Microsoft Word - ETASR_V11_N2_pp7018-7022 Engineering, Technology & Applied Science Research Vol. 11, No. 2, 2021, 7018-7022 7018 www.etasr.com Hashmi: A Novel Drone-based Search and Rescue System using Bluetooth Low Energy Technology A Novel Drone-based Search and Rescue System using Bluetooth Low Energy Technology Anas M. Hashmi Department of Electrical and Electronic Engineering University of Jeddah Jeddah, Saudi Arabia ahashmi@uj.edu.sa Abstract-Drones are widely known for their mobility and ease of use and represent a significant technological breakthrough with various applications. In this paper, a novel inexpensive Search and Rescue (SAR)-based approach for application in indoor environments is presented. The usage of Bluetooth Low Energy (BLE) has been evaluated with respect to other technologies and a conceptual view of the complete setup has been presented. Besides the cost of the drone and the locator devices, the other hardware is relatively inexpensive costing only a fraction of a US dollar. The system is believed to cover a wide area in a small-time frame ranging a few minutes, for instance, a 3600m 2 surface area could be scanned in less than 5 minutes. The system is tested by attaching a BLE device in the payload to evaluate the presence of target beacons. Potential upgrades in the system are also proposed, including design modifications for outdoor use and the application in locating missing objects. This system can confidently replace search parties dealing with missing children in public places or venues, with minimal human interaction while bearing the potential for complete automation. Keywords-BLE; wearable electronics; UAVs; localization I. INTRODUCTION During the last decade, the usage of Unmanned Aerial Vehicles (UAVs) has risen exponentially [1]. It is estimated that the global market share of multi-rotor drones is expected to reach $36.9 billion by 2022 from $6.8 billion in 2016 [2]. Drones are directly or indirectly used as an effective tool in agriculture, industrial maintenance, logistics, military, entertainment, and traffic control [3, 4]. In most of the aforementioned applications, drones have proven to be an effective replacement of the existing technologies, mainly in terms of cost maintenance, time, training, and automation. One of the most significant applications of UAVs lies in the domain of Search and Rescue (SAR)-based missions. Usually, such operations are performed in conventional matters, which are time consuming and costly [5]. For example, a successful 19- day rescue operation in July 2020 to locate two hikers lost in New Zealand cost the New Zealand authorities over 80k NZD ($57.8k) [6]. Using an autonomous UAV would potentially cut the cost and time of a SAR operation by 70% in comparison with classic operations [7]. Typically, drone-based SAR operations are carried out manually by continuously monitoring live transmitted broadcast imagery through technologies, such as the First-Person View (FPV). Some advances in the automation of the search process have been implemented which involve AI-based protocols to identify pre-trained objects or shapes, or face recognition. In confined places, such as events held at parks, theatres, or concerts, parents or guardians that lose their children usually seek the assistance of a dedicated lost child office to locate them. The proposed UAV system can potentially help in such cases or in similar cases regarding elder patients (e.g. patients with Alzheimer’s or autism). The search process in these cases can often be costly and time consuming or may involve the risk of a child/patient leaving the dedicated event area [8]. This paper discusses a novel inexpensive approach of utilizing multirotor drones as a SAR tool whereby it scans passive pre-defined passive receivers. This device overcomes the characteristic challenges of events being held in large venues, allowing object/person finding applications. Potential future upgrades for transforming the system to form a completely autonomous decision support system tool for event management may even be possible. Moreover, these developments may be uploaded to upgrade the system for open space usage with certain limitations. II. EXISTING TECHNOLOGY There are various existing drone-based SAR applications and these attempts have been featured to be utilized in various types of technologies as described in the existing literature. Table I enlists some common technologies used in a typical SAR operation. As shown in Table I, most of the technologies are focused on searching wireless radio activities emitted by target mobile phones. However, usually minors and individuals with special needs are not expected to carry mobile phones, therefore technology dependent on mobile phones may be difficult to implement in such cases. Other technologies achieve this task by different means which include live streaming footage, QR codes, or audio signals. Cellular network-based systems received enormous attention when it comes to SAR application due to the increased network coverage and enhanced communication capabilities, including the rapid expansion of 5G networks [16, 17]. Corresponding author: Anas M. Hashmi Engineering, Technology & Applied Science Research Vol. 11, No. 2, 2021, 7018-7022 7019 www.etasr.com Hashmi: A Novel Drone-based Search and Rescue System using Bluetooth Low Energy Technology TABLE I. SELECTED EXISTING SAR DRONE-BASED APPLICATION S Product Technology Features Drawbacks Ref. AltiGator IR thermal imaging Manual inspection of live footage. Manual inspction requires additional human power, thermal cameras can be expensive. [9] Intelligent drone navigation for SAR operations Computer vision QR codes are used for indoor localization. Uses GPS for localization, difficult to operate indoors [10] SARDO Mobile network and IR cameras Locates a mobile phone in 3 minutes in an open field. Mobile network-based solution, requires a phone carried by the target [11] SLAM 3D map generation from stitching still images Compares movement between two frames of the images. Images are difficult to be evaluated in crowded environemnts [12] MOBNET Mobile phones Searches for mobile phones, signal of trapped people. Mobile network-based solution, requires a phone carried by the target [13] GuideLoc Wireless signals from mobile phones Localization system uses target guiding technology based on region division. Mobile network-based solution, requires a phone carried by the target [14] Audio-based SAR with a drone Drone-embedded sound source localization Estimates the Time Difference Of Arrival (TDOA) of sound waves in a microphone pair in addition to the angle of arrival. Difficult to operate in crowded environments due to various sources of sound [15] The possibility of utilizing electronic ID tags, such as Bluetooth Low Energy (BLE) devices has been explored. We believe that featuring such inexpensive tags will make the task less complex and more achievable comparing to the methods proposed in Table I. III. CONCEPT AND METHODOLOGY The concept of the design is to re-design the drone payload in order to be able to carry the required electronics and means of communication for accurate identification of the ID tags. In order to design the system appropriately, various elements of the design have been investigated. A. ID Tags The ID tags are an important element of the process of development since they act as the primary point of communication. Tags should communicate with the host once close proximity is detected, and may be operated actively or passively. They should feature reliable operation, rapid response, inexpensive price tag, reasonable energy consumption, lightweight arrangement, and acceptable communication range. If the total weight of those tags is low, they can be manufactured as wearable electronics. The recent advances in the development of short to medium range wireless technologies provided multiple options for the UAV designed in this study. There are various communication protocols meeting the aforementioned criteria, which are enlisted in Table II [18, 19]. Referring to Table II, it can be seen that the BLE technology features have low cost, low weight, moderate communication range, and relatively low energy consumption. Conventional Bluetooth has stable connection, reasonable price, and moderate IPv6 communication energy overhead with a slow data transfer rate. As the name suggests, BLE is designed to optimize power consumption compared to the classic Bluetooth, while maintaining a similar protocol stack profile. Usually BLE devices are powered by a single coin battery that lasts up to two years. It was found that BLE consumes very little energy compared to competing low-energy communication technology, such as ZigBee [20]. It uses a 2.4GHz ISM Band with 2MHz channel spacing, and is governed by IEEE 802.11 and IEEE 802.15 standard protocols [21]. The low energy operation of BLE is achieved by the potential trade off of the following features: • Communication channels: Instead of using multiple channels for discovery or connections, only 3 broadcasting channels would be used. • Memory: Only a few bytes are featured with BLE, while additional memory would require additional hardware, increasing the chances of current leakage. • Data packets: BLE sends shorter data packets which results in fewer calibration iterations drawing less leakage current. • Sleep time: Classic Bluetooth takes over 100ms for waking from sleep, transmitting a data packet, and then going to sleep mode, while BLE performs this process 15 times faster. This will reduce scanning time for BLE saving more battery, especially if small data segments are broadcast. TABLE II. EXISTING SAR DRONE-BASED TECHNOLO GIES Technology Cost Weight Range Energy consumption Radio Frequency Identification (RFID) Very low Low Read range of 100m, unidirectional Passive Near Field Communication (NFC) Very low Low 10cm, bidirectional Very low or passive Bluetooth Medium Low 100m (class dependent), bidirectional Medium (class dependent) BLE Low Low 10-50m, bidirectional Very low Wifi High High 100m High ZigBee Moderate Low 10-20m Low BLE operates in star network topology where the nodes can be configured as slaves connected to a single master, or a master connected to multiple slaves. The master scans the designated channels to search for slave devices. Once a device is discovered, synchronous connection is established between them resulting in the change of state of activity mode from ‘sleep’ to ‘wake up’. Another option would be to set the slave devices to sleep mode by default to save energy [22]. Engineering, Technology & Applied Science Research Vol. 11, No. 2, 2021, 7018-7022 7020 www.etasr.com Hashmi: A Novel Drone-based Search and Rescue System using Bluetooth Low Energy Technology B. Data Transmission and Positioning Data are transmitted in BLE using two service profiles, namely Find Me Profile (FMP) and Proximity Profile (PXP), which take into consideration the slave/master (target/locator) positions. As the name suggests, FMP makes the localization of devices a significantly straightforward process. The target seeks the search command broadcast by the locator. Once the message is delivered, the target device is triggered, usually in the form of an audio signal or other alert mechanisms. On the other hand, PXP detects any drop in the connection with the target or initiates a connection with a different locator. This profile is effective only if a connection is initiated. Other available protocols found in the domain of consumer-based electronics include the iBeacon which operates between pre- programmed manufacturer’s devices. In terms of internal hardware arrangements in BLE devices, one point to consider is the directionality of their chip antennas, which are characterized by non-isotropic radiation. This is significantly challenging to mitigate due to the size of the additional antenna. There are various fundamental concepts when considering the positioning of BLE devices. The simplest approach of positioning is determining if both the target and the locator are within the radio coverage range, while providing a binary decision about their presence within that range [23]. In order to determine the distance between devices in free space, the following equation that relates distance (d) and the Received Signal Strength Indicator (RSSI) was utilized. � � 20�RSSI � (1) where is an offset based on the maximum RSSI of the locator. The equation could also be used to determine the lateration based on multiple locators. Unlike outdoors, the indoor environment usually exhibits multipath effect due to the presence of obstacles that may vary in form and arrangement. This significantly worsens the approximation taken into consideration in (1). However, the utilization of the Time-of- Flight (ToF) approach has proven to produce more accurate results in indoor locations. Unfortunately, ToF estimation requires additional electronics (e.g. oscillators) which increase power consumption and cost. It is also important to note that since the locator is fixed in a moving station (i.e. drone), the angulation process may be ignored. In the same context, it is highly challenging to utilize fingerprinting which identifies the pre-recorded target RSSI values of both the target and locator, since they tend to exhibit continuous change in their locations. C. Locator Movement and Beacon Analysis One of the most challenging issues when attempting to locate the target in indoor environments is the multipath propagation effect. This issue can worsen if several devices are available in a confined space and moving objects are present. Usually, no accurate RSSI model in case of indoor environments is available and a rather simplistic way of estimation involves the use of ToF values. Therefore, a worst- case scenario should be taken into consideration when planning such a system. In some commercial applications, such as the iBeacon, it is recommended that the advertising signal intervals should be set below 100ms, when the tag is operated by a coin cell battery for maximizing the battery life to a few months. In addition, when the intervals are set just below 1000ms the battery may last up to 3 years [23]. BLE devices continuously broadcast their data packets with a unique identifier known as the Universally Unique Identifier (UUID). This UUID acts as a network MAC address for the manufactured chipset and consists of 128-bits, with a 48-bit mandatory advertiser address. For BLE target finder, an open- source program, Raspberry Pi 3 BLE sniffer was used [24]. This code allows the detection of various compatible devices from different manufacturers, with all data being analyzed in a dedicated dashboard. While initiating the software, a signal search interval of 200ms was set. The advertising signal interval was set to 500ms, so that the target could operate confidently using a coin cell battery for one full year. In Figure 1, we see the Raspberry Pi (locator) attached to a drone payload for locating advertising channels. Fig. 1. Basic arrangement of target localization with scan and synchonization with the advertising beacons. IV. EXPERIMENTAL SETUP AND ROUTE PLANNER For route estimation purposes, the maximum Bluetooth coverage radius was assumed to be 10m. To ensure that all targets in the designated search zone are discoverable and that their discoverability is enhanced, the detection range was reduced to 5m. Moreover, the drone should fly in intervals whereby the movement was paused for 1000ms for each 5m flown. This interval of inactivity ensures that the locator exhibited at least 4 iterations of advertising signals broadcast by the target. For doubling the chances of discovery, each drone’s vertical/horizontal straight parallel route should have a separation of <5m. Raw experimental data are shown in Table III. TABLE III. RAW EXPERIMENTAL D ATA Flight duration Wind speed AGL Route length Stop mode 5m Vertical movement: 12×55m = 660m Horizontal lines: 11×5m = 55m Total route: 715m 1000ms every 5m flown 0m/s (indoor) For demonstration purposes, the exhibition hall Dhahran Expo, located in Dhahran, Saudi Arabia and covering a surface Engineering, Technology & Applied Science Research Vol. 11, No. 2, 2021, 7018-7022 7021 www.etasr.com Hashmi: A Novel Drone-based Search and Rescue System using Bluetooth Low Energy Technology area of 3600m 2 (60m×60m), was chosen [25]. To estimate the planned route, an open-source code named Reinforcement Learning for Autonomous navigation for UAVs [26] was used. This code was designed to implement reinforcement learning algorithms for autonomous navigation of drones in indoor environments. Figure 2 shows the exhibition hall and the planned route using predefined hall edges. The total horizontal route of the drone was 60m×12m=720m 2 , with an additional 5m×11m=55m 2 surface area of horizontal connecting lines, which makes the total route approximately around 775m 2 . Considering the average speed of the drone to be 6.7ms -1 (15mph), the total area can be covered in around 115s, with an additional 1s per 5m taken into consideration. It would take a total of 227s to completely evaluate the whole area of 3600m 2 . (a) (b) Fig. 2. (a) A raw image of hall #3 in Dhahran Expo, Dhahran, Saudi Arabia. (b) Raw image processed using planned routing with the aid of open- source code named Reinforcement Learning for Autonomous navigation for UAVs. The triangle on the left-hand side demonstrates the drone’s starting point, while the traingle on the right is suggestive of the end point of the flight. The dashed lines demonstrate the main path of vertical movement, and the solid lines represent the path of the horizontal flight. A. Accuracy and Operationality In order to ensure the maximum accuracy of the system, repeated movements may be performed. Once the target is located, the movement of the drone is locked around this area, while it continuously performs search operations in its proximity. If the connection drops, the drone should potentially move in a spiral path to reconnect with the target. The RSSI value could be taken into consideration when determining the static positioning of the drone. The proposed model was validated by using a BLE device as a target moving backward and forward from the host. The signal loss was evaluated to ensure scan signal being continuously broadcast within the expected Bluetooth range. B. Potential Applications The technological development reported in this study may be used in applications involved in locating objects or persons in halls or exhibitions. This development is easy to implement, since the BLE target devices require inexpensive hardware. BLE devices could come in different forms, such as wristbands, keychains, tags, pins, each costing less than one US dollar. Therefore, the tool shows promising potential for commercialization. One potential upgrade of this application would be its outdoors potential use. In that case, uploading a suggested route plan with details of streets or landmarks may be pertinent. A swarm drone technology may be applied to cover a wider area. V. CONCLUSIONS A BLE-based SAR system with indoor applications was proposed which potentially could be applied to locate lost individuals. The proposed BLE-based, cell-battery powered device has a year-long operational durability, and bears relatively inexpensive electronic tags in convenient packaging such as a wristband or a keychain. The drone is able to scan an indoor area of 3600m 2 with high accuracy in just below 5 minutes. This time can be further minimized with tradeoff with battery lifetime. This system can be used for additional purposes such as event management and enhanced localization capability based on beacons from target or high-density spots based on attendees’ mobile signals. Dedicated AI-based software was utilized in order to estimate the UAV route. This system could be upgraded for outdoor operations using swarm technology. REFERENCES [1] S. Maharana, "Commercial Drones," International Journal of Advances in Science Engineering and Technology, vol. 5, no. 1 Suppl. Is. 3, pp. 96–101, Mar. 2017. [2] D. Gonzalez-Aguilera and P. Rodriguez-Gonzalvez, "Editorial: Drones—An Open Access Journal," Drones, vol. 1, no. 1, Dec. 2017, https://doi.org/10.3390/drones1010001. [3] Z. Liu, Z. Li, B. Liu, X. Fu, I. Raptis, and K. Ren, "Rise of Mini-Drones: Applications and Issues," in Proceedings of the 2015 Workshop on Privacy-Aware Mobile Computing, New York, NY, USA, Jun. 2015, pp. 7–12, https://doi.org/10.1145/2757302.2757303. [4] M. Almutiry, "UAV Tomographic Synthetic Aperture Radar for Landmine Detection," Engineering, Technology & Applied Science Research, vol. 10, no. 4, pp. 5933–5939, Aug. 2020, https://doi.org/10.48084/etasr.3611. [5] G. Remy, S. M. Senouci, F. Jan, and Y. Gourhant, "SAR drones: drones for advanced search and rescue missions," Journées Nationales des Communications dans les Transports, vol. 1, pp. 1–3, 2013. [6] "Cost of missing tramper rescue operation $80,000 and counting," Stuff, Jul. 21, 2020. https://www.stuff.co.nz/national/122198869/cost-of- missing-tramper-rescue-operation-80000-and-counting (accessed Mar. 24, 2021). [7] W. Stecz and K. Gromada, "UAV Mission Planning with SAR Application," Sensors, vol. 20, no. 4, Jan. 2020, Art. no. 1080, https://doi.org/10.3390/s20041080. [8] S. A. Riggins and D. P. Sharp, "Missing child reporting, tracking and recovery method and system," US20100262367A1, Oct. 14, 2010. [9] "Drones for search & rescue missions," AltiGator Drone & UAV Technologies, Sep. 17, 2014. https://altigator.com/en/drones-for-search- rescue-missions/ (accessed Mar. 24, 2021). [10] D. Yavuz, H. Akbıyık, and E. Bostancı, "Intelligent drone navigation for search and rescue operations," in 2016 24th Signal Processing and Engineering, Technology & Applied Science Research Vol. 11, No. 2, 2021, 7018-7022 7022 www.etasr.com Hashmi: A Novel Drone-based Search and Rescue System using Bluetooth Low Energy Technology Communication Application Conference (SIU), Zonguldak, Turkey, May 2016, pp. 565–568, https://doi.org/10.1109/SIU.2016.7495803. [11] A. Albanese, V. Sciancalepore, and X. Costa-Perez, "SARDO: An Automated Search-and-Rescue Drone-based Solution for Victims Localization," IEEE Transactions on Mobile Computing, 2021, https://doi.org/10.1109/TMC.2021.3051273. [12] F. R. Heukels, "Simultaneous Localization and Mapping (SLAM) : towards an autonomous search and rescue aiding drone," M. S. thesis, University of Twente, Enschede, Netherlands, 2015. [13] M. Półka, S. Ptak, and Ł. Kuziora, "The Use of UAV’s for Search and Rescue Operations," Procedia Engineering, vol. 192, pp. 748–752, 2017, https://doi.org/10.1016/j.proeng.2017.06.129. [14] A. Wang et al., "GuideLoc: UAV-Assisted Multitarget Localization System for Disaster Rescue," Mobile Information Systems, vol. 2017, Mar. 2017, Art. no. e1267608, https://doi.org/10.1155/2017/1267608. [15] A. Deleforge, D. D. Carlo, M. Strauss, R. Serizel, and L. Marcenaro, "Audio-Based Search and Rescue With a Drone: Highlights From the IEEE Signal Processing Cup 2019 Student Competition [SP Competitions]," IEEE Signal Processing Magazine, vol. 36, no. 5, pp. 138–144, Sep. 2019, https://doi.org/10.1109/MSP.2019.2924687. [16] S. K. Khan, A. Al-Hourani, and K. G. Chavez, "Performance Evaluation of Amplify-and-Forward UAV Relay in Millimeter-Wave," in 2020 27th International Conference on Telecommunications (ICT), Bali, Indonesia, Oct. 2020, https://doi.org/10.1109/ICT49546.2020.9239435. [17] S. K. Khan, M. Farasat, U. Naseem, and F. Ali, "Performance Evaluation of Next-Generation Wireless (5G) UAV Relay," Wireless Personal Communications, vol. 113, no. 2, pp. 945–960, Jul. 2020, https://doi.org/10.1007/s11277-020-07261-x. [18] S. J. Danbatta and A. Varol, "Comparison of Zigbee, Z-Wave, Wi-Fi, and Bluetooth Wireless Technologies Used in Home Automation," in 2019 7th International Symposium on Digital Forensics and Security (ISDFS), Barcelos, Portugal, Jun. 2019, pp. 1–5, https://doi.org/ 10.1109/ISDFS.2019.8757472. [19] N. Chhabra, "Comparative Analysis of Different Wireless Technologies," International Journal of Scientific Research in Network Security and Communication, vol. 1, no. 5, pp. 13–17, Dec. 2013. [20] M. Siekkinen, M. Hiienkari, J. K. Nurminen, and J. Nieminen, "How low energy is bluetooth low energy? Comparative measurements with ZigBee/802.15.4," in 2012 IEEE Wireless Communications and Networking Conference Workshops (WCNCW), Paris, France, Apr. 2012, pp. 232–237, https://doi.org/10.1109/WCNCW.2012.6215496. [21] R. Heydon, Bluetooth Low Energy: The Developer’s Handbook, 1st ed. Pearson, 2012. [22] C. Gomez, J. Oller, and J. Paradells, "Overview and Evaluation of Bluetooth Low Energy: An Emerging Low-Power Wireless Technology," Sensors, vol. 12, no. 9, pp. 11734–11753, Sep. 2012, https://doi.org/10.3390/s120911734. [23] A. a. M. K. Abuelgasim and K. M. Yusof, "High Speed Mobility Management Performance in a Real LTE Scenario," Engineering, Technology & Applied Science Research, vol. 10, no. 1, pp. 5175–5179, Feb. 2020, https://doi.org/10.48084/etasr.3245. [24] "diyActive: Make a Raspberry Pi Hub," ReelyActive. https://reelyactive. github.io/make-a-pi-hub.html (accessed Mar. 24, 2021). [25] "Exhibition Halls," Dhahran Expo. https://www.dhahranexpo.com.sa/ venue-facilities-exhibition-halls.aspx (accessed Mar. 24, 2021). [26] M. Liaq and Y. Byun, "Autonomous UAV Navigation Using Reinforcement Learning," International Journal of Machine Learning and Computing, vol. 9, no. 6, pp. 756–761, Dec. 2019, https://doi.org/ 10.18178/ijmlc.2019.9.6.869.