Estimating Position of Bio Electric Potential Dataset as A Natural Sensor using Time Series Approach Imam Tahyudin1,2 , Hidetaka Nambo1 1 Artificial Intelligence Laboratory, Graduate School of Natural Science and Technology Division of Electrical Engineering and Computer Science, Kanazawa University, Japan 2 Department of Information System, STMIK AMIKOM Purwokerto, Indonesia imam@blitz.ec.t.kanazawa-u.ac.jp Abstract of which is the death that is not known by others, whether caused by accidents in the home or other factors such as mur- The application of plants as natural sensors to de- der. Based on research by the same number of deaths caused tect human behavior is very interesting to investi- by accidents in the home because it was not helped as much gate. One benefit is to know the position of elderly as 12.5%. This condition be attention for all parties, includ- people living alone in a house in order to avoid acci- ing the researcher. One of the measures being initiated is to dents resulting in death because of not immediately examine the installation of cameras in their house so that ac- helped. The previous authors already use some cidents that occur immediately known by a neighbor or an au- method to estimate the position such as classifica- thorized officer so it can be helped. However, this solution is tion and multilayer perceptron methods. However, less widely accepted because of privacy concerns, therefore, to find the best estimation is so difficult. There- the study conducted to make plants as a camera to monitor fore, this study tried to use time series approach to the location of the elderly activity at home. It is known as solve bioelectric potential dataset problem because Bio electric potential. the data type is numeric and data stored are based on time. The time series model which is used is Au- Basic use leafy plants are because it can be treated with the toregressive (AR) model. The purpose of this study installation of electrodes on the leaf that can produce low- is to estimate the position based on the distance voltage electrical signal but it also can be used as a room of training dataset to AR model. The result per- freshener that impact both on the health of its inhabitants formed that the AR model selected is AR of order 3. like to reduce stress. Plant bio-electric potential generates In addition, the estimation accuracy is pretty good an electrical signal low because the activity of the plant such of 75% compared with the other methods, such as as photosynthesis and transpiration, but it is also due to envi- multi layer perceptron or decision tree. ronmental factors such as temperature, humidity living things and human behavior around [Shimbo and Oyabu, 2004]. The use of bio electric potential as a natural sensor is an 1 Introduction innovation in an effort to detect the human behavior like ac- In 2014, a publication of the aging society published by the cident in order to prevent the eldery death who live alone Japanese cabinet office, announced in October 2010 and Oc- in a home because of accidents are not helped immediately. tober 2013 which there are 23% and 25.1% of the elderly A previous attempt to use the camera a lot of rejection be- population respectively [Nomura et al., 2016]. Then, from cause the monitor in place of privacy as the bathroom and the the resulting this percentage, the number of elderly people bedroom [Shimbo and Oyabu, 2004], [Nomura et al., 2014], in Japan the average age is more than 65 years. This con- [Nambo and Kimura, 2017], [Nambo, 2015], [Nambo and dition is the highest proportion in the world [Nomura et al., Kimura, 2016]. Besides, the use of infrared sensors tested to 2016][Chen et al., 2016]. solve this problem, although the results were pretty good but Based on the research of Nomura et al [Nomura et al., costly due to capture human behavior requires many sensor 2016], the condition of the elderly is mapped into two groups: cells, so that are not economically [Jin, 2014]. Then the other the elderly who live with their families and the elderly who solutions tested using the sense of odor but the results are not live alone. Based on data from samples taken in one of so good because there is often noisy when the data records the major provinces in Japan Kyoto mention that the num- [Jin, 2014]. Hence, the use of bio-electric potential could ber of the first group in 1990 is 284.013 (86.7%). After be the solution to these problems because it is friendly to 20 years in 2010, it increased nearly two-fold as many as monitor the behavior of people in privacy place and may also 495.343 (81.8%). Next, the second group in 1990 is by be a producer media of oxygen to reduce stress (for healing) 43.416 (13.3%) then in 2010 increased to 110.366 (18.2 [Shimbo and Oyabu, 2004], [Nomura et al., 2014], [Nambo There is an interesting phenomenon of the second group, and Kimura, 2017], [Nambo, 2015], [Nambo and Kimura, that the proportion of elderly people living alone in 2010 is 2016]. high at 18.2%. These conditions lead to various problems one Furthermore, based on the results obtained from previous studies that bioelectric potential has the ability to capture hu- say that the scientific vision of Artificial Intelligence (AI) man behavior well. Research conducted by Hirobayashi et can be successfully synthesized by the words of Pat Lang- al [Hirobayashi et al., 2007], states that human activities like ley: ”AI aims to understand and reproduce computing sys- stepping around the plants produce a strong correlation with tems of various intelligent behaviors observed by humans” changes the signal by using plant bio-electrical potential. An- (Langley, 2012) [Lieto and Radicioni, 2016]. Other studies other study conducted Nomura et al [Nomura et al., 2014], discussed Artificial Cognitive systems, by A. Lieto and M. Shimbo et al [Shimbo and Oyabu, 2004] with using machine Cruciani. This paper presents the AI collaborative studies of learning method that shows the results of human behavior many disciplines such as computer scientists, psychologists, such as talking, moving, walking, opening the door can be engineers, philosophers, linguists and biologists. This collab- detected using bioelectrical plant potential. Subsequent re- oration led to its influence on the study of natural and artificial search conducted by Jin et al [Jin, 2014] using artificial neu- systems. The author describes the use of many AI cognitive ral network algorithm successfully detects a distance of per- system frameworks such as SOAR and A-SOM as well as son within the plant of bioelectric potential. Then another the development of research on Artificial Cognitive systems study conducted by Nambo et al utilize bioelectric potential [Lieto and Cruciani, 2017]. Paper ”Cognitive System: Argu- for determining the position in a room. This study uses sev- mentation and Cognition” was performed by A. Kakas and eral algorithms including decision tree (J48) for the classifica- Michael. This paper discusses the relationship between argu- tion point and multi layer perceptron locations to determine ment and cognition from a psychological and computational the position and then make a regression model to matching point of view. In addition, this paper also investigates how process. The results obtained show that a person’s position the synthesis of work on reasoning and understanding of nar- can be estimated with an accuracy rate of 60% [Nambo and rative texts from the Cognitive Psychology of work which is Kimura, 2017], [Nambo, 2015], [Nambo and Kimura, 2016]. based on computational arguments from AI that can offer a Utilization of plants as a natural sensor is a breakthrough scientific and pragmatic basis for building human cognitive to help some problems in daily life of human activities. systems in performing everyday tasks [Kakas]. An evolu- Recorded data using plant media, can be used as input learn- tionary study of architectural cognitive frameworks, ICARUS ing by using various methods such as machine learning, was presented by D. Choi and P. Langley. The paper men- statistics and data mining. Furthermore, this research is ex- tions that two early versions of ICARUS explain in more de- pected as an effort to face a new era in data computing, known tail the third incarnation that has stabilized over the past 12 as AI cognitive system. That is a technologically advanced years. These include modules for conceptual inference based system that has learning features and can continue to adapt on perception, goal-based reactive execution, problem solv- just like a human brain. ing with shield analysis, skill acquisition of new solutions, The previous researches about artificial intelligence cogni- and achievement of top-level objectives [Choi and Langley, tive system are able to outline as follow. J. Suchan and M. 2017]. Bhatt study about Semantic Q and A with Video and eye- The paper ”The level of knowledge in cognitive architec- tracking data. This study is the foundation of AI for human ture: current limitations and developments” is written by A. visual perception which is studied based on Cognitive Film Lieto et al. The authors say that the level of cognitive archi- Studies. By using a demonstration of major technological ca- tecture (CA) is not only a technological issue but also episte- pabilities aimed at investigating the effects of attention and mological one, because they limit the comparison of knowl- recipients on motion pictures; These results have a high de- edge representation and CA processing mechanism to that of gree of analysis of the subject’s visual fixation patterns and humans in their daily activities. In addition, they say that it correlations with the semantic analysis of the dynamic visual should be tackled to build artificial agents that are capable of data [Suchan and Bhatt, 2011]. Research on cognitive pro- demonstrating intelligent behavior in a common scenario [Li- gramming is conducted by L. Michael et al. They explain eto et al., 2017]. J. Rosales et al., presented the integration of that by following a vision where humans and machines share cognitive computing model of planning and decision making the same level of common sense. They have proposed cog- by considering affective information. The contribution of this nitive programming as a means to build cognitive systems. research is that the resulting model considers affective infor- Cognitive programming adopts a machine view as a personal mation and motivation as a fundamental and essential trig- assistant. The point is that humans demand completion of ger in the planning and decision making process; In addition, tasks, perhaps without specifying fully and clearly determin- the model is capable of mimicking the internal human brain ing what is needed, but relying on the experience of the as- as well as human external behavior [8]. F. Martinez et al., sistant, and finally, the machine is able to perform the task. studied computational analysis of general intelligence tests to Cognitive programming aims to bring traditional program- evaluate cognitive development. They mentioned that to un- ming flexibility to existing technology users, enabling them derstand the role of basic cognitive operational construction to view their personal devices as novice assistants, who can (such as identity, difference, order, calculation, logic, etc.) re- receive training and personalization through natural interac- quired intelligence testing and serves as a proof of evaluation tions [Michael et al.]. concept on other developmental issues [Martnez et al., 2017]. A. Lieto and D.P Radicioni study about the Cognitive AI The paper ”Enabling the social intelligence of robots by the system. They reviewed the major historical and technological engineering of human social-cognitive mechanisms” is writ- elements that characterize the recent rise, fall and resurgence ten by T.J. Wiltshire et al. This study explains the basic tech- of the cognitive approach to Artificial Intelligence. They niques of human social cognition to illustrate how the embod- ied social robot can be designed to function autonomously as an efficient co-worker. Adopting the engineering human so- cial cognition (EHSC) as an approach to modeling the socio- cognitive mechanisms in robots, not only provides a robust, flexible, sophisticated, cognitive, perceptual, motor and cog- nitive architecture, but also enables a more direct understand- ing of, and natural interaction with the environment and hu- man colleagues. It also provides a mechanism for better un- derstanding human behavior and mental states as well as en- abling the prediction and interpretation of new and complex social situations [Wiltshire et al., 2017]. D.L. Dowe and M.V. Herna discussed a universal psychometric that measures cog- nitive abilities in the machine kingdom. This paper presents the measurement of cognitive ability, and the creation of a Figure 1: Bio electric potential new foundation for redefining and mathematically formalizes the concept of cognitive tasks, evaluable subjects, interfaces, task choices, difficulties, agency response curves, and so on. ment of bio electric potential, positioning coordinates, exper- The authors explain that the AI Evaluation may receive im- iment design and datasets. section 3 presents results and dis- portant impacts from the cognitive system view characterized cussion. Then conclusions and future work are described in by its diversity and cognitive ability level, analysis of the rela- section 4. tionship between the spectrum of capabilities, the difference between characteristics and measuring tools, or borrowing of 2 Proposed Method item response theory, as well as other theories and concepts 2.1 Measurement of Bio electric Potential developed in psychometrics [Dowe and Herna, 2014]. The paper ”Advanced user assistance based on AI plan- To perform measurements using a data logger. Specifications ning” is conducted by S. Biundo et al. This study presents a data logger used is GRAPHTEC GL400-4. It measures the hybrid planning approach in detail and demonstrates its po- low voltage at an average altitude of sampling (approximately tential by describing the realization of various aid functions 1 kHz). This tool has four channels so that it can simulta- based on complex cognitive processes such as generation, neously measure voltage. For the measurement of electrical improvement, and explanation of the Author’s plan that the potential of plants by attaching electrodes on two different user’s instructions are given based on an action plan synthe- leaves then measured the voltage generated between both the sized by the hybrid planning system. If a particular action leaves. The measurements are stored on a PC in real time via execution fails due to some unexpected environment change, the local network (figure 1). for example, the system can help the user out of the situation by starting the plan improvement process. The resulting plan 2.2 Positioning coordinates overcomes a failing and stable situation by simply pointing P1 is plant of bio electric potential and M1-M3 are the loca- out the deviation from an indispensable starting plan. Finally, tions of the experiment. For the position coordinates of the an explanation of the plan can be given based on the analysis plants and experiment point is seen in Table 1. of the structure of the rich knowledge plan generated by the planner and also the planning process itself [Biundo et al., 2011]. Table 1: Coordinate position of observation point and plant This research attempts to estimate the position of human bio electric potential in a room. The method to solve this problem is using time Observation point X-Coordinate Y-Coordinate series method. The time series approach to bioelectric poten- tial dataset is interesting to perform because it is stored based M1 260 475 on time and indicate the specific pattern. The use of time M2 200 490 series for prediction has been done by Abdurahman (2014) M3 140 430 about time series algorithms which combined with Particle P 160 340 Swarm optimization [Nambo and Kimura, 2016]. In addition, the research about the discovery knowledge in time series databases. This study aims to predict an important attributes and extract rules in association analysis [Schluter, 2012]. Fur- 2.3 Experiment Design thermore, in 2015 conducted research on time series analysis The design of this research can be seen in Figure 2. Referring through AR modeling. In this study shows various types of this figure, the observed bioelectric potential data is divided AR models such as univariate and multivariate AR models, a into two types, namely training and testing dataset. In the radial base function autoregressive model and so on [Ohtsu et process of training data analysis are used Autoregressive time al., 2015]. series method. Through this method selected AR model that The research is described in the following sections: section is based on the level of feasibility and size of the standard 2 is talking about a research method includes the measure- error. Table 2: Standar error of AR model Parameter Position 1 Position 2 Position 3 AR(1) AR(2) AR(3) AR(1) AR(2) AR(3) AR(1) AR(2) AR(3) p-value 0 0 0 0 0 0 0 0 0 standar eror 0.045 0.038 0.033 0.044 0.038 0.033 0.045 0.038 0.033 Table 3: The component of each model Component Position 1 Position 2 Position 3 Intercept (a0 ) -9.31485E-05 -8.39632E-05 -9.24376E-05 X Variable 1 (a1 ) 0.665990568 0.66779352 0.668736298 X Variable 2 (a2 ) -0.061332529 -0.062073254 -0.062234636 X Variable 3 (a3 ) -0.497219399 -0.497083196 -0.496436779 Figure 2: The research design position 1 to 3 are almost the same (0.033). Therefore, the selected AR model is used to create a model at each position. Furthermore, to determine the position by using data test- ing which find the difference of the actual to the estimated Constructing AR Model value of AR model. According to the selected AR order, the models are con- structed as follows: 2.4 Dataset The table 3 explains the modeling component which is ob- Data were obtained by using plant of bioelectric potential in a tained at each position. The first component is constant value room of size 3 m x 4 m. The position of observation there are then the next three values are coefficients for the three vari- three points and the plant used there are two trees. The pro- able values of the previous time series in sequence. Models cess of recording data is done by someone by walking around that are formed in general are as follows: each position point for 30 seconds. Once the people move Yt =a0 +a1 Yt 1 +a2 Yt 2 +a3 Yt 3 on each observation point, the recording process starts from both bioelectric potential plants. Data obtained is spectrum Where : data format and can be converted into numerical data using Yt = signal value of bioelectric potential at t data logger. Therefore, from the experiment obtained three Yt 1 = signal value of bioelectric potential at t-1 dataset point position of two potential bioelectric plants so Yt 2 = signal value of bioelectric potential at t-2 that there are six total datasets. The first data collection is Yt 3 = signal value of bioelectric potential at t-3 used as training data. Next the second experiment in the same a0 ,a1 ,a3 = coefficient value for signal value of bioelectric po- way is used for data testing. tential at t, t-1, t-2, and t-3 Determination of data testing positions 3 Results and Discussion In this testing process there are five datasets are used. From 3.1 Experimental Setup each dataset we select the sequential data at time t to the order of data to 1500, 4000, 7500, 11500, and 13000 respectively. The data used is Bioelectric potential dataset from three po- The determination of this sequence is the basis of consider- sitions using one plant. Data analysis was performed using ation to seek and to cover the overall pattern in each dataset a MacBook Pro with specification: 2.7 GHz Intel Core i5, 8 (15000 observations). After that, the sequence is tested to the GB 1867 MHz, and DDR3. selected AR (3) model. Next, the calculation of the difference Best Model Determination of the estimation result and the actual test value. The smallest difference is selected as the estimated dataset position result. This research experiment was conducted in three positions. Here is the design and test results: To determine the best AR model performed in each position Based on the table 4, the estimation of the five datasets of the experimental object, by comparing the p-value and er- tested, there are three appropriate datasets and the rest are ror standard of each AR model (table 2). Based on the analy- sis results obtained p-value and standard error as follows: Based on the table 2, it is found that all p-values on each model of AR are zero because all of the p-value is smaller Table 4: The Position estimation Experiment position Position 1 Position 2 Position 3 Estimation result than alpha value (0.05) so that all models are selected as can- Position 1 0.039034078 0.039049273 0.039019109 Position 3 didate models. Next is a comparison of standard error values Position 2 0.025319258 0.025360625 0.025391858 Position 1 Position 3 0.038997821 0.03898176 0.038939225 Position 3 in each selected position and the result of the 3rd order of Position 1 0.021176831 0.021190519 -0.021182437 Position 1 AR model because the standard error value is smallest com- Position 2 0.020541127 0.020538297 0.020574728 Position 2 pared to the other AR order. The selected error values from missed. The first and the second experiment are fail because tion, we thank for anonymous reviewers who gave input and the estimation result is not true and the third to five experi- correction for improving this research. ment are accurate. Therefore, the value of accuracy obtained by 75%. 3.2 Discussion The research related to plant of Bio electric potential to esti- mate position is conducted in the simulation experiment. That experiment is performed in a small room size 3 m x 4 m. Then the number of plants used as much as two trees and the point of observation positions are three points. This means that if this research is successful it is possible to use in a house with several room as the observation positions and the number of bio electric potential plants that are used more than two trees. Based on the results which was conducted by using time se- ries approach, AR model has been able to estimate a person’s position with accuracy level of 75%. This level of accuracy is very important to be improved further with different methods approach. However, this accuracy level is quite competitive when compared with previous research like using multilayer perceptron and decision tree methods. The important point of this research that must be consid- ered is the contribution points. That this research contributes to determine someone’s position. Especially it is used to de- termine the position and movement of an elderly which liv- ing at home alone. We wish can help the elderly people for something unexpected happens. Utilization of plants as a me- dia for human activity aid oriented are more deeply studied. Included in the study of AI cognition system. Hopefully, the topic of bio electric potential can be developed further on the topic of AI cognition design like a smart robot. The plant is being smart by utilizing the data recording as a learning mate- rial to be used as a knowledge. Furthermore, with the knowl- edge possessed utilized for various human interests such as measurement of room temperature, distinguishing objects of life, counting the number of people in a space, measure the burning of calories and so forth. 4 Conclusion This study analyzes the time series method using AR model on Bioelectric potential dataset. Based on the analysis result obtained the best model that was AR (3). Furthermore, the selected model is used to estimate the position of data testing. Finally, this research obtained the estimation result accuracy of 75%. This research is exciting to be developed further by using optimization methods such as Steepest Ascent, Newton Raphson, or Particle swarm optimization, in order to increase an accuracy value. In addition, it is better to try using the others model of time series methods such as MA, ARMA and ARIMA. Acknowledgments This research was supported by various parties. We would like to thank Kanazawa University, Japan and Ministry of Re- search, Technology and Higher Education (RISTEKDIKTI) for scholarship program and thank to STMIK AMIKOM Pur- wokerto, Indonesia for all support. This research also sup- ported by JSPS KAKENHI Grant No. 17K00783. In addi- References Kenta Nomura, Hidetaka Nambo, and Haruhiko Kimura. De- velopment of Basic Human Behaviors Cognitive System Susanne Biundo, Pascal Bercher, Thomas Geier, Felix Mu, using Plant Bioelectric Potential. IEEJ Transactions on and Bernd Schattenberg. Advanced user assistance based Sensors and Micromachines, 134(7):206–211, 2014. on AI planning Action editor : Ute Schmid. 12:219–236, 2011. Mayuko Nomura, Stuart Mclean, Daisuke Miyamori, Ya- suhiro Kakiuchi, and Hiroshi Ikegaya. Science and Justice Brian K. Chen, Hawre Jalal, Hideki Hashimoto, Sze-chuan Isolation and unnatural death of elderly people in the aging Suen, Karen Eggleston, Michael Hurley, Lena Schoe- Japanese society. Science & Justice, 56(2):80–83, 2016. maker, and Jay Bhattacharya. Forecasting Trends in Dis- ability in a Super-Aging Society: Adapting the Future El- Kohei Ohtsu, Hui Peng, and Genshiro Kitagawa. Time Series derly Model to Japan. The Journal of the Economics of Modeling for Analysis and Control. 2015. Ageing, pages 1–10, 2016. Tim Schluter. Knowledge discovery from time series. (April Dongkyu Choi and Pat Langley. Evolution of the ICARUS 2012), 2012. Cognitive Architecture. Cogn. Syst. Res., 2017. Tatsuyo Shimbo and Takashi Oyabu. Statistical Analysis of Plant Bioelectric Potential for Communication with David L Dowe and M Victoria Herna. Universal psychomet- Humans. IEEJ Trans. on Sensor and Micromachines, rics : Measuring cognitive abilities in the machine king- 124(12):470–475, 2004. dom. 27:50–74, 2014. Jakob Suchan and Mehul Bhatt. Semantic Question- Shigeki Hirobayashi, Yusuke Tamura, Student Member, and Answering with Video and Eye-Tracking Data : AI Foun- Tatsuo Yamabuchi. Monitoring of Human Activity Using dations for Human Visual Perception Driven Cognitive Plant Bioelectric Potential. IEEJ Trans. on Sensor and Mi- Film Studies. pages 2633–2639, 2011. cromachines, 127(4):258–259, 2007. Travis J Wiltshire, Samantha F Warta, Daniel Barber, and Xingyi Jin. Recognition of the Distance between Plant and Stephen M Fiore. Enabling robotic social intelligence by Human by Plant Bioelectric Potential. APIEMS, pages engineering human social-cognitive mechanisms. Cogn. 602–606, 2014. Syst. Res., 43:190–207, 2017. Antonis Kakas. Cognitive Systems : Argument and Cogni- tion. pages 1–7. Antonio Lieto and Marco Cruciani. Introduction to cognitive artificial systems. 0091(July):1–3, 2017. Antonio Lieto and Daniele P Radicioni. From Human to Ar- tificial Cognition and Back : New Perspectives on Cogni- tively Inspired AI Systems. 2016. Antonio Lieto, Christian Lebiere, and Alessandro Oltramari. The knowledge level in cognitive architectures : Current limitations and possible developments. Cogn. Syst. Res., pages 1–17, 2017. Fernando Martnez, Cesar Ferri, Jose Hernandez, and M.J Ramirez. A computational analysis of general intelligence tests for evaluating cognitive development. 43:100–118, 2017. Loizos Michael, Antonis Kakas, Rob Miller, and Gyorgy Tur. Cognitive Programming. Hidetaka Nambo and Haruhiko Kimura. Estimation of Res- ident s Location in Indoor Environment Using Bioelec- tric Potential of Living Plants. Sensor and Materials, 28(4):369–378, 2016. Hidetaka Nambo and Haruhiko Kimura. Development of the Estimation Method of Resident s Location using Bioelec- tric Potential of Living Plants and Knowledge of Indoor Bookshelf. In The Tenth International Conference on Man- agement Science and Engineering Management, 2017. Hidetaka Nambo. A Study on the Estimation Method of the Resident s Location using the Plant Bioelectric Potential. APIEMS, pages 1896–1900, 2015.