Multi-Hop Linear Network Based on LoRa

Transcription

1 , pp Multi-Hop Linear Network Based on LoRa Cong Tan Duong 1 and Myung Kyun Kim Dept. of IT Convergence, Univ. of Ulsan, Daehak-Ro 93, Nam-Gu, Ulsan, Korea mkkim@ulsan.ac.kr Abstract. Long Range (LoRa) communication is known as new technology communication for Internet of Things (IoT). LoRa technology offers a very compelling mix of long range, low power consumption and secure data transmission. Networks system using this technology can provide coverage that is greater in range compared to that of existing cellular networks. By multi-hop communication, we argue that LoRa is potentially very useful for the monitoring system for wide area with long distance. In this paper, we propose a protocol with multi-hop communication for linear network including network constructing period and data operating period using LoRa. The experimentation performed in the university area shows a delivery reliability of above 90% Keywords: LoRa, Multi-hop Communication, IoT, Linear Network. 1 Introduction New Low-Power Wide-Area Network (LPWAN) technologies [7] such as LoRa [6], Sigfox [5], RPMA [4] and Weightless are emerging which enable power efficient wireless communication over very long distances. Besides improved communication range, the LPWAN technologies have unique features stemming from the used modulation schemes. Therefore, the existing Medium Access Control (MAC) protocols and routing mechanisms are not efficient to simply adapt these LPWAN technologies. In this paper, we use LoRa as technology for building an IoT system. We propose a protocol with multi-hop transmission for linear network. It is suitable for monitoring application in which need coverage long distance and every monitored point is almost placed as line style such as gas line system, high voltage line system, etc. To evaluate performance of protocol, we perform an experimentation using 5 LoRa nodes (1 sink node and 4 sensor nodes) equipped with a Semtech SX1272 transceiver. LoRa [1] is a proprietary spread spectrum modulation technique by Semtech. It is a derivative of Chirp Spread Spectrum(CSS) with integrated Forward Error Correction (FEC). A LoRa transceiver can decode transmissions 19.5 db below the noise floor, thus, enabling very long communication distances. LoRa key properties consist of long range, high robustness, multipath resistance, doppler resistance and low power. The remainder of this paper is organized as follows. The design of protocol is presented in detail in Section 3. Section 4 evaluates the performances of protocol base on result of experimentation. Finally, we conclude the paper. ISSN: ASTL Copyright 2018 SERSC

2 2 Protocol Design 2.1 Protocol Overview In this section we describe the design of protocol with multi-hop communication for linear network in detail. The protocol structure consists of two periods: Network Initialization Period and Data Operation Period. Fig. 1. Network Initialization Period Fig. 2. Data Operation Period Network Initialization Period The Network Initialization Period will be performed at the beginning of life time to forming linear network by every node and perform from sink node. The purpose of network initialization period (NIP) is to make sure that every node in the network is placed at the position that forming a linear network is possible and after performing network successfully, every node will go to Data Operation Period at the same time. In this protocol, we assume that maximum depth of network is fixed. This period will be performed in certain time interval T NIP. To performing initialization network, a Request Init Message (REQ_INIT) is used. It includes sender Id, maximum depth of network and passed time. The passed time information is to notify receiver that how many REQ_INIT packet duration time the initialization process already undergone. Base on passed time, the node can calculate the remaining time of initialization period. The Fig. 1 shows how the network initialize in detail. After powering on, every sensor node goes to listening mode to overhear REQ_INIT from its previous grade node. And the sink node start forming network by broadcast REQ_INIT including maximum depth of network and passed time is 1. It is supported re-transmission mechanism by performing overhear to track the REQ_INIT of next grade node after sending to know the next grade node received REQ_INIT message or not. Every node will perform similarly after when they receive proper REQ_INIT. A node will go to sleep and finish initialization process if it tracks successfully or after sending REQ_INIT in case leaf node. 30 Copyright 2018 SERSC

3 2.1.2 Data Operation Period In the data operation period, every sensors node forward data packet in pipeline fashion to send data from leaf node to sink node. A node can combine the received data packet from next grade node with its own data and send them together. Every node use slot with certain duration time dur data to receiving or sending data packet. The dur data has to be considered greater than duration time for transmitting data packet with maximum length after combining data at node 1. To receiving the data packet, every node have to wake up at the same sending time of next grade node. In the first cycle, every node the wake-up time based on its hop-count and maximum depth of network as below: T wake_n = (MaxDepth depth n 1) dur data (1) The most important thing to insure protocol operating well is time synchronization from second cycle and so on. Because of clock drift, every node will be loss synchronization after long enough time. To avoid it, a node can be seen as time source of previous grade node. Right after receiving data packet from next grade node, it will calculate wake-up time for next cycle after duration time T wake as below: T wake = dur cycle dur tx T short (2) where dur tx is time on air of packet which is sent by next grade node and it is calculated by formula [3], T short is short time the MCU need to actually change to TX mode in Radio Chip to send packet. 3 Performance Evaluation In this section, performance of the protocol will be evaluated under result of experimentation in real environment. The protocol is deployed on MultiConnect Mdot which is LoRa modules using STM32F411RET Processor and radio chip SX1272 and is developed by MultiTech. The implementation is based on LMiC library [8], which is developed by IBM for LoRaWAN in C. We deployed a 5 node network on the university campus as shown in Fig. 3. We run the test for every SF with BW125, TX power of 14 dbm, frequency is 922MHz. Nodes are set to transmit a data packet with 5 bytes of its own data and combining the received data from next grade node. The packet is sent with fixed preamble length of 8 symbol. In this experiment, we implemented a listen period of 8 symbol time in each slot for receiving. The table 1 shows duration of data and cycle duration for each spreading factor after calculating the time on air for maximum length as 20 bytes and adding more guard time to avoid conflict among tasks in MCU. For each SF, we run a test with duration of 200 data cycles time. Copyright 2018 SERSC 31

4 Fig. 3. Location Deployment Table 1. Duration parameters SF12 SF11 SF10 SF9 SF8 SF7 dur data (ms) dur cycle (ms) In our evaluation, the delivery reliability and the throughput of entire network for each spreading factor are shown as the Fig. 4 and Fig. 5. The delivery reliability achieves above 90%, and they trend decreasing from SF12 to SF7 when we use same node location deployment. Fig. 4. Delivery reliability of entire network Fig. 5. Throughput for entire network 4 Conclusion By multi-hop communication, LoRa is potentially very useful for the monitoring system for wide area with long distance. We contributed a protocol with multi-hop transmission for linear network including network construction period and data operating period. With the result in performance evaluation part, this protocol is applicable for monitoring application in which need coverage long distance and every 32 Copyright 2018 SERSC

5 monitored point is almost placed as line style such as gas line system, high voltage line system, etc. Acknowledgments. This research was supported by Basic Science Research Program through the National Research Foundation of Korea(NRF) funded by the Ministry of Education (KNRF ). References 1. Martin Bor, John Vidler and Utz Roedig: LoRa for the Internet of Things 2. Wei Ye, Fabio Silva, and John Heidemann: Ultra-Low Duty Cycle MAC with Scheduled Channel Polling 3. Datasheet SX1272: Accessed: 24/11/ Ingenu RPMA. Accessed: Sigfox. Accessed: LoRa. Accessed: Usman Raza, Parag Kulkarni, and Mahesh Sooriyabandara. Low Power Wide Area Networks: A Survey. 8. LMiC. Accessed: Copyright 2018 SERSC 33