Energy Optimization in Directional Advanced Intruder Handling AODV Protocol in MANET

Transcription

1 Energy Optimization in Directional Advanced Intruder Handling AODV Protocol in MANET S.Hemalatha 1, P.C.Senthil Mahesh 2 1,2 Professor,Panimalar Institute of Technology Chennai, Tamil Nadu 2 Professor,Annamachariyar Instittue of Technology,Andra Pradesh, 1 slaechemalatha@gmail.com, 2 senthilmagesh@gmail.com Correspondent author P.C.Senthil Mahesh Abstract Mobile Ad-hoc network is a collection of nodes which tries to communicate each other without any fixed infrastructure. In this network, nodes can move freely and dynamically from self-organized into arbitrary topologies. Due to self-organizing, the network is vulnerable to attack by an intruder who attempts to gain unauthorized access and damage data on communication medium. Transmitting of packet from source to destination is one of the greatest challenges because the packet should reach the destination without disturbances like delay, packet loss and intruder etc. Adhoc On- Demand Distance vector protocol is designed for transmitting of packet by finding a new route when it s needed. Even though this protocol is creating a path on demand, protocol functionalities has limitations on route redirection, security and energy consumption. This research article is worked to develop algorithm to identify the failure and Black hole attacker nodes in the network. The algorithm uses directional antenna transmission to optimize the energy consumption as energy factor is an important challenges in MANET. The developed protocol is named as Directional Advanced Intruder Handling Ad-hoc On-Demand Distance Vector protocol (Directional Advanced Intruder Handling Ad-hoc On-Demand Distance Vector). The algorithm are simulated in NS2 and compared with Ad-hoc On-Demand Distance Vector protocol. The result of analysis shown, proposed research algorithms performed 50% better than AODV protocol. Keywords: MANET, Black hole attacker, Failure node, Divide and Conquer, Directional Antenna. 1. Introduction A Collection of nodes formed a network under the working principles of move freely, organized themselves arbitrarily and without any administration is called Mobile Adhoc network (Wikipedia 2004). In a common, a route between the source to destination 935

2 1 through the Ad-hoc network is established by the routing protocol. The packets have followed this route to transfer the data. Packets are moved from a node to another node called the hop, until to reach the destination. Mobile Ad Hoc Network (MANET) is assortment of Multi-hop wireless movable nodes that correspond with each other without central control or recognized infrastructure. Nodes in this network can move freely, so this network is more prone to error. The MANET architecture with three nodes communication as shown in figure 1. As a result, routing in MANET is a Critical task due to highly dynamic of mobile environment. In recent years, numerous routing protocols have been proposed for mobile ad hoc Networks and outstanding among them are DSR, AODV and TORA. Challenges in MANET Fig.1 MANET A Mobile Ad Hoc Network (MANET) is a gathering of wireless mobile nodes form a momentary network without any fixed infrastructure. All the node are configured among them to move freely.every node is capable of performing dual role as a router and host. This natural history of MANET leads to come across the challenges are 1) Dynamic network topology 2) Frequency of updates or Network overhead 3) Energy efficient/power aware routing 4) Secure routing This research article is developed to provide solution for dynamic topological challenges due to failure, identify black hole attacker node and directional antenna transmission used for energy optimization. 2. Literature Survey 936

3 2 A Routing protocol in an Ad-hoc network (Wikipedia, 2004) is divided into two main categories of proactive and reactive protocol. In proactive protocol nodes maintain routing information for all other nodes in the network and it is stored in routing table. So this protocol is also named as a table driven protocol. In reactive protocol, route information is established when a packet is transferred between the nodes. In the table driven protocol is classified into different types Destination Sequenced Distance Vector Routing (DSDV), Cluster Gateway switch routing protocol (CGSR), Optimization link state routing protocol, Topology dissemination Based on reverse path (TBRPF), Fish Eye state Routing Protocol (FSR).In source initiated routing protocols are classified into different types of protocols such as Ad-hoc On-Demand Distance Vector (AODV) (Perkins, & Royer 1999), Dynamic source routing protocol (DSR). In Dynamic Source Routing Protocol each node maintains a route cache contains a route learned by the node. Source node only initiates route discovery process enters into a route cache continuously updated. AODV node creates a route on demand to maintain a complete a route using DSDV algorithm. TORA is another source initiated on Demand protocol, in a concept of link reversal of directed Acyclic Graph. TORA has the capacity of routing repair. ABR routing protocol (Giannoulis et al 2007) is on demand protocol route selection is based on the signal strength in the link. Intruder detection is a one of the challenges in MANET ( Tiranuch Anantvalee & Jie Wu 2006; Ioanna Stamouli, et al; Dorothy Denning, 1987 Yongguang Zhang & Wenke Lee 2000). Different methodologies were proposed for identifying an intruder in MANET from the year 2001to 2003, these methodologies were based on the technique of Knowledge-based intrusion detection (Farooq Anjum et al 2003) signature based intruder detection, sensor based intruder detection (Kachirski & Guha 2003), anomaly based intruder detection (Md. Safiqul Islam & Syed Ashiqur Rahman 2011) collaborative intruder detection (Ningrinla Marchang & Raja Datta 2008)and zone based intruder detections Sun, B, Wu, K & Pooch, U 2003.Identification of an intruder was done by defining architecture in MANET during the year 2005 to 2007, based on corporative based intruder detection architecture and RIDAN architecture was developed. Different types of MANET attacks were identified using attacks detection techniques (Ranjana & Rajaram 2007), warm hole attacks, critical node identification (Karygiannis et al 2006; Rajaram & Palaniswami 2010), fabrication attacks (Ranjana & Rajaram, 2007), consumption attacks, packet dropping attacks, black whole attacks were detected based on attack detection techniques. the Gain(gd) of the any antenna can be defined as the product between the directivity and efficiency. g d = ƞg d When gt transmission gain of the directional antenna Gr is a reception gain And go is a omni-directional gain. From the property of directional antennas, gt g0 937

4 3 gr g0 gain is expressed in terms of decibels (db). If g is any given gains and the equivalent gain in decibels(g db) is given by: g db = 10 log 10 g From the above literature survey, the routing protocol only performs routing the packets, none of the protocols have been proposed for identifying a balck hole in the Ad hoc network. There is need for an algorithm for efficient delivery of packet to the destination. In this research AODV protocol is modified for efficient packet delivery. Performances of AODV modified algorithms are compared with AODV and simulation results depict that proposed algorithms out performs existing AODV. The research article is organized into five sections. Section 2 discusses a brief literature survey. Section 3 provides an overview about research methodologies. The design and results discussion are explained in section 4 and, 5. Finally Section 6 draws conclusions 3. Research Methodology A novel protocol is designed and implemented that incorporates identifying failure node and black hole attacker with directional antenna transmission. This algorithm is implemented by modifying existing AODV protocol as AODV lacks ability to identify failure nodes and black hole attacker nodes. The algorithm is further improved for power optimization using directional antenna transmission. Methodology Used 3.1 Divide and Conquer Divide and conquer is a recursive algorithm and it is used to give solution to a problem by working through divide the problem in to two or more sub problem. The final solution of sub problem gives solution to the original problem give solution. There are five main steps for a divide and conquer solution Step 1: Define your recursive sub-problem. Step 2: Define your base cases. Step 3: Present your recursive cases. Common Running Times Let T (n) denote the running time of your algorithm on input of size n. If T (n) = 2T (n /2) + O(n) and T(1) O(1) then T(n) O(n log n). 938

5 4 In this thesis is defined to used Divide and Conquer strategy on the packet acknowledgement for finding out Failure node and Intruder node in MANET. So the above five major steps are defined as Step 1: Recursive sub problem Calculate the number of hops between the needed nodes. The number hop between the nodes is divided by two. Finding out the middle node Step 2: Define base cases: Case 1: Check whether the acknowledgement received from the middle node If (Yes) Conclude Up to the beginning to middle node packet is forwarded. Else if (Process middle node is a failure node) Else ( Find out the new middle and repeat the process ) Step 3: Present your recursive cases Case 2: If middle node received Acknowledgment Take the second half Check the acknowledgment received from middle node go to Step 1. Case 3: If consecutive more than one nodes acknowledgement is not received for example i,j,k etc, Check whether the previous node of first consecutive node forwarded the packet, then conform failure node is first i node by checking following stages. (i) (ii) (iii) (iv) Send router request to the acknowledgement missing nodes. Select the failure node which router reply is not received. If more than one Router reply missing go to (i) More than two chances router reply missing conform all the missing node as a failure node. 3.2 Algorithm for identifying failure node Procedure(Source, Dest, G) - Divide and Conquer strategy Consider the ordered Set G={1..N}. Step 1: Initialize source = 1, dest = N. Step 2: Calculate middle = No of hops (source to dest)/2. Step 3: (i) Check whether the packet is passed through the middle node. If (yes) Calculate the new middle node from the middle node to the destination then go to step 2. Else 939

6 5 Calculate the middle node from source node to middle node then go to step 2. Repeat the process. If there is no flow of data then the node may be Victim or failure node. Process whether the middle node is victim node If True Set Victim = Middle and initiate route discovery process. Step 4: Process to confirm Victim node Send route request to the Victim node. If there is no reply then confirm Victim node. Step 5: Process to retransmit the data through stage 1. Step 6: Send alert message about the Victim node or failure node. Step 7: Stop. 3.3 Algorithm for identifying black hole attacker node Algorithm is designed for identify an black hole attacker in the MANET by adding two more stages in algorithm of failure node identification Procedure (Source, Dest, G) Consider the ordered Set G={1..N} Step 1: Initialize source = 1, dest = N. Step 2: Calculate middle = No of hops (source to dest)/2. Step 3: Check whether the packet is passed through the middle node If (yes) Calculate the new middle node from the middle node to the destination then goto step 2 Else Calculate the middle node from source node to middle node then go to step 2 Repeat the process If there is no flow of data then the node may be the black hole attacker or failure node. Process whether the middle node is black hole attacker. If True Set attacker = Middle and initiate route discovery process. Step 4: Process to confirm node is a black hole attacker. Send route request to the black hole attacker node. If there is a reply then confirm as an attacker node Otherwise node may be the failure node. Step 5: Process to retransmit the data through stage 1. Step 6: Send alert message about the Intruder node. Step 7: Stop. 3.4 Probability of failure node and attacker node When a packet is transferred from source node to a destination node, the probability of exhasive event is whether to get the failed node or link failed node. The probability of failed node = Number of failed node / total number of nodes 940

7 6 If m is a number of failed node (absolutely failed node is 1) N is a total number of nodes then The probability of failed node f= m/n. f= 1/N (eq 1) The probability of failed node is denoted by f and the probability of not failed node is denoted by a. Then (Total number of nodes) - (number of Failed nodes) = number of cases the event will not happen to failure. (N) (1) = (N-m) Here m is 1 So N-1 = N-1. The probability of not failed node = number of node which can not failed / total number of node a = (N-m)/ N a = (N/N)- (m/n) (m is 1) a = 1-(1/N) from eq 1 a = 1-f a+f = 1 hence 0 f 1 and 0 a 1 When a packet is transferred from source node to a destination node, the probability of exhasive event is whether to get the black hole attacker. The probability of black hole attacker = Number of black hole attacker node / total number of nodes 941

8 7 If b is a number of black hole attacker (absolutely failed node is 1 because one black hole attacker is possible to find ) N is a total number of nodes then The probability of black hole attacker node bf= b/n. bf= 1/N (eq 2) The probability of black hole attacker node is denoted by bf and the probability of not black hole attacker node is denoted by nb. Then (Total number of nodes) - (number of black hole attacker nodes) = number of cases the event will not happen to black hole attacker. (N) (1) = (N-b) Here m is 1 So N-1 = N-1. The probability of not black hole attacker node = number of node which can not black hole attacker / total number of node nb = (N-m)/ N nb = (N/N)- (m/n) (m is 1) nb = 1-(1/N) from eq 2 nb = 1-bf nb+bf = 1 hence 0 bf 1 and 0 nb

9 8 4. Directional Advanced Intruder Handling Adhoc On-Demand Distance Vector (DAIHAODV) protocol design The stages of failure node and black hole attacker identification of Directional Advanced Intruder Handling AODV (AIHAODV) protocol design stages are as follows: A. Decide the path using AODV protocol. B. Transmit Packet. C. Apply algorithm for Identify failure or attacker node (i) Suspect the black hole attacker node (ii) Confirm the black hole attacker node. D. Route redirection. E. Send alert message to other nodes about intruder. The algorithm design stages are implemented using NS2 with following simulation set up. AODV protocol is taken for adding the algorithm features. Simulation is done with directional antenna transmission and Omni antenna transmission to prove the comparison result of energy efficiency. PARAMETER OF THE SIMULATION Channel Type Wireless Channel Radio Propagation Model Two Ray Ground Antenna type Omni Antenna/ Directional Interface queue type Drop Tail /Pri Queue Maximum Packet in Queue 50 Network interface type Phy/Wireless Phy MAC type 802_11 Topographical Area 500 X 300 sq.m txpower 0.5W rxpower 0.1W idlepower 0.01W Initial energy of a Node Joules Routing protocol AODV Number of mobile nodes 10, 20,30,40,50,60,70,80 Mobility 0 or 20m/s Decide the path using AODV protocol In this stage, existing AODV protocol is used to identify route between source node to the destination node. Source node wants to transmit the packet to the destination, it send Router Request to all the nodes. Upon receiving a router Reply from black hole attacker who role is sending the a router reply to the requested node. Router Request in NS2 is shown in Figure

10 9 Fig.2 Decide the path Transmit Packet Once the path is identified between source to destination using AODV protocol, source node starts sending packet to the destination node through the identified path. Packet transmission in NS2 is shown in Figure 3. Fig.3 Packet Transmit Apply algorithm for identifying failure and black hole attacker node Suspect the node as an attacker In this stage, identify the whether any node is not forwarding the packet to the next hope is noted as a suspected node and it brought in to surveillance. All the activity 944

11 10 about that node will be noted and recorded. The alert message will not be send until the node is confirmed as a black hole attacker. Confirmation of attacker node In this stage, source node sends a RREQ to the suspected node, If there is no RREP from suspected node to source then confirm suspect node is attacker node and if there is False RREP from intruder to source then confirm node is black hole attacker. Otherwise the Node is failure node. Route Re-Direction The source node discovers new route to retransmit the packets when the node is identified a black hole attacker or failure node. Send Alert Message Once a suspect node is confirmed as a black hole attacker or failure node, the source node will send the alert message to the other node in the entire network. Sending alert message is shown below figure 4. Fig.4 Send alert message about failure node / Black hole attacker node POWER IDENTIFICATION From the above output screen the yellow circled nodes are having residual power after identification of the black hole attacker node 40. Here packet transmission is based on the Omni-Directional Antenna.Power status is shown in figure 5 and

12 11 Fig.5 Power Status Fig.6 Power Status From the above output screen the yellow circled nodes are having residual power after identification of the black hole attacker node The Red circled nodes are having no power after identification of the black hole attacker node. Here packet transmission is based on the Omni-Directional Antenna and Directional antenna. 4. Result NS2 is used to simulate and compare the results with AODV. Three performance parameters such as throughput (rate of packet delivered successfully through medium), packet delivery ratio (packet send / packet received), End to End delay (total time taken/connections in channel)are considered for comparison. The algorithm is implemented in antenna transmission of Omni and directional antenna. Transmission. in Omni antenna transmission performance measures are named as AIHADOV protocol for comparison. The simulation results depict that proposed DAIHADOV protocol better in all aspects.the result simulations are shown in figure 7, 8, and

13 12 Through Put Fig.7 Throughput Packet Delivery Ratio The ratio of the number of delivered data packet to the destination. This illustrates the level of delivered data to the destination. Number of packet receive / Number of packet send Fig.8 Packet Delivery Ratio The average time taken by data a packet to arrives in the destination. It also includes the delay caused by route discovery process and the queue in data packet transmission. Only the data packets that successfully delivered to destinations that counted. ( arrive time send time ) / Number of connections 947

14 13 Fig.9 End To End Delay The AIHAODV algorithm is further modified for optimization of congestion in traffic, delay and power consumption using directional antenna transmission. NS2 is used to simulate and compare the results with AODV and AIHAODV directional transmission and Omni antenna transmission. The performance parameters are based on energy consumption and residual energy. The simulation results depict that proposed DAIHADOV algorithm better in all aspects. The result simulations are shown in figure 7 and 8. Fig.10 Energy Consumption 948

15 14 Fig.11 Residual Energy 6. Conclusion An Adhoc network is a combination of different nodes, created for communicating each other without any infrastructure. Transmitting of packet from source to destination is one of the greatest challenges because the packet should reach the destination without disturbances like delay, packet loss and security breach. Adhoc On-Demand Distance vector protocol is designed for transmitting packet by finding a new route when it s needed. Even though this protocol is creating a path on demand, protocol functionality limits on route redirection, security and energy consumption. AAODV algorithm designed to identify failure node overcome it limitation of reliable packet delivery. Simulation results show that the algorithm performs better than exiting AODV. DAIHAODV algorithm modified AODV algorithm in safer delivery of packets when nodes are under threats by an attackers. Comparative analysis with existing algorithm shows the modified algorithm proves better.efficient energy consumption show that DAIHAODV outperforms existing AODV algorithm. The Proposed research articles modifies an existing AODV protocol and gives solution to identify failure and attacker node in MANET with directional antenna transmission. The proposed algorithms are simulated by ns2 and outperform the existing ADOV. References [1] Dorothy Denning, E 1987, An Intrusion-detection Model IEEE Transaction on Software Engineering, vol.13, no.7, pp [2] Farooq Anjum, Dhanant Subhadrabandhu & Saswati Sarkar, 2003, Signature based Intrusion Detection for Wireless Ad-Hoc Networks: A Comparative Study of Various Routing Protocols. [3] Giannoulis, S, Antonopoulos, C, Topalis, E & Koubias, S 2007, ZRP versus DSR and TORA: A comprehensive survey on ZRP performance, IEEE Transactions on Industrial Informatics vol.3, no.1, pp

16 15 [4] Ioanna Stamouli, Patroklos G. Argyroudis & Hitesh Tewari, 2005 Real-time Intrusion Detection for Ad hoc Networks Proceedings of the Sixth IEEE International Symposium on a World of Wireless Mobile and Multimedia Networks WoWMoM 05), /05. [5] Karygiannis, A, Antonakakis, E & Apostolopoulos, A 2006, Detecting Critical Nodes for MANET Intrusion Detection Systems, pp [6] Kachirski, O & Guha, R 2003, Effective Intrusion Detection using Multiple Sensors in Wireless Ad hoc Networks, In Proc. 36th Annual Hawaii Int l. Conf. on System Sciences (HICSS 03), pp [7] Ningrinla Marchang & Raja Datta 2008, Collaborative techniques for intrusion detection in mobile ad-hoc networks, Ad Hoc Networks, vol.6(4), pp [8] Perkins, CE & Royer, EM 1999, Ad-hoc On-Demand distance vector routing, in: Proceedings of 2nd IEEE Workshop on Mobile Computing Systems and Applications, pp [9] Rajaram, A & Palaniswami, S 2010, Malicious node detection system for mobile ad hoc networks. (IJCSIT) International Journal of Computer Science and Information Technologies, vol.1(2), pp [10] Ranjana, R & Rajaram, M 2007, Detecting Intrusion Attacks in Ad-hocNetworks, Asian Journal in Information Technology, vol.6(7), pp , ISSN: 1682:3915. [11] Md. Safiqul Islam & Syed Ashiqur Rahman 2011, Anomaly Intrusion Detection System in Wireless Sensor Networks: Security Threats and Existing Approaches. International Journal of Advanced Science and Technology vol. 36. [12] Sun, B, Wu, K & Pooch, U 2003, Zone-based Intrusion Detection for Mobile Ad hoc Networks [13] Tiranuch Anantvalee & Jie Wu 2006, A Survey on Intrusion Detection in Mobile Ad Hoc Networks, pp [14] Wikipedia, 2004, The free encyclopedia Mobile_ad-hoc_ network. [15] Yongguang Zhang & Wenke Lee 2000, Intrusion detection in wireless ad-hoc networks. In MOBICOM, pp

17 IoT Based Self Monitoring System For Chronic Disease Patients Golda Jeyasheeli. P., Professor, Department of CSE, Mepco Schlenk Engineering College Sivakasi, India Chandra Mohan. B., Associate Professor, School of CSE, VIT, Vellore, India Raji. A., PG Student, Department of CSE, Mepco Schlenk Engineering College Sivakasi, India Abstract Nowadays people suffered from chronic diseases like asthma, chronic obstructive pulmonary disease, diabetes, and hypoxemia. The vital signs like respiratory rate, blood glucose and oxygen saturation has to be monitored periodically for the people who are suffering from these diseases. As the health care cost is increasing, people cannot afford to go to hospital regularly, and in a few of the hospitals, the healthcare professionals and patient ratio are comparatively low. Hence, it is unable to monitor all the patients on continuous monitoring by the healthcare professionals. With the lack of immediate help, patients are suffering lot. In order to solve this issue, this paper proposed a self monitoring system for both inpatients and outpatients using Internet of Things technology. With the help of our proposed system, the doctors will continuously monitor the patient s health status and will advise the patients remotely. Thereby, the patient visits to the hospital, hence the transportation cost and the travelling burden of the patients are reduced. It also improves the quality of care and early disease detection. Keywords Respiratory rate; Blood glucose; Oxygen Saturation; health care; wearable sensor; vital signs. I. INTRODUCTION According to Global Burden of disease study, the people in India are suffering from asthma, chronic obstructive pulmonary disease, diabetes and hypoxemia. People suffered from these diseases are needed to monitor their vital signs. The vital signs are the basic functions of the body. Among the six vital signs, the respiratory rate, blood glucose and oxygen saturation are considered in this project. The respiratory rate means the number of times the chest rises and finds the breaths per minute for a patient at rest. Diabetes are diseases in which person has high levels of sugar in blood and also insulin production in the body is inadequate. Diabetes patients need long term monitoring and blood glucose level varies before meals, after meals and also in bed time. Oxygen saturation in blood refers how much oxygen is carried in a blood. Hypoxemia means the low oxygen level in blood. People suffered from chronic diseases are needed to monitor their vital signs regularly. Every time people come to the hospital and do a check up. All the people cannot afford to go to hospital regularly because of increasing health care cost. Sometimes people may miss their regular check up and also undergone stress and mental problems. And also in hospital healthcare professionals and patient ratio is comparatively low. All the patients are not monitored by healthcare professionals (Chandramohan et al, 2011, 2012, 2015). With the help of advent of science and technology, we use Internet of Things technologies to monitor the vital signs of a patient more comfortable. The Internet of Things technologies emerged from wireless sensor networks, embedded system and pervasive computing. All the medical devices are built by using embedded system technology. Transfer the data from one to another node and thus forming an adhoc network by using wireless technologies. By applying IoT techniques, each medical device has assigned same IP address. Using these IP Address, the sensor data may see from anywhere using internet connectivity. 951

18 The self monitoring system is designed with the help of IoT technology. The system is used for both inpatient and outpatient. The outpatient health status is known by doctor remotely and patients get correct solution in an emergency case. It reduces the hospitalization cost, patient need to visit the doctor only on need base and long term stay in hospital also reduced. The inpatients know the health status without healthcare professionals help (Saikiran et al, 2014; Chandramohan et al, 2011). These IoT medical devices produce lots of data and all the data are not useful to us. By using data mining techniques, analyze the data. The model is built using supervised learning (Chandramohan et al, 2008). By applying classification algorithm, the model is trained and tested to find the health status of a patient. For every vital sign, different class labels are maintained. It gives the health status. For unseen data also, the model gives the correct result. And also in hospital system, ontology is built for these vital signs. Every vital signs has causes, symptoms and medicines. Anybody can see these details from ontology by sparkle queries (Arulmozhi et al, 2011; Chandramohan et al, 2011). II. EXISTING WORK Shichiri et al (2010) developed invasive technique for glucose monitoring. In every time the finger pricking is needed to withdraw the blood sample and it is placed in a test trip to measure glucose using a glucometer. For continuous glucose monitoring the multiple finger prints are not desirable which is very painful and higher risk of affections.hanady et al (2015) has developed ECG remote monitoring system. They used NRF24l01 for wireless transmission. It is very low cost. Adhoc network architecture is formed and the sensor data is transferred from one node to another node. Marietta et al (2016) developed a respiratory rate monitoring system based on PPG technique. The IR emitter and receiver is placed in a pillow which observes the respiratory rhythm. These respiratory rate algorithms are developed on signal processing techniques and didn t give better results for all aged people and also in sleep apnea cases. These systems, respiratory rate output are affected by many external parameters. Swathi et al (2016) developed respiratory rate monitoring system using thermistor temperature sensor. One temperature sensor is used to calculate the respiratory rate. Single temperature sensor is used here. It didn t give better results. The respiratory rate is affected by many other external parameters. They didn t concentrate on those parameters. Supriya et al (2016) developed a facial tracking respiratory rate. The thermal image of a patient is captured from a thermal camera. And applying image processing techniques to find the respiratory rate. By using these methods, the accuracy is not achieved. Sonia et al (2015) developed Arduino based web server for temperature sensor using an Ethernet shield. By mounting the Ethernet shield into an Arduino, they assigned IP address for Arduino. By using the IP address, they identified their Arduino device and displayed the value in a web browser. It can be accessed in a particular network. They used red, orange, yellow, green and IR rays to find the glucose level. They found as the voltage value increases based on the glucose concentration value. IR and LED emitter is placed on one side and photodiode is placed on other side. They used two finger to find the pulse oximetry. In our paper, the better algorithm for respiratory, glucose and pulse oximetry is discovered. This algorithm perfectly works for all patients and it concentrates on external parameters also. And outpatients are remotely monitored by doctors and in hospitals patient s data is published on a web server. In high emergency case, the system generates an automatic call to the ambulance. III. PROPOSED WORK We propose a remote health monitoring system for both inpatient and outpatient. Patient who suffer from diseases wear a wearable sensors. These sensors are fitted in a things like nebulizer mask, ring and clip. These things may sense the vital signs of a patient and transmitted to medical center through wireless communication. From the webpage, the data is accessed by all doctors. By using the system in hospital nurses take care the patients well and reduce the long term stay in hospital. The doctor give these wearable sensors to the patient discharging time which may reduce the hospitalization cost of the patient and need basis patient visit the hospital. The early disease detection is also possible. From the collected sensor data, we do some analysis to find the healthy and unhealthy fields by classification algorithms. And ontology is also created to understand the causes, symptoms and medicines for the particular diseases. Fig 1 represents the self monitoring system. To implement the system in real time, we choose Arduino microcontroller. And the sensors used are room temperature LM35 sensor, IR Photo transmitter and receiver, Red LED. A. Data Acquisition Sensor is a device that measures a physical quantity and converts it into a signal which can be read by an observer. To acquire physiological data from the patient, we use body temperature, heart rate and blood pressure sensors. Arduino is an open source prototyping platform. It is easy to use hardware and software. Arduino boards are able to read input from sensors and turn it into an output. This is defined by a set of instructions programmed through Arduino IDE. It has 54 digital and 16 analog input/output pins. We can give 6-20V voltage. It has 256KB flash memory and it has 10 bit Analog to Digital converter. 952

19 1. Respiratory Sensor To sense the temperature value LM35 sensor is used. The LM35 is temperature sensor, whose output voltage is linearly proportional to the Celsius temperature. The LM35 sensor accuracy is ±1 4 C. The range of the temperature sensor is 55 to +150 C. The output from the sensor gives the analog voltage. This output is linearly proportional to the Celsius (centigrade) temperature. Every 10mV is considered as 1 C. So divide the output mv by 10 gives the temperature value in C. Fig 2 represents the smart nebulizer mask. LM35 (1) represents the nostril temperature and LM35 (2) represents the external temperature. The difference gives the respiratory signal. Figure 1. Self Monitoring System 2. Glucose Sensor Figure 2. Smart Nebulizer Mask Non-invasive blood glucose meter designed and fabricated in this study composed of a circuit consisting two LEDs of the same wavelength (LED pair) with one acting as a photo-emitter and the other as a photodiode photo-detector. The nearinfrared LED (NIR-LED) pair with wavelength of 1450nm is used. The finger is placed in between photo emitter and a photo detector. As the glucose concentration increases and the voltage value also increases. Fig 3 represents the smart ring. 953

20 Figure 3. Smart Ring 3. Pulse Oximeter Sensor Pulse Oximeter is used to monitor person s oxygen saturation. Oxygen saturation in blood represents how much oxygen in blood. Pulse Oximeter consists of RED LED, IR LED and photodetector. The RED LED and IR LED is placed on one side. The finger is placed in between LED and photodetector. The area of minimum light absorption to maximum light absorption can be attributed entirely to blood. By using normalized Ratio, intermediate value may calculate. Then SPO 2 is calculated. Fig 4 represents the smart clip. Figure 4. Smart Clip The equation (1) represents the absorption rate and spo 2 is calculated from equation (2) where A & B are constant. A= and B= B. Outpatient Monitoring 1. Transfer Sensor data via Bluetooth Once the data is sensed and then the data is transferred to mobile using Bluetooth. HC 05 Bluetooth device is used. The data is displayed in android GUI. Bluetooth serial Monitor app is used to display the sensor data. 2. Automatic call to ambulance The data is greater than threshold value means, generate an automatic call to ambulance. 3. Track the Patient Geographical Information By using android location manager to track the latitude and longitude information of a patient. 4. Transfer the sensor data and geographical data using NRF The sensor data is transmitted from the patient s home to medical center /hospital using NRF24L01. The NRF24L01 is IEE standard, low cost, the distance is >1000m and the data rate is 1Mbps. (1) (2) C. Inpatient Monitoring 1. Receive the data using NRF Once the data is sensed and then the data is transferred to mobile using Bluetooth. HC 05 Bluetooth device is used. The data is displayed in android GUI. Bluetooth serial Monitor app is used to display the sensor data. 2. Display the data in webpage 954

21 Finally the data is displayed in a webpage. By using Ethernet shield, the sensor data is stored in mysql database and also published in webserver. In a webpage, track the outpatient geographical information also. And also the patient information is put in Thingspeak cloud. 3. Data Analysis Ontology and classification model is also developed for checking the vital signs be normal or not without nurse help. Ontology is created by using protégé tool. Classification model is developed in WEKA. 4. SMS to patient s mobile Finally causes, symptoms, medicines to patient s mobile by SMS. SIM900A GSM is used. By using AT commands to send message to patient. 1. Experimental Setup IV. RESULTS AND DISCUSSION Table I shows the values of nostril temperature, external temperature and difference between the two temperatures. Table 1 Temperature Difference Time Room Nostril Diff 16:16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: :16: Fig 5 shows automatic call and gps screenhots. Fig 6 represents the respiratory Web Server. Fig 7 represents Respiratory Web page. Fig 8. Shows the sensor data in mysql. 955

22 Table 2 represents the respiratory rate for five different patients. Table 3 represents the blood glucose for five different patients before meals and after meals. Table 4 represents the oxygen saturation for five different patients. Table 2 Respiratory Rate for different patients Patients Age Respiratory Condition Rate Normal Normal Normal Abnormal Normal Table 3 Blood Glucose for different patients Patients Age Blood Glucose Level Before Meals After Meals Table 4 Oxygen Saturation for different patients Patients Age Oxygen Saturation Fig 5. Automatic call and gps screenhots 956

23 Fig 6. Respiratory Web Server Fig 7. Respiratory Web page Fig 8. Store the sensor data in mysql Fig 9. Shows the patient gps latitude and longitude information displayed in a google map in a web application. Fig 9. Track the patient location Table 5 shows the percentage of healthy and unhealthy fields for respiratory rate. Table 5 Healthy and Unhealthy fields for respiratory rate 957

24 Classification Algorithm Correctly Classified Instance Incorrectly Classified Instance Naïve Bayes 74.38% 25.62% LIBSVM 71.90% 28.10% J % 28.10% b. Performance Analysis To measure the performance of the system, we have to use a percentage error measure. It is the difference between an approximate or measured value and an exact or known value expressed as percentage. The formula for %error is given in equation 3. %error = ((Experiment value Theoritical value) / Theoritical value ) 100% (3) RMS Error is used for performance analysis of respiratory rate. In RMS error as in equation 4, n represents the number of observations for a single subject, y calib represents the respiratory rate value checked manually by doctors or any healthcare professionals and y pred correspond to the respiratory rate value attained by our method. Here for a single subject, we have taken 5 samples. The number of observations is 5. Table 6 shows the Performance Analysis of respiratory rate by applying RMSError. Patien t Table 6: Performance Analysis for Respiratory Rate Patients Age Respiratory RMS Error Rate Table 7 Performance Analysis for Blood Glucose Ag e Blood Glucose Level (Our Prototype) Blood Glucose Level (Original Device) Percentage Error BM AM BM AM BM AM Table 7 shows the performance analysis for blood glucose by applying percentage error. Table 8 shows the performance analysis for pulse oximeter by applying percentage error. Table 8 Performance Analysis for Pulse Oximeter Patients Age Oxygen Oxygen Percentage 958

25 Saturation (Our Prototype) Saturation (Medical Device) Error V. CONCLUSION AND FUTURE WORK In this paper, we have monitored the inpatient and outpatient vital signs. The collected data are stored in remote places and if any abnormal conditions happened, the system generates an alarm message and automatic call to the ambulance. And then the data are transmitted wirelessly to the medical Centre via NRF / GSM. In the medical Centre, the data are stored for further analysis and also give information about the patient status. And then the patient data are published locally using a web server and the data are seen by all the doctors in the hospital. The system is flexible to all users, the hospitalization cost is reduced and prevent the disease earlier. IoT devices produce enormous amounts of data. To handle the huge data in a technological way, we have to use Big Data Analytics. Data storage through cloud computing and remote access to stored data improves the computing performance. Acknowledgment The authors wish to thank the Management and Principal of Mepco Schlenk Engineering College, for the support in carrying out this research work. References [1] Sonia, Nafiz Ahmed Chisty, Wireless Sensor Tree Network monitoring system through Web Server, International Journal of Scientific & Engineering Research, Volume 6, Issue 1, 2015, pp.395. [2] Hanady S.Ahmed Abduladhem Abdulkareem Ali, Wireless Sensor Network for Medical Applications, Iraq J. Electrical and Electronic Engineering, Vol.11 No.1, [3] Shichiri Y, Y.Yamasaki, R.Kawamori, N.Hakui and, H. Abe (2010). Wearable Artificial Endocrine Pancreas with Needle- Type Glucose Sensor. Lancet, vol. 320, no. 8308, pp [4] Chandra Mohan, B. and Baskaran, R. A Survey: Ant Colony Optimization based recent research in various engineering domains Expert System with Application, Elsevier, Vol. 39, No. 4, pp , [5] Chandra Mohan, B. Restructured Ant Colony Optimization routing protocol for next generation network, International Journal of Computer Communication and Control, Vol.10, No.4, pp , Agora University Press, 2015 [6] Chandra Mohan, B. and Baskaran, R., Reliable transmission in network centric military network, European Journal of Scientific Research, Vol. 50, No. 4, pp , 2011 [7] Sai Kiran Chebbiyyam, and B. Chandra Mohan, Review on Secured Cloud Environment Using Cryptographic Schemes, International Journal of Applied Engineering Research, Vol 9, No 24, pp , 2014 [8] Chandra Mohan, B., Sandeep, R. and Sridharan, D. A Data Mining approach for Predicting Reliable Path for Congestion Free Routing using Self-Motivated Neural Network, Studies in Computational Intelligence (Vol. 149), Springer-verlag, pp , [9] Chandra Mohan, B. and Baskaran, R. Reliable Barrier-free Services in Next Generation Networks, International Conference on Advances in Power Electronics and Instrumentation Engineering, Communications in Computer and Information Science (Vol. 148), Springer-Verlag Berlin Heidelberg, pp , [10] Chandra Mohan, B. and Baskaran, R. Energy Aware and Energy Efficient Routing Protocol for Adhoc Network using Restructured Artificial Bee Colony System, International Conference on High Performance Architecture and Grid Computing, Communications in Computer and Information Science (Vol. 169), Springer-Verlag Berlin Heidelberg, pp , [11] Swati Jaiswal and Chandra Mohan B, A Survey: Privacy and Security to Internet of Things with Cloud Computing, International Journal of Control Theory and Applications, Vol.9, No.42, pp ,2016 [12] Supriya Sarkar and Chandra Mohan B, Review on Recent Research in Data Mining based IoT, International Journal of Control Theory and Applications, Vol.9, No.42, pp ,

26 [13] Marietta J, and Chandra Mohan B, Review on Recent Research In IOT and its Various Applications, International Journal of Pharmacy & Technology, Vol.8, No.4, pp , 2016 [14] Arulmozhi K.S., R. Karthikeyan, B. Chandra Mohan, Optimizing Resource Sharing in Cloud Computing, International Conference on Advances in Power Electronics and Instrumentation Engineering, Communications in Computer and Information Science (Vol. 149), Springer-Verlag, pp.50-55,