Keywords

1 Introduction

In a general hospital environment, in case of emergency, it is needed to send the emergency information in real time to the persons in charge by using SMS or telephone call for respond the situation. The persons in charge with the corresponding situation have to respond as soon as possible. Based on the general emergency call method, if the emergency information cannot be timely conveyed to the persons in charge, there will delayed the deployment of the persons in charge and cannot respond be in time. A system that can send emergency information to persons in charge by using multiple communication methods when an emergency situation occurs and the persons in charge can respond to the emergency in real time is proposed. The system is also can record and manage all emergency activity information.

Figure 1 shows the multiple communication concept of the proposed system.

Fig. 1.
figure 1

The communication concept of the emergency deployment system

Full size image

2 Previous and Related Studies for Indoor Positioning

In this study, the management server uses one-way broadcast communication to send emergency alarm message to persons in charge, according to the emergency reaction protocol, then change to two-way communications by confirming from the client to exchange emergency information with clients, all the activities and position of the responder are tracked in real time.

Bluetooth LE (Bluetooth Low Energy), visible light communication technology, SMS message can be used as one-way broadcast communication method, Wi-Fi and cellular mobile communications can be used as two-way TCP communication methods in this system.

2.1 Bluetooth LE

By growing importance of IoT (Internet of Things), Bluetooth LE application is also diffused. Compared with the Bluetooth, it does not need to maintain a long connection, therefore power consumption is also greatly reduced.

BLE does not need pairing that can send data to multi clients in one-way, this process is called Advertising. All the terminals which Bluetooth enabled can receive the advertising data. When a message is transmitted via advertising, the message can be receiving by multi users in the same time.

The BLE device for advertising is called beacon and it is possible to indoor positioning by using RSSI (Received Signal Strength Indication) value of beacon signal [1]. For indoor positioning using beacons, Cell ID method is mainly used. It is a method of judging in which cell the position of its own is located based on the information received from the fixed node. This method is easier and simpler than other methods, but has the disadvantage of low accuracy. Indoor positioning using beacons is mainly used for hospital, department store and etc.

2.2 Two-Way Communication and Indoor Positioning Based on Wi-Fi

Socket communications have two major transport layer protocols, TCP (Transfer Control Protocol) and UDP (User Datagram Protocol).

The UDP communication has no handshaking dialogues among the processes for session, and cannot provide any reliability control so there is no guarantee of delivery, ordering, or duplicate protection of datagrams. The TCP communication provides reliable, ordered and error-checked delivery of a stream of octets between processes which are running on hosts based on IP network. In this study, TCP is used for socket communications.

A TCP communication is requested by using the IP addresses and port numbers of two processes. The client starts the connection and if the server does not accept the response, the connection will be failed. If connection is created, data can be exchanged through the socket between the server and client. The TCP can provide error control and flow control, if the data not received properly, it can be requested again [2].

GPS (Global Positioning System) is the most widely used outdoor positioning system. But in the indoor environment, GPS cannot work properly. Therefore, many research of indoor positioning are in progress. For example, visible light communications [3], BLE signal strength [1, 9], Wi-Fi signal strength and so on can be used [4, 5]. For finding the coordinate of mobile object in indoor environment, fingerprint and triangulation methods are widely used.

The triangulation method is a geometric method of computing a coordinate by calculating each distance from three reference points. It is necessary to be able to obtain the correct distance from the wireless AP as reference point so that the correct position can be traced [10].

Fingerprint method is RSSI (Received Signal Strength Indication)-based, but it simply relies on the previously recorded data of the signal strength from several reference access points in the proper range. Storing this information in a database along with the known coordinates of the tracking device is clone in an offline phase. During the online position decision phase, the current RSSI vector at an unknown location is compared to those stored in the fingerprint DB and the closest matching position is returned as the estimated user location [8].

Wi-Fi or BLE signal strength can be used for calculating distance for using the fingerprint method.

2.3 Emergency Deployment System in a General Hospital

When a patient meet in an emergency situation, this system can send the emergency notice to the persons in charge on various departments, then by the confirming to the notice which means the corresponding departments can give an initial solution. Indoor positioning is used to guide the persons in charge to the current location of the patient, then lead those on the most appropriate route to the operating room.

In an emergency situations, excessive noise in a hospital has been attributed to negative clinical outcomes to patients and perhaps give negative performance and stress as well to all staffs. It is essential to use a plain language emergency code only for an appropriate individuals to initiate an immediate and appropriate response.

When initiating an emergency code call by the control center in a general hospital setting, the notification process for specific doctors and nurses will be initiated with single-way multiple task communication. Once the emergency code message has been effectively sent to emergency responders, staff will press button to open bilateral one to one communication and exchange information about the target locations and specific emergency operational plan with server. Start with confirmation response from emergency response staffs, initiating protocol information will be recorded in server.

Server with location tracking system manage staff’s current position, moving route and the time required after confirmation response in real time. If delayed response is occurred, server sends urgent messages and emergency operational plan to emergency response staff. But if emergency response staff cannot response emergency call, server initiates a contingency plan. It provides alternate medical staff and notifies other emergency response staff about current situation. Emergency code should be finished when situation has been managed or resolved.

Every process such as the time required from beginning to end, staff location per time, situation status and so on is recorded in database. With these data, further analysis and improvement of the system will be managed.

3 Comparison of Energy Consumption on Wi-Fi and BLE

Different communication methods not only have difference on data rate and communication distance, power consumptions are also different. Therefore, the power consumption tendencies are seriously considered in the proposed system.

3.1 Required Energy for Operating Hardware Modules

The WL18xxMOD WiLink 8 Single band Combo Module developed by TI which is widely used in portable mobile devices has be selected for power consumption analysis of Wi-Fi device. One important selection reason is the chip on the device uses the same antenna for Wi-Fi and Bluetooth communications.

The chip’s operating power of Wi-Fi communication showed on Table 1 [6]. It can be seen that, depending on the negotiation data rate and the encoding method, the operating power of chip at 49 mA to 85 mA when the chip used as a Wi-Fi receiver. As a Wi-Fi transmitter, the operating power of chip surged to 238 mA to 420 mA.

Table 1. Operating current of Wi-Fi communication
Full size table

The chip’s operating powers of Bluetooth BR (Basic Rate) and EDR (Enhanced Data Rate) communication showed on Table 2 [6], The operating power of each case in a range from 178 µA to 33 mA.

Table 2. Operating currents of Bluetooth
Full size table

Table 3 showed the chip’s operating powers of Bluetooth LE (BLE). The current values are from 124 µA to 266 µA [6].

Table 3. Operating current of Bluetooth LE(BLE)
Full size table

The power consumption of Wi-Fi is a thousand times than that of Bluetooth LE.

3.2 Energy Consumptions Based on Software Structure

The power consumption of TCP/IP based two-way communication protocol via Wi-Fi, one-way Bluetooth LE communications, and using Google Cloud Message (GCM) Push [7] are tested. The confirmation of power consumption software used in this study is done based on the BroadcastReceiver function of the Android operating system. Figure 2 is the flow chart of this program. When an event is transmitted from the Android system, the receiver application which is corresponded can be executed.

Fig. 2.
figure 2

Flow chart of power consumption analysis program

Full size image

In power consumption evaluation, data package sends total 60 times in 10 min. Then the power consumption results are measured after completion of communications. Samsung Galaxy S3 is used as receiver, ipTime A3004 wireless router is used as Wi-Fi transmitter, Raspberry Pi 3 is used as BLE transmitter in this experiment. The results of power consumption experiment are shown on Table 4.

Table 4. Result of Power consumption experiment
Full size table

As shown at the tables, Bluetooth LE communication and GCM Push consumes less power than Wi-Fi communication in both tests.

It is definite that the needed operating power based on hardware modules of Wi-Fi is a thousand times higher than those of Bluetooth LE communications. From a software point of view, the difference of power consumption is also obvious. Therefore, in this system, GCM Push and BLE communication are preferred for sending an emergency one-way notification to conserve power, and Wi-Fi communications is used for ending the specific solution among the server and clients because a large amount of data exchange is required in this process.

3.3 The Process for Switching Between One-Way and Two-Way Communications

BLE requires less energy but, Wi-Fi requires a relatively large amount of energy. For the robustness of the device which should always wait for communication, BLE is used for one-way communication. And Wi-Fi is used for two-way communication for indoor positioning and emergency information transmission.

Figure 3 shows the process for switching between one-way and two-way communications.

Fig. 3.
figure 3

The process for switching between one-way and two-way communication

Full size image

4 Indoor Positioning Based on Fingerprint

For indoor positioning using fingerprint, The tracking data must be collected in advance. We store the pre-collected tracking data based on RSSI value at various indoor positions in the database.

Figure 3 shows an experimental environment with 4 APs in corridor. Tracking data is collected at 108 points.

4.1 Methods of Collecting Tracking Data

For accurate tracking data, accurate RSSI values should be collected. But RSSI values have errors and are measured differently each time. Mean, median, Kalman filter can be used to collect accurate RSSI value.

The mean is the average value of all data. The operation is simple, but if there is extremely large error, the value can be distorted.

The median is the middle value among the collected data sorted by size. If the number of data is odd, select the middle value, and if it is even, average the two numbers on both sides of the middle.

The Kalman filter is based on measurements made according to time and can give more accurate results than other measurements [11]. It can process data containing noise, and can optimal statistical prediction.

Table 5 shows RSSI values of AP1 measured at any point, and Fig. 4 shows comparison of results of mean, median, and Kalman filter.

Table 5. RSSI values of AP1
Full size table
Fig. 4.
figure 4

An experimental environment in corridor

Full size image

4.2 Calculating Similarity Distance

Fingerprint uses not the absolute distance but the similarity distance. The similarity distance should be calculated using the RSSI value at each test point with the pre-collected RSSI value in the tracking data. We use the K-NN (K-Nearest Neighbor) algorithm to calculate the similarity distance.

$$ D_{i} = (\sum\nolimits_{j = 1}^{n} {\left| {S_{j} - S_{ij} } \right|^{q} } )^{1/q} $$
(1)

In K-NN algorithm of Formula (1), ‘n’ is the total number of wireless APs, ‘j’ is the ID number of wireless APs, and ‘i’ is the test point number where the signal strength is collected. ‘q’ is a distance constant, that has 1 for Manhattan distance method and 2 for Euclidean distance method. ‘Di’ is the similarity distance, ‘Sj’ is the signal strength of the jth AP at the positioning point, and ‘Sij’ is the signal strength of the jth AP at the ith collection point [12].

The nearest point can be found by comparing the similarity distance obtained at each point. But it is inefficient to calculate the similarity distance of all 108 points and then compare them.

After finding the nearest AP using the Cell ID method, it is possible to reduce the amount of calculation by comparing only the data around the AP. Figure 5 shows the range where the signal intensity of each AP is the strongest, and we can select from 18 to 36 points for each AP.

Fig. 5.
figure 5

Comparison results of Mean, Median and Kalman filter

Full size image

When calculating using the selected data, the amount of computation is reduced to 1/3 compared to the total data, and the computation time also decreases.

4.3 Position Tracking and Error Analysis

Tracking data is collected by filtering 30 RSSI values at each tracking point. And then, as shown in Fig. 6, position tracking was performed 5 times at 10 positioning points (Fig. 7).

Fig. 6.
figure 6

The range where the signal intensity of each AP is the strongest

Full size image
Fig. 7.
figure 7

10 positioning point in experimental environment

Full size image

The errors in each positioning point are as shown in Table 6.

Table 6. The errors of indoor positioning using fingerprint in each point
Full size table

The errors of indoor positioning using fingerprint in each point are from 0 m to 16.2 m. The average error is about 4.39 m. For the purpose of positioning a moving object in the building, the error is tolerable.

5 Conclusions

An Emergency Deployment System based on multiple-communications. For practical alarm system model is designed and some experiments for checking the performance of unit technologies are performed.

For the multiple communication methods, energy consumptions of 1 to N one-way communication for sending one way codes and communication method for guiding and tracking the responders have been analyzed. The GCM Push and BLE communications are preferred for sending an emergency one-way notification to conserve power. Wi-Fi communications are preferred for two-way communications among the server and clients because a large amount of data exchange is required in this process.

The effects of indoor positioning based on fingerprint is analyzed. The experiment result shows that the proposed system can be used as practical one.

For improving accuracy of positioning, some better methods have to be studied. One candidate method is UWB (Ultra Wide Band) for better accurate positioning decision.