Abstract: Wireless sensor networks (WSNs) are target perception and monitoring systems built upon sensing, communication, and computer technologies. Due to their immense application potential, they have attracted widespread attention from academia and industry. Localization is a crucial issue in the design and application of sensor networks. The location information of sensor nodes has significant theoretical and practical implications for sensor network protocol design, network management, inter-node cooperation, and target tracking. This paper details the implementation strategy of a sensor network localization algorithm based on acoustic ranging, analyzes the specific implementation mechanism, and provides relevant conclusions. Keywords: sensor network, range-based localization, acoustic ranging [b][align=center]Ranging and Localization in Wireless Sensor Network Liu Gang, Zhou Xingshe, Ma Junyan, Wang Gang[/align][/b] Abstract: A wireless sensor network integrates sensor, MEMS, and communication networks. More and more academic researchers and insiders are engaged in developing wireless sensor networks due to their bright future with various applications. In this paper, range-based localization is discussed in detail and our acoustic-based ranging approach is presented, then the algorithm performance and some conclusions are provided. Keywords: wireless sensor network, range-based localization, acoustic-based ranging 1. Introduction The continuous advancement of processor, storage, and wireless communication technologies has opened up new development space for networked sensors. A networked sensor is not just a simple sensor; it consists of a processor, memory, wireless transceiver, battery, and a set of sensors, possessing multiple functions such as sensing, computing, and communication. Moreover, with the continuous innovation of manufacturing processes, the size of networked sensors is becoming smaller and smaller, hence the figurative term "intelligent dust." Wireless sensor networks composed of these "smart dust particles" are widely used in various fields, such as environmental monitoring and military reconnaissance. Localization is a crucial issue in the design and application of sensor networks. The location information of sensor nodes has significant theoretical and practical implications for sensor network protocol design, network management, inter-node collaboration, and target tracking. 2. Ranging-Based Localization 2.1 Hardware and Software Environment and Feasibility Analysis The nodes we used are CrossBow's MPR400CB, and the processors used on the nodes are ATmega128L. The ATmega128L processor has a clock speed of approximately 7.4MHz and a CPU clock cycle of 136ns. Furthermore, the ATmega128L processor uses a reduced instruction set and pipelined technology; most instructions execute in one CPU clock cycle, while interrupt response takes four CPU clock cycles. If we measure a distance of 10cm, assuming the speed of sound is 340m/s, then ΔT should be approximately 0.294ms. Assuming each instruction executes in 136ns, the node can execute approximately 2161 instructions within ΔT, indicating the processor's processing speed is sufficient. Furthermore, the processor's counter crystal oscillator can reach a frequency of up to 7373 times per millisecond, providing sufficient accuracy for timing. Therefore, there are no issues with timing in acoustic ranging. The sensor board we use is the CrossBow MTS310CA. The MTS310CA integrates both a buzzer and a microphone. The buzzer is a simple 4kHz piezoelectric resonator. Its drive and frequency control circuitry are internally integrated; operation requires only a switch. The MTS310CA provides a microphone with two main uses: sound ranging and sound recording and measurement. The microphone has a built-in filter and sound decoder, which can be used to monitor the 4kHz sound emitted by the buzzer on the MTS310CA sensor board. The operating system running on the node is TinyOS. TinyOS is an open-source operating system designed by the University of California, Berkeley for low-power embedded network sensors. It is an event-driven embedded operating system that supports the concurrent intensive operations required by networked sensors with minimal hardware configuration. It is a micro-threaded, event-triggered operating system. It retains many characteristics unique to networked sensors, efficiently manages the hardware system, and supports concurrent intensive operations, achieving efficient modularity and robust operation in a sense. 2.2 Algorithm Implementation To use sound waves for ranging and improve the minimum distance resolution to at least 10 cm, we must modify the TimerC provided by TinyOS. Considering that some other important components of TinyOS, such as the inter-node wireless communication module GenericComm.nc and the analog-to-digital converter ADCC, all use the TimerC component provided by the system, if we want to modify TimerC, we must ensure that the original functionality remains unchanged during the modification. Another approach is to utilize another hardware resource in the processor, namely Timer/Counter1, and write a dedicated Timer. Comparing the two methods, the former saves hardware resources by slightly improving upon the original approach; the latter is simpler, but it uses a dedicated timer/counter on the processor to implement only one Timer, which is wasteful from a hardware perspective, although it can achieve a very high timing speed and can be used for high-frequency sampling. The basic idea of ​​this method is: the Beacon node first transmits a ready-to-measure message to the Listener node via a wireless transmitter at fixed time intervals, which includes a timestamp. When a ready-to-measure message cannot be sent in one go, the Beacon node increments the timestamp by 1. Here, the Beacon node uses two timers: the first timer controls the time interval for sending the ready-to-measure message to the Listener, and the second timer controls the beeping time. Here, the two nodes achieve synchronization through the SendMsg.sendDone(TOS_MsgPtr sent, result_t success) event of the SendMsg interface used by the Beacon and the event TOS_MsgPtr ReceiveMsg.receive(TOS_MsgPtr m) of the ReceiveMsg interface used by the Listener. That is, for the Beacon, when SendMsg is completed, a SendMsg.sendDone event is generated, and the Beacon will turn on its buzzer for a period of time; for the Listener, when ReceiveMsg generates a ReceiveMsg.receive event, the Listener node starts timing. 2.3 Result Analysis The program runs on the sensor node. The data received by the base station (as shown in Figure 1) is the measurement result when the distance between the two nodes is 20 cm. The value sent back to the base station multiplied by 10 cm is the actual distance between the two nodes. [align=center] Figure 1 Raw data received by the base station[/align] The horizontal lines in Figure 1 represent the distance measured by the nodes, with a scale factor of 10 cm. The data stream reveals unstable measurement results with significant drift. However, a correct result of 02 (20 cm) is observed. Similarly, data with distances of 30 cm, 40 cm, 50 cm, 60 cm, 70 cm, and 80 cm show similar characteristics. Careful analysis of the error distribution in acoustic ranging reveals a rather arbitrary distribution that cannot be fitted using commonly used distributions such as the normal distribution. However, a simple minimum value filter yielded good results. Careful observation of the received data stream shows that while the distribution of measurement deviations at the same distance appears irregular, within a certain time window (e.g., 10 consecutive measurements at the same distance), there is always a correct measurement that matches the true distance, and this measurement has the smallest value. Based on the characteristics of the data analyzed above, we decided to use the minimum value filtering method to improve the measurement. This involves finding the smallest value among a set of continuous data (experiments showed that the length of 10 consecutive data points is the shortest distance while ensuring an accuracy rate greater than 98%) to represent the ranging result over a period of time. Thus, a single ranging measurement between two nodes is obtained by performing 10 consecutive ranging measurements on each node and applying minimum filtering. However, this method has a significant drawback: it increases the time required for ranging between nodes. If the original distance measurement was performed every second, with this algorithm and a filtering length of 10, it becomes every 10 seconds, meaning a positioning measurement every 10 seconds. This reduces the real-time performance of the measurement. To improve real-time performance, the time interval between each Beacon transmission of the ranging signal can be shortened, such as every 100 milliseconds. This reduces the positioning and ranging time to once per second under the same algorithm. The algorithm is improved by obtaining a single ranging value between two nodes through 10 consecutive ranging measurements and applying minimum filtering. The improved program was burned onto the sensor nodes, and the distance between the two nodes was set to 10 cm, 20 cm, 30 cm...200 cm, and the data received by the base station was observed. The received data, i.e., the measured value, was relatively stable, and the maximum error between it and the actual distance between the two nodes was within ±10 cm. The reason for using minimum value filtering here is that our time reading can never be faster than the arrival time of the sound, so using the minimum value estimate should be the closest to the true value. 3. Conclusion In sensor networks, time synchronization is a core scientific problem that needs to be solved, and it is also a difficult problem. In the process of node ranging and positioning, we need accurate time synchronization. By using the method of interrupting a specific sound frequency (the sound emitted by a 4kHz buzzer) to determine whether the sound has arrived and thus to measure the distance, we have drawn the following conclusions: l The external environment has a significant impact on ranging, especially external noise and the echo of the buzzer sound; l When the distance is far, if it is indoors, the echo will have a great impact on the ranging result, and even if the minimum value filtering method is used, a stable value will not be obtained. To eliminate the influence of echo, the buzzer's sound interval should be shortened, and the time interval between two sounds should be lengthened to eliminate the indoor echo effect. After eliminating the echo effect using the above method, stable and accurate data can be obtained in long-distance measurements. The distance resolution can be controlled by changing the trigger interval of the Listener timer, i.e., calling Timer.start2(TIMER_REPEAT, 10). Decreasing the value of the second parameter can improve the resolution, but it should not be too small, as this will increase the load and make the measurement results inaccurate. Furthermore, due to the microphone's lack of sensitivity, setting its value too small has little impact on the measurement results. Experiments show that 10 cm resolution is a good value, and most international sound ranging methods also have an error of 10 cm. Due to hardware limitations, the error cannot be further reduced. This method has a relatively long positioning time, which is not as good as signal strength positioning. It is suitable for positioning in network topologies where the relative positions of nodes are relatively stable or do not change rapidly. The author's innovation lies in achieving approximate time synchronization through a simple and effective timestamp and performing minimal filtering on the measurement data to achieve ranging and positioning. References: [1] Bahl P, Padmanabhan VN. RADAR: An in-building RF-based user location and tracking system. In: Proc. of the IEEE INFOCOM 2000. Vol.2, Tel Aviv: IEEE Computer and Communications Societies, 2000. [2] Savarese C, Rabaey JM, Beutel J. Locationing in distributed ad-hoc wireless sensor network. In: Proc. of the 2001 IEEE Int'l Conf. on Acoustics, Speech, and Signal. Vol.4, Salt Lake: IEEE Signal Processing Society, 2001. [3] Doherty L, Pister KSJ, Ghaoui LE. Convex position estimation in wireless sensor networks. In: Proc. of the IEEE INFOCOM 2001. Vol.3, Anchorage: IEEE Computer and Communications Societies, 2001. [4] Piyush Gupta, PRKumar, The Capacity of Wireless Networks. IEEE Transactions on Information Theory, 2000, 46(2) [5] Dai Qingyu, Ji Zhenzhou. Design of Sound Sensor Nodes. Microcomputer Information, 2006, 5-1, pp. 146-147