Touch switch sensor for cognitive body sensor networks

Journal Pre-proof Touch switch sensor for cognitive body sensor networks Yujie Li, Huimin Lu, Hyoungseop Kim, Seiichi Serikawa

PII: DOI: Reference:

S0140-3664(19)30567-5 https://doi.org/10.1016/j.comcom.2019.07.019 COMCOM 5910

To appear in:

Computer Communications

Please cite this article as: Y. Li, H. Lu, H. Kim et al., Touch switch sensor for cognitive body sensor networks, Computer Communications (2019), doi: https://doi.org/10.1016/j.comcom.2019.07.019. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier B.V.

Journal Pre-proof

Touch Switch Sensor for Cognitive Body Sensor Networks Yujie Li, Huimin Lu*, Hyoungseop Kim, Seiichi Serikawa

of

Fukuoka University, Japan Kyushu Institute of Technology, Japan

pro

Corresponding author: [email protected] Abstract

With the global popularity of Internet of Things (IoT) technology, increasingly numbers of digital mobile products have been developed, and they have increased the productivity of people’s daily lives. These electronic products are used in all aspects of life, such as medical care, office life, home services, and sports. However, most of these products are designed for healthy people with high literacy rates.

re-

For disabled people, these products cannot be widely used. In this paper, new, differently shaped touch sensors are proposed for body sensor network-based devices. This touch sensor can be formed into any shape because of the use of conductive fabric adhesive tape as a switch. That property is why the sensor can change positions in the body sensor network in which the human body is used as a trigger

urn al P

to safely activate the touch switch. The number of switch sensors can easily be increased or decreased without changing the wiring of the central controller. The number of sensors in a switch sensor system is greater than that in other touch switch systems, and the accuracy is higher. Keywords

Touch switch sensor, Body sensor network, Data fusion 1. Introduction

In recent years, with the rapid development of information technology, body sensor networks (BSNs) has been widely applied in the fields of health-care, smart cities, security, and sports [1-5]. New information technologies, especially the Internet of Things (IoTs), are growing in importance around the world. These IoTs, especially the BSN applications, may have special benefits for social life.

Embedded systems in BSNs must have the features of small size, good performance, low power

Jo

consumption, and high reliability. Currently, a variety of single device monitor-based commercial BSN systems have been developed [6-8]. These systems primarily use signal processing techniques and machine learning algorithms. However, most of these systems ignore the multiple devices that can monitor the various body statuses. By fusing the multi-sensor signal data, the BSN systems can make better decisions. In addition, conventional equipment is easily affected by the environment. A much cheaper and energy-conserving touch switch sensor needs to be developed for BSN systems in order

Journal Pre-proof

to meet future needs. A touch switch sensor is the most common electronic device for BSN. Traditional touch sensors have two mechanical contacts. The contacts have to be connected for an electric current to flow between them and form a loop, which means that the switch is “closed”. Most mechanical switches

of

will produce a noise; furthermore, they are limited by low reliability, high failure rates and easy wear, which mean that the sensors cannot meet the technological needs and application requirements. Given the development of electronic equipment, traditional mechanical switches are unable to meet

pro

the requirements of some kinds of the modern equipment for disabled people. Smart touch sensors have been gradually developed to replace mechanical switch sensors. There are many kinds of intelligent switch sensors, such as voice controlled, light sensing, touch sensing, and pressure detection. These touch sensors have more functions than normal mechanical switches but still have some drawbacks. Voice-activated switches can be activated automatically in the case of a vocal command. However, the accuracy of voice recognition is insufficient, and it may function incorrectly. If the user

re-

needs to add another switch in that kind of system, they need to add another electric wire from the control circuit to the switch. In addition, with the increased functionality of the device, it requires more buttons to control the different functions of equipment, such as adjusting a lamp. Touch sensors are a new type of switch that can be manually controlled by tapping them. Touch

urn al P

sensors are designed using a touch sensitive chip, and the accuracy is higher than for normal smart switches. When a finger touches the switch, the human body acts as a conductor. There are several types of touch sensors, including capacitance touch sensors, resistance touch sensors and piezo touch sensors [9]. These sensors need the human body to work as a part of the circuit for the sensor to function. Another kind of touch sensor uses infrared transmission and receiving circuits to identify the touch operation [10]. A touch switch can be embedded in walls and have different shapes, but it should be on the same plane.

To have more applications for irregular planes, a bendable touch switch has been proposed [11]. This touch sensor can be placed at any place and designed into any shape, including sticking to petals [12]. The number of switches can also be increased or decreased. However, the maximum number of touch sensors is limited by the accuracy and noise of the system. In a smart home system, the touch sensors are more numerous than this design.

To improve the systems, a new kind of touch sensor is proposed in this paper. In the proposed system,

Jo

each switch uses a metal film as the touch part. The switch uses a kind of conductive fabric adhesive tape such that the shape can be flexibly changed. That property is why the sensor can change positions. In this study, the human body is used as a trigger to activate the touch switch, and it is highly safe. The number of sensors can easily be increased or decreased without changing the wiring of the central controller. The number of sensors in the system can be more than that in a conventional touch switch system, and the accuracy is also higher. The proposed touch sensor can be applied in smart home

Journal Pre-proof

devices, remote medical assistance, and Internet of Things. 2. Related Works Fig. 1 shows the block diagram of a touch panel switch using the conventional method [11]. Part A in

of

Fig. 1 shows the detection circuit, and Part B shows the switch part. Part A and Part B are connected by only two lines. If a human touches the metal film of a touch panel switch in Part B, the currents of i1 and i2 flow from detection circuit A. Fig. 2 shows the equivalent electric circuit. On the backside of the metal film

pro

of each touch panel switch, a resistance R is connected. When a finger touches the metal film, the equivalent electric circuit is shown in Fig. 3. The equivalent electric circuit of the human body is in the dashed circle in Fig. 3.

The detection method is calculated according to the ratio of i1 and i2.

A

Metal film

i2 i1

Detection Circuit

i1  i2 (1) i1  i2

re-

Rate 

B

urn al P

Touch Panel Switches

Fig. 1 Block diagram of a touch panel switch in the conventional model A

V

R

i1

V

n R i2 R R

V’

Touch points n

i1

m

R

R R

R

Rx Cx

B

R R

R R

R

R

V

i2

Touch panel switches m

Metal film

Fig.3.4 Equivalent electric circuit of Fig.3.3 Fig. 2 Equivalent electric circuit of Fig.1

Assuming that the resistance from the touch point in the left side of Fig. 2 is n, and the resistance on

Jo

the right side of Fig. 2 is m, then the expression of i1 and i2 holds.

i1  i2 

V V  nR V V  mR

(2) (3)

Based on Equations (2), (3), the ratio of the currents is shown in the following equation.

Journal Pre-proof

Rate 

i1  i2 m  n (4)  i1  i2 m  n

R

R

i1

R

R

R

R

of

R

i2

Rx

pro

Cx

Fig.3 Equivalent electric circuit when a finger touches a panel

3. System Design

In this paper, the system is broadly built using discrete components first [13]. The model is based

re-

on the switch system shown in Fig. 4. The detailed circuit of the switch part is shown in Fig. 5. All of the connected resistances in the switch part have same value 1 kΩ. In this instance, the clock signal is a PWM wave with a frequency of 10 kHz, which is produced by a PSoC chip. Vcc is 5 V, and the GNDs are produced by a PSoC MiniProg. The switch number detection has two steps. First, we detect the

R C

urn al P

V R

-

i1

R

+

V R

R

R C

-

i2

+

Sub circuit Metal film CLK Vcc GND

Detection circuit

Fig.4 Touch switch system model

Jo

current, calculate the results and judge the position. The first part is realized on board, and the second part is realized by programming the microcomputer chip. The IV converter circuit and the voltage

Journal Pre-proof

CLK

Vcc R4 + R2 Touch panel

+

D

_

_ R3

C

of

R1

R5

pro

GND

R

i1 i2

Fig.5 Model of switch part

re-

detection are based on the model shown in Fig. 6. This system is built using discrete components, and only a small number of switches are connected and tested in this experiment. Fig. 7 shows the on board circuit diagram. Each switch has the same subcircuit part; therefore, several switches are made on board. An illustration of the detection part is given in Fig. 8. The on

V+

urn al P

board LED lights are analog peripherals.

+

V-

R

-

-

+

R

+ -

R

R

Jo

Fig.6 Differential voltage detection circuit model.

Vo

Journal Pre-proof

re-

pro

of

Fig. 7 Switch part of the proposed panel switch sensor.

Fig. 8 Detection part of the proposed panel switch sensor.

The bright or dark lights represent that the devices are on or off. Different combinations of lights

urn al P

represent different devices. The amplifier is an LM358, and the PSoC chip is a CY8C29466-24PXI in this experiment. After the testing and adjustment, the system is determined to function successfully. 3.1 Different Design of Sensors

Fig. 9 shows that the switch part and the touch panel is a quadrilateral. In principle, we should verify that the shape of switch can be freely determined. In this case, we just use one switch as an example. If this switch can be successfully used, then the other switches can also be successfully used. The metal film tape is cut into different shapes, similar to the picture shown in Fig. 9. There are pentagonal, heart-shaped, circular and triangular shapes. If we can successfully test these switches one by one on the board, then the circuit worked. In other words, the shape of the switch has no influence

Jo

on the switches’ performance.

pro

of

Journal Pre-proof

3.2 Different Number of Touch Sensors

re-

Fig. 9 Picture of the different shapes of the switch panels

After testing the shapes, more switch parts are connected to the switch system, as shown in the circuit in Fig. 4. Since only one or two switches make it difficult to observe the linearity, three and

urn al P

more than three switches are tested in this experiment. To show the result, an LCD screen is used to show the values that are calculated by the microcomputer. To test the addition or subtraction of a switch, a push button is used to change the working state. Theoretically, the detection principle is represented by Equation (5). However, the actual detection is represented by Equation (6). After getting the values from the LCD, the two values are compared.

Rate 

Rate 

2n  N (5) n2

V2  V1 V2  V1

(6)

The program that is used to detect the voltage and judge the switch positions is finished in the computer and then downloaded to the PSoC chip. A Digital Dual A/D converter model is used to model the PSoC, PGA, PWM, and LCD models. The block diagram of the main program is shown in Fig. 10, and the interrupt program is shown in Fig. 11. In the main program, the A/D converter is a dual

Jo

A/D converter, which can realize two A/D converters at the same time. At the start of programming, it is necessary to set a total number and a start number of switches. We use ten measurements and then take the average, which can reduce the noise and improve the accuracy. After calculating the Rate using the voltage value, we compare it with the ideal Rate that is calculated using the total number of switches and the positions; thus, the switch ID can be identified.

Journal Pre-proof

Main

pro

Setting theoretical Rate

of

Initialization

Start PWM, LCD, PGA, and ADC

re-

Open interrupt enable signal

ADC finish?

No

urn al P

Yes

Get the data and transfer No

10 times? Yes

Average the data and calculate

Jo

Compare ideal Rate and judge

Fig.10 Block diagram of main programming.

If pushing the button means adding a switch in system, the ideal Rate should be recalculated. We

judge the current number of switches. If it is less than the prescribed maximum number, then we return to main program. If not, then we show the result on the LCD. Similarly, if we are decreasing the number of switches, the steps are the same as with adding switches.

Journal Pre-proof

Interrupt

No

Maximum?

pro

Yes Add a switch and change ideal Rate

of

No

Button

Yes

Show result

Continue

re-

Return

Fig.11 Block diagram of interrupt part. 3.3 Sensor Data Fusion

In this paper, we take Mbed NXP LPC1768 MCU for fusing the sensor data. The Mbed

urn al P

microcontroller is process at the edge layer of the human body and send the signals to the display by WiFi. Mbed microcontroller is a development board and prototyping platform based on ARM. It is with low cost and fast and includes three parts, free software library (SDK), hardware reference designs (HDK) and online tools. The Mbed Microcontroller is made for prototyping all sorts of devices; include USB, Ethernet, SD and Wifi communication. Figure 8 is the terminal layout of Mbed NXP LPC1768.

We use a 32-bit ARM Cortex-M3 core with 96MHz Mbed microcontroller that includes 512KB Flash memory, 32KB RAM and a lot of interfaces including built-in Ethernet, USB host and devices, and other interfaces. The pins (pin 5 to pin 30) can be used as DigitalIn and DigitalOut interfaces. The Mbed is powered using a 5V battery, which also connect with the touch screen. If press the touch bottom or screen, the Mbed will be work. Pin 15-Pin 18 are set as the signal input pins. Mbed based touch screen is much cheaper than the traditional one. Figure 12 shows the sensor data fusion

Jo

architecture in the proposed system.

pro

of

Journal Pre-proof

re-

Fig. 12 Sensor data fusion in edge layer.

4. Experimental Results

From the experiments, we obtain some data on the rate for when the switch is closed, as shown in Equation (5). In any case, in reality, errors cannot always be avoided. Therefore, in order to reduce the errors, we calculate a range for the errors and then test the validity of the data. The I/V converter in

urn al P

Fig. 4 produces the voltages V1 and V2. If these two voltages include the errors V1 and V2 respectively, according to Equation (6), the maximum error rate R can be calculated as follows.   Rate V1  Rate V2 V2 V1 (7) 2V2 2V1  V1  V2 V2  V1 V2  V1

R 

We measured V1 and V2 using an analog oscilloscope, and V1 and V2 are obtained from the LCD. The data are shown in Tables 1-4 when the total number of switches is 3, 4, 5, and 6, respectively. Table 1 Three switches in this system

V1/[V]

V2/[V]

V1 /[V]

V2 /[V]

1

4.25

1.16

0.0210

0.0140

2

2.30

1.58

0.0138

0.0160

3

1.56

2.32

0.0150

0.0140

Jo

Switch ID

Journal Pre-proof

Table 2 Four switches in this system V1/[V]

V2/[V]

V1 /[V]

V2 /[V]

1

4.23

0.92

0.021

0.015

2

2.35

1.15

0.016

0.0125

3

1.56

1.55

0.0155

4

1.19

2.30

0.0130

of

Switch ID

0.0165

pro

0.0170

Table 3 Five switches in this system V1/[V]

V2/[V]

1

4.30

0.78

2

2.35

0.94

3

1.56

1.17

4

1.17

5

0.94

V1 /[V]

V2 /[V]

0.0220

0.0200

0.0160

0.0175

0.0115

0.0135

re-

Switch ID

1.55

0.0135

0.0140

2.32

0.0170

0.0160

urn al P

Table 4 Six switches in this system Switch ID

V1/[V]

V2/[V]

V1 /[V]

V2 /[V]

1

4.65

0.71

0.0180

0.0120

2

2.19

0.81

0.0120

0.0160

3

1.60

0.96

0.0160

0.0110

4

1.20

1.23

0.0105

0.0150

5

0.95

1.6

0.0110

0.0140

6

0.80

2.36

0.0120

0.0180

Tables 5-8 show the results for the rate, the ideal rate and the error rate in different cases. The ideal rate is calculated based on Equation (5). The rate is obtained using Equation (6). The error

Jo

rate is calculated using Equation (7). Correspondingly, Fig. 13 to Fig. 16 show the rate and switch ID in each graph, respectively.

Journal Pre-proof

Table 5 Three switches in this system Rate

Ideal Rate

△R

1

-0.605

-0.6

0.0057

2

-0.195

-0.2

0.0077

3

0.206

0.2

0.0075

pro

0.3 0.2 0.1 0

Rate

-0.2 -0.3

urn al P

-0.4

2

3

re-

1

-0.1

of

Switch ID

Switch ID.

-0.5 -0.6 -0.7

Fig.13 Switch number is 3

Table 6 Four switches in this system Rate

Ideal Rate

△R

1

-0.679

-0.667

0.0062

2

-0.329

-0.33

0.0078

3

0.004

0.0

0.0102

4

0.342

0.33

0.0082

Jo

Switch ID

Journal Pre-proof

0.6

0.2

0 2

3

-0.2

-0.4

re-

-0.6

4 Switch ID

pro

Rate

1

of

0.4

-0.8

urn al P

Fig.14 Switch number is 4

Table 7 Five switches in this system Rate

Ideal Rate

△R

1

-0.717

-0.714

0.0080

2

-0.427

-0.429

0.0103

3

-0.136

-0.143

0.0093

4

0.15

0.143

0.0101

5

0.435

0.428

0.0102

Jo

Switch ID

Journal Pre-proof

0.6 0.4

of

0.2

1

2

3

4

5 Switch ID

pro

Rate

0 -0.2 -0.4

re-

-0.6 -0.8

Fig.15 Switch number is 5

urn al P

Table 8 Six switches in this system Rate

Ideal Rate

△R

1

-0.752

-0.75

0.0047

2

-0.493

-0.5

0.0099

3

-0.232

-0.25

0.0101

4

0.007

0

0.0104

5

0.27

0.25

0.0094

6

0.508

0.5

0.0085

Jo

Switch ID

Journal Pre-proof

0.6 0.4

of

0 1

2

3

4

5

6

pro

Rate

0.2

-0.2

Switch ID.

-0.4

re-

-0.6 -0.8 -1

urn al P

Fig.16 Switch number is 6

In these figures, the black short line in each data point refers to the biggest error rate, which shows the maximum error of each point. Compared with the rate, the error is very small. As observed from the value in tables, the greatest error among all the points is 0.0104. In each figure, to assess the linearity between the ID and the Rate, we calculate the correlation coefficient . This coefficient is calculated using the following equation.



 ID  ID Rate  Rate   ID  ID   Rate  Rate  i

(8)

i

2

2

i

i

In these figures, we can get the correlation coefficients in each case, and the approximate values are shown in Table 9, where N represents the total number of switches in the system. Table 9. Correlation coefficients in different cases 3

4



0.9999

1

Jo

N

5

6

1

0.9998

A correlation coefficient is a quantitative measure of the correlation and dependence (the statistical

Journal Pre-proof

relationships) between two or more random variables or observed data values. Any of the correlation coefficient measures that assess how well a statistical model fits the observations by summarizing the discrepancy between the observed values and the expected values from the model has to be between –1.0 and 1.0. Correlations equal to 1 correspond to data points lying exactly on a line.

of

From this result, we can determine that the rate and switch ID have a linear relationship. The greatest error in the rate of each point is 0.0104, and the range of the rate is from -1 to 1. Therefore, we can estimate the maximum number of switches. It is roughly estimated as (range of rate)/(maximum

pro

error). Therefore, the value is 1   1 / 0.0104  200 . Thus, we can connect 200 touch switches in this system, which can meet the needs of BSN systems. This number is considerably greater than that obtained with the conventional method and is sufficient for practical applications.

5. Conclusions

In this paper, we proposed a new kind of touch switch system and its structure. The detailed

re-

circuits are analyzed in other sections. This study’s contributions are as follows: (1) an energyconserving touch switch sensor for BSN Systems is developed and experimentally tested, (2) a flexible and arbitrarily shaped touch switch is designed, and (3) it has low costs with the best battery life. Our touch switch system has high accuracy, since the human body is used as a trigger to activate the switch

urn al P

circuit. After verifying the feasibility of the touch switch through simulations, a real circuit is built on the board. Differently shaped touch panels are used in this experiment, and the results show that the performance of the switch has nothing to do with the shape of the switch. In the experiments, several tests are carried out with 3, 4, 5 and 6 switches. After detecting the output voltage, the current rate can be calculated using the proposed detection method. All of these results show that the rate has a linear relationship with the switch ID when the total number of switches is certain. This finding means that adding switches is possible. Since circuit errors are inevitable, voltage errors are also detected when the switch works. The maximum error of the rate is used to estimate the maximum number of switches that can be connected to the system. As the number of switches increases, the accuracy is only slightly influenced, which means that the number of switches can be much more than the conventional method.

For future work, the power consumption of switches needs to be reduced. In this switch model, we focus on improving the accuracy of the touch switches and test the switches using discrete

Jo

components when the number of switches is less than 6. In the future, we should test more switches and make an integrated circuit with more concise switches.

References

[1] Y Zhang, H Wen, F Qiu, Z Wang, H Abbas, “iBike: Intelligent public bicycle services assisted by

Journal Pre-proof

data analytics”, Future Generation Computer Systems 95, 187-197, 2019. [2] Y Zhang, X Ma, J Zhang, MS Hossain, G Muhammad, SU Amin, “Edge Intelligence in the Cognitive Internet of Things: Improving Sensitivity and Interactivity”, IEEE Network, 33(3), 58-64, 2019.

of

[3] X Ma, J Zeng, L Peng, G Fortino, Y Zhang, “Modeling multi-aspects within one opinionated sentence simultaneously for aspect-level sentiment analysis”, Future Generation Computer Systems 93, 304-311, 2019.

pro

[4] Y Li, Y Zhang, X Xu, L He, S Serikawa, H Kim, “Dust removal from high turbid underwater images using convolutional neural networks”, Optics & Laser Technology 110, 2-6, 2019. [5] Y Zhang, Y Qian, D Wu, MS Hossain, A Ghoneim, M Chen, “Emotion-aware multimedia systems security”, IEEE Transactions on Multimedia 21 (3), 617-624, 2018.

[6] Y Zhang, R Gravina, H Lu, M Villari, G Fortino, “PEA: Parallel electrocardiogram-based authentication for smart healthcare systems”, Journal of Network and Computer Applications 117, 10-

re-

16, 2018.

[7] S Xiao, H Yu, Y Wu, Z Peng, Y Zhang, “Self-evolving trading strategy integrating internet of things and big data”, IEEE Internet of Things Journal 5 (4), 2518-2525, 2018. [8] H Lu, Y Zhang, Y Li, Q Zhou, R Tadoh, T Uemura, H Kim, S Serikawa, “Depth map reconstruction 7122, 2017.

urn al P

for underwater Kinect camera using inpainting and local image mode filtering”, IEEE Access 5, 7115[9] P.T.Krein and R.D.Meadows, “The electro-quasistatics of the capacitive touch panel",IEEE Transaction on Industry Applications,Vol.26,No.3,pp.529-534,1990. [10] C. Jiang, L. Liu, H. Kiang and S. Lu, “The control system of the Blind Hole Processing Machine based on PLC and touch panel”, Proc.of 2010 International Conference on Computer, Mechatronics, Control and Electronic Engineering, Vol.4, pp.556-558, 2010. [11] S. Serikawa, L. Zhang, “Proposal of a touch panel switch with the function of bending and addition”, 2nd International Conference on ICICI-BME, pp. 8-13, 2011 [12] Miyata K., Shiyuan Yang, Serikawa S., Nakashima S., “Proposal of Flexible Touch Panel Sensor”, SNPD, 2013 14th ACIS International Conference on, pp 615-620 [13] Yong Huang, “Simulation analysis and design of the circuit based on PSPICE”, Journal of Hubei

Jo

Institute of Education, Vol. 24, No. 2, pp. 35-37, 2007.

Journal Pre-proof

Conflict of Interest The authors whose names are listed immediately below certify that they have NO affiliations with or involvement in any organization or entity with any financial interest

of

(such as honoraria; educational grants; participation in speakers’ bureaus membership, employment, consultancies, stock ownership, or other equity interest; and expert testimony or patent-licensing arrangements), or non-financial interest (such as personal

Jo

urn al P

re-

materials discussed in this manuscript.

pro

or professional relationships, affiliations, knowledge or beliefs) in the subject matter or