rates (5.7 IR Detector Characteristics and Performance - HST User Documentation, n.d.). Conclusion It can be concluded that the percent accuracy of the SwiftCounter showed a full 100% accuracy in all trials in all setups, which implies that the SwiftCounter is efficient in identifying every denomination of the Peso paper bill. There is a significant difference in the efficiency of the SwiftCounter in terms of the response time in milliseconds among all trials in all setups which means the system does not possess the same potential in the response time in identifying the correct denomination of every peso paper bill identifying the correct denomination of every peso paper bill since the calculated p-value of.007 is less than 0.05 level of significance. It implied that one setup overpowered the other one in terms of response rate having Setup 2 (Image sensor and 4 IR sensor) as the most efficient with a lowest average response rate of 637.42 ms compared to Setup 1 with 639.55. Thus, both setups 1 and 2 possessed the same accuracy but the most efficient setup with fastest response rate is Setup 2 (Image sensor and 4 IR sensor). 23 | P a g e  Recommendation To improve the study, materials, and equipment should be ready before conducting the experimentation. During the testing, recommended that the laptop is suitable for the Arduino IDE application, and for the coding, it is recommended that a data cable should be used for uploading the code. It would also be rational to consider adding a color sensor for a more accurate detection and efficient response. To add to the functionality of the device, it is also recommended to add an ultraviolet (UV) sensor for counterfeit detection. Lastly, it is also advantageous to use the HP deskjet ink advantage 2060 model printer for its roller is suitable for an easy and smooth passing through of the peso paper bills. 24 | P a g e  References Fiorino, S. T., Raut, Y., Slabaugh, L., Erickson, A., Schmidt, J., Keefer, K., & McCrae, J. (2022). Quantifying surface layer moisture flux from MWIR imagery. Semantic Scholar. https://www.semanticscholar.org/paper/Quantifying-surface-layer-moisture- flux-from-MWIR-Fiorino-Raut/2e152bbc46d17fb2b0a645e5c2e8fda927df83f9 Huang, Y. (2022, December 25). The ultimate guide to money counting machines. RIBAO TECHNOLOGY USA. https://www.ribaostore.com/blogs/news/the-ultimate-guide-to- money-counting-machine LE RAYON INTERNATIONAL. (2020). Quotation for good quality counting machine. Penny. (2024, September 19). The ultimate guide to money counting machines | MUNBYN Blog. MUNBYN® Business. https://pos.munbyn.com/blog/the-ultimate-guide-to- money-counting- machines/#:~:text=Money%20counting%20machines%2C%20also,identify%20count erfeits.&text=authenticity%20and%20count%20the Ruksana, S., Imran, S., Farheen, S. F., Kireeti, K. M., & Sailaja, K. (2024). IoT-enabled covert surveillance camera identification through IR sensor. Semantic Scholar. https://www.semanticscholar.org/paper/IoT-Enabled-Covert-Surveillance-Camera- through-IR-Ruksana-Imran/755b76ff5223a889295df8e03a882bd6d61524d0 York, T. (n.d.). Fundamentals of image sensor performance. https://tinyurl.com/y9hk6std 5.7 IR detector characteristics and performance - HST user documentation. (n.d.). Hubble Space Telescope. https://hst-docs.stsci.edu/wfc3ihb/chapter-5-wfc3-detector- characteristics-and-performance/5-7-ir-detector-characteristics-and- performance#:~:text=Chapter%202%3A%20WFC3%20Instrument%20Description  Appendix A Figure 6. Assembling of the device Figure 7. Uploading of Code a) Assembling of Casing b) Uploading of Code to the device Figure 8. Testing the Efficiency Figure 9. Testing the Efficiency c) Testing the efficiency of the device d) Testing the accuracy of the device  Appendix B Table 7 Number of Correct and Incorrect responses in Setup 1 (Image sensor and 2 IR sensors) and in Setup 2 (Image sensor and 4 IR sensors) Correct and Incorrect response of the SwiftCounter System Setup 1 Setup 2 Trials 20php 50php 100php 200php 500php 1000php 1 / / / / / / 2 / / / / / / 3 / / / / / / 4 / / / / / / 5 / / / / / / 6 / / / / / / 7 / / / / / / 8 / / / / / / 9 / / / / / / 10 / / / / / / 11 / / / / / / 12 / / / / / / 13 / / / / / / 14 / / / / / / 15 / / / / / / Note. / - Correctly identified Peso paper bill X – Incorrectly identified Peso paper bill Table 8 Group Statistics of The SwiftCounter Setup N Mean Std. Std. Error Deviation Mean Values Setup 1 6 639.5467 1.09489.44699 Setup 2 6 637.4233 1.05947.43252 Table 9 Levene’s Test for Equality of Variances F Sig. Values Equal variances assumed.010.921 Equal Variances not assumed Table 10 t-test for Equality of Means t df Sig. (2- Mean Std. Error 95% Confidence tailed) Difference Difference Interval of the Difference Lower Upper Values Equal 3.414 10.007 2.12333.62199.73745 3.50922 variances assumed Equal 3.414 9.989.007 2.12333.62199.73724 3.50942 Variances not assumed SwiftCounter: An Advanced Portable Device for Accurate Peso Paper Currency Counting RESEARCH PLAN A. Rationale In the past, banks and other financial industries were fully reliant on human operations to count money. However, this resulted in numerous counting errors in businesses and financial institutions all over the world. This resulted in the invention of the money-counting devices in the 1920s. Money counting machines, also known as currency counters or bill counters, are electronic devices used to verify the authenticity and count the number of banknotes (Penny, 2024c). It is not possible to offer an exact assessment of the overall accuracy of money-counting machines since various devices will have various error rates. Many money-counting devices, on the other hand, have an error rate of less than 1% which is regarded as quite with an accuracy rate of 99.99% (Ribao Technology, 2022). High maintenance cost of currency counting machines and the continuous power supply required to operate these machines are expected to hamper the market growth (Currency Counting Machines Market, 2023). For the past years, several types of money-counting devices have been made to prevent errors in the finance industry and make everything easy. They count the money using infrared sensors and can even identify between various denominations. Like automatic detection with UV(Ultraviolet) and MG (Magnetic), IR, MT, DD (LE RAYON INTERNATIONAL, 2020) This ensures that counterfeit bills are identified and prevented. The Philippine Peso refers to the Philippines’s official currency and is represented by ISO code PHP (Team, 2024b). They can count money at an average rate of 600 to 2000 bills per minute or higher. Money counters provide a level of accuracy not possible when counting by hand (Ribao Technology, 2022). Counting one cash denomination at a time can be time- consuming and insufficient, especially at the finance industry, banks, and businesses. B. Research Question The main goal of the study is to develop a Peso paper bill portable amount counter. Specifically, it aims to answer the following questions: 1. What is the circuit design of the SwiftCounter system? 2. What is the flowchart of the SwiftCounter system? 3. What is the percent accuracy of the SwiftCounter system in detecting and counting the amount of the introduced peso paper bills in the following setups: a. Image sensor with 2 Infrared sensors; and b. Image sensor with 4 Infrared sensors? 4. What is the efficiency of the SwiftCounter system in terms of the response time in milliseconds among all trials in all setups? 5. Is there a significant difference in the efficiency of the SwiftCounter in terms of the response time in milliseconds among all trials in all setups? C. Hypotheses There is no significant difference in the efficiency of SwiftCounter counting system in terms of the response time in milliseconds among all trials in all setups. D. Procedures Research Locale The study will be conducted within the vicinity of Koronadal City, specifically in Koronadal National Comprehensive High School (KNCHS) Science Laboratory in all phases of the study. The laboratory is wide enough to cater the process particularly the assembly of the device and the testing for accuracy and efficiency. The place also has all the necessary materials needed for the testing of the device. Research Design Developmental research design, posttest experimental research design, and will be used in this study. Assembly of the SwiftCounter will be the development stage, testing the effect of using two different sensors on the efficiency of the SwiftCounter is the experimental stage and post-test design will be used for the testing of the SwiftCounter’s response time in detecting different types of bills. Two setups will have 15 trials each, with a total of 30 trials to test the efficiency of the SwiftCounter. The first phase will be the conceptualization and identification of the problem. The second phase will be the planning of the study. During this phase, a cost analysis of the materials will be done. The third phase will be the development of the SwiftCounter. In this phase, the gathering and assembling of materials will be done. The testing and gathering of results will be the fourth phase of the study. The efficiency of the device will be tested in this phase. In the fifth phase, the analysis of data from the experimentation will be done. The results and discussion of the study will be done in the sixth phase. During this phase, the conclusion will be discussed and the hypotheses will be approved or rejected. Gathering of Materials Commonly used materials such as the Arduino Nano, infrared sensor, battery, and wires will be purchased from a local store within Koronadal City. Materials not available within Koronadal City, like the Huskylens, DC motor gear, LCD display, and 10K resistor, were bought online. Circuit Design There will be two circuit designs for this study given that there will be two setups. For the first circuit design of the study includes an image sensor, DC motors, 5V relays, LCD display, DC - DC buck converter and 2 IR sensors which will be then connected to the Arduino Nano according to the corresponding pins as shown in the figure below. Figure 1. Circuit Design of Setup 1  The second circuit design of the study includes an image sensor, DC motors, 5V relays, LCD display, DC - DC buck converter and 4 IR sensors which will be then connected to the Arduino Nano according to the corresponding pins as shown in the figure below. Figure 2. Circuit Design of Setup 2 Interfacing the Huskylens AI Image Sensor The VCC pin of the Huskylens AI Image Sensor will be connected to the VIN of the board. Then, the ground (GND) pin will be connected to the ground (GND) of the board. The SDA pin of the sensor will connect to the Digital input pin D7 of the board, and the SCL pin of the sensor will connect to the Digital input pin D8 of the board. This will enable data reception and transmission between the Image Sensor and the Arduino. Interfacing the Four (4) IR Sensors The VCC pins of all four IR Sensors will be connected to the 5V pin of the Arduino Nano board. The ground (GND) pins of all four IR Sensors will connect to the ground (GND) pin of the Arduino Nano board. The OUT pins of the four IR Sensors will connect to the Analog pins A0, A1, A2, and A3 of the Arduino Nano board. Interfacing the Two (2) DC Motors and Relays The negative terminal of the first DC Motor will be connected to the IN+ of the Buck Converter, while the negative terminal of the second DC Motor will connect to the OUT- of the Buck Converter. The GND pin will connect to any ground of the circuit. The positive terminal of the first DC Motor will be connected to the Normally Closed (NC) Contact of the 5V Relay, and the Common Contact of the 5V Relay will connect to the IN- of the Buck Converter. Similarly, the positive terminal of the second DC Motor will connect to the NC Contact of the 5V Relay, and the Common Contact of the 5V Relay will connect to the OUT+ of the Buck Converter. The IN of the first 5V Relay will be connected to the D10 pin of the Arduino Nano, and the VCC of the 5V Relay will connect to the OUT+ of the Buck Converter. The IN of the second 5V Relay will be connected to the D11 pin of the Arduino Nano, and its VCC will also connect to the OUT+ of the Buck Converter. Interfacing the Charging Module and Buck Converter The positive terminal of the 12V Lithium-Ion Battery will be connected to the on/off switch, leading to the VIN of the Arduino Nano board, while the ground of the battery will be connected to the ground of the Arduino Nano board. The positive side of the 12V Lithium-Ion Battery will connect to the IN+ of the Buck Converter, and the negative side will connect to the IN- of the Buck Converter. This will power the IR sensors, Huskylens, DC Motors, and LCD Screen. Interfacing Other Components The GND of the LCD screen will connect to the ground of the Swift Counter, and the VCC of the LCD screen will connect to the VIN of the Arduino Nano. The SDA of the LCD screen will be connected to the Analog pin A4, and the SCL will connect to the Analog pin A5 of the Arduino Nano. This setup will allow the display of the types of peso paper bills counted and the total amount. After connecting all the wires and pins, the Arduino Nano board with the Arduino IDE application will be programmed, and the code will be uploaded. The Swift Counter will then be switched on, and different types of peso paper bills will be placed for testing the recognition of the bills and overall functionality based on the programmed instructions. Flowchart of the Design The device starts when the command is given. If the IR sensor detects any presence of a peso bill, then the peso bill then proceeds into the roller and passes the image sensor. The image sensor detects a 20 denomination it will count as 20php. If the device did not detect an 20 denomination it proceeds to detect another denomination. For example, the 50 denomination counts as 50php. The 100 denomination counts as 100php, the 200 denomination is the 200php, 500 denomination is the 500php bill. And lastly, the 1000 denomination counts as 1000php. If the system does not detect any Peso bill, the device will automatically stop and then show the total amount and number of bills in each denomination. Which was displayed on the LCD. After that, the process will repeat. Figure 3. Flowchart of the Design Creating of Casing The casing of the device is 12 inches long, 16 inches wide, and 2.5 inches tall. The opening of the bin was 11.25 inches long, and 6 inches tall. The illustration of the casing is shown in the figure below: Figure 4. Cutting of the casing The device will be installed inside the casing without the wires being removed. The physical design of the case will consist of 3 major parts. The first part will be the opening made from an unused tray of a printer where the peso paper bills will be placed. The second part will be where the batteries, Arduino Nano, IR Sensors, Motors, and other components will be covered with acrylic sheets that open left and right, attached with hinges. The uncovered part will have an acrylic sheet where the peso paper bills will be sensed by the Image Sensor. Assembling of the Device The setup 1 of the device will consist of an Arduino Nano, 2 IR sensors, and an image sensor. The IR sensor will determine the presence of bills. The Image sensor will distinguish the different types of bills. An enclosure will be put to the device. This enclosure is the main component that held everything in it. The sensors will be connected to the pins. The first IR sensor will be connected to the A3 pin, the 5v pin and the ground (GND) pin. The second IR sensor will be connected to A2 pin, the 5v pin and the ground (GND) pin. The Image sensor will be connected to the D7, D8, the ground (GND) pin and the VIN input pin.  The setup 2 of the device will consist of Arduino Nano, 4 IR sensors, and an image sensor. The IR sensor will determine the presence of bills. The Image sensor will distinguish different types of bills. An enclosure was put to the device. This enclosure is the main component that held everything in it. The sensors will be connected to the pins. The first IR sensor will be connected to the A3 pin, the 5v pin and the ground (GND) pin. The second IR sensor will be connected to the A2 pin, the 5v pin and the ground (GND) pin. The third IR sensor will be connected to the A1 pin, the 5v pin and the ground (GND) pin. The fourth IR sensor will be connected to the A0 pin, the 5v pin and the ground (GND) pin. The Image sensor will be connected to the D7, D8, the ground (GND) pin and the VIN input pin. Testing for Accuracy The data taken in this test will focus on the device’s ability to correctly identify a peso paper bill introduced to it. The number of correct responses will be counted to calculate the percent accuracy, which will be done by dividing the number of correct responses by the total number of trials for each peso paper bill and multiplying by 100 as follows: Figure 5. Percent Accuracy Formula Testing for Efficiency The response time will be calculated using the oscilloscope. The ground (GND) pin of the Arduino Nano is connected to the ground (GND) terminal of the oscilloscope using the probe of the oscilloscope. The input channel of the oscilloscope is connected to any digital input pin of the Arduino Nano also using the probe of the oscilloscope. The efficiency will also be measured by the response time in milliseconds. The process will begin once the command is given. The data will be calculated by taking the average response time in milliseconds of the device. The oscilloscope will start to read the response time if the device will correctly count the exact amount of the paper bill. This process will be repeated for 15 trials in 2 setups. The data will be presented in a table showing how much time each setup took and will be calculated by taking the average time of all trials in all setups. E. Risk and Safety When handling the circuitry of the system, it is important to adhere in wearing appropriate personal protective equipment, including electrical gloves, to prevent electrical shock and ensure safety. Utilize the correct tools specifically designed for installing pins in the Arduino board to avoid damaging components and to ensure precision. All activities involving circuitry should be conducted under the supervision of a research supervisor to ensure compliance with safety protocols and to provide guidance during the installation process. F. Data Analysis The initial step in our data analysis will start by conducting a Kolmogorov-Smirnov normality test to know whether the collected data follows a normal distribution. If the test reveals a normal distribution, subsequent analysis will involve parametric tests. Given by the 2 setups, we will apply both Levene's Test and the T-test to determine the differences. On the other hand, if the data fails the normality test, indicating a non-normal distribution, it will proceed with a non-parametric test. Specifically, the Mann-Whitney U Test will be applied to compare the 2 setups. G. Bibliography Fiorino, S. T., Raut, Y., Slabaugh, L., Erickson, A., Schmidt, J., Keefer, K., & McCrae, J. (2022). Quantifying surface layer moisture flux from MWIR imagery. Semantic Scholar. https://www.semanticscholar.org/paper/Quantifying-surface-layer-moisture- flux-from-MWIR-Fiorino-Raut/2e152bbc46d17fb2b0a645e5c2e8fda927df83f9 Huang, Y. (2022, December 25). The ultimate guide to money counting machines. RIBAO TECHNOLOGY USA. https://www.ribaostore.com/blogs/news/the-ultimate-guide-to- money-counting-machine LE RAYON INTERNATIONAL. (2020). Quotation for good quality counting machine. Penny. (2024, September 19). The ultimate guide to money counting machines | MUNBYN Blog. MUNBYN® Business. https://pos.munbyn.com/blog/the-ultimate-guide-to- money-counting- machines/#:~:text=Money%20counting%20machines%2C%20also,identify%20count erfeits.&text=authenticity%20and%20count%20the Ruksana, S., Imran, S., Farheen, S. F., Kireeti, K. M., & Sailaja, K. (2024). IoT-enabled covert surveillance camera identification through IR sensor. Semantic Scholar. https://www.semanticscholar.org/paper/IoT-Enabled-Covert-Surveillance-Camera- through-IR-Ruksana-Imran/755b76ff5223a889295df8e03a882bd6d61524d0 York, T. (n.d.). Fundamentals of image sensor performance. https://tinyurl.com/y9hk6std 5.7 IR detector characteristics and performance - HST user documentation. (n.d.). Hubble Space Telescope. https://hst-docs.stsci.edu/wfc3ihb/chapter-5-wfc3-detector- characteristics-and-performance/5-7-ir-detector-characteristics-and- performance#:~:text=Chapter%202%3A%20WFC3%20Instrument%20Description SwiftCounter: An Advance Portable Device for Accurate Peso Paper Currency Counting PROJECT DATA LOGBOOK SwiftCounter: An Advanced Portable Device for Accurate Peso Paper Currency Counting FORMS