A Method to Locate Tree Positions Using Ultrawideband Technology

Tree position plays an important role in research on forest resources and ecological functions, and quickly and accurately obtaining tree position data has long been the focus of investigators. However, the classical method is time-consuming and laborious; thus, a convenient method of measuring tree position is needed. The primary achievements of this study include the following: (1) a device was designed for precise location of trees; (2) a new location algorithm was proposed for pentagonal localization based on the received signal strength indication and ultrawideband technology; and (3) a PC software application was developed for automatically storing and uploading tree position data. The device was applied to 10 circular plots with a diameter of 24m to test the positioning speed and accuracy. The results showed that the tree positions could be accurately estimated. On the xand y-axes, the biases were -3.94 and 3.36 cm, respectively, and the root mean square errors (RMSEs) were 28.39 and 28.53 cm, respectively. The mean error (Ed) between the estimated and reference distances was 36.13 cm, and the standard deviation was 16.67 cm. The device is inexpensive and easy to use and carry in the field; thus, it is suitable for locating trees in environments with complex terrain.


Introduction
Tree position measurements are of great significance in forest resource investigations. The rapid and accurate acquisition of tree position in sample plots is helpful for predicting the growth of tree DBH and population development trends and has important ecological significance for revealing the relationship between trees and tree species [1,2]. Traditional measurements use a box compass combined with tapes or a laser distance meter. The measurement process is timeconsuming and laborious, and the data are read and recorded manually, which is inefficient and easily confounded, thus seriously restricting the quality and efficiency of acquiring tree positions in sample field surveys [3][4][5][6][7]. Developing a method of improving the quality and efficiency of tree position information acquisition has long been an issue for forestry researchers [8][9][10].
In recent years, many devices and methods have been developed for locating trees in sample plots. The Haglöf Postex® Laser uses ultrasonic solution for tree position measurement; however, due to the short distance of ultrasonic measurement, the size of the sample plot will be limited [11]. The global navigation satellite system (GNSS) [12][13][14][15] can provide positioning coordinate information in most environments. However, signal attenuation or even signal disappearance occurs in stands with high canopy density, which is also greatly affected by the density of trees in the sample plot [16][17][18]. Total stations are a type of precise electronic survey system with a mapping instrument that integrates rangefinders, electronic theodolites, and microprocessors [19]. Such systems provide an effective method for locating trees on the sample plot but have some disadvantages, such as inconvenience, complicated operation, and expensive price. Terrestrial laser scanning 2. System Design 2.1. Design of the Main Device. The device is mainly composed of a base station and a mobile station, as shown in Figure 1. The base station consists of a supporting tripod and five 1 m long poles with a UWB module on top. The poles and supporting tripod can be folded and shrunk, which facilitates the carrying of the device by survey crews in the field. Mobile station is the location tag. Table 1 describes the components used by the device and their main properties. The circuit of the base station is mainly composed of a microprocessor, power module, gyroscope module, compass module, UWB module, and display screen. The circuit of the mobile station is mainly composed of a microprocessor, power module, UWB module, secure digital memory card (SD card), Bluetooth module, button, and display screen. The device uses an 8-bit microprocessor, which is high speed and low cost and has strong anti-interference. The power module is composed of a lithium battery, a power management chip, and a switch. The lithium battery is used as the supply power to the system components, and the power management chip can protect the lithium battery and other circuits from overcurrent. The gyroscope module is used to collect the location information of the base station. The compass module is used to collect the azimuth information. The base station is equipped with five UWB modules, and the mobile station is equipped with one UWB module to measure the distance between the mobile station and the base station. The display screen displays the measurement information, the SD card stores the distance and location coordinate data collected by the mobile station, and the Bluetooth module transmits the data between the mobile station and host computer.

Principle of UWB Ranging.
To accurately measure the distances, the UWB sensor based on the double-sided twoway ranging (DS-TWR) principle [46,47] was selected. Due to the error introduced by clock deviation in single-sided two-way ranging (SS-TWR) [48,49] measurement, DS-TWR adds another round of communication on the basis of SS-TWR to compensate for the error introduced through two communication times and increase the accuracy of ranging. The principle is shown in Figure 2, and the formula is shown in where Node A and Node B are communication nodes, D is the distance between Node A and Node B, T prop is the time for the wireless pulse signal to travel in the air, c is the speed of light, T round1 is the total time of Node A receiving and sending pulses in the first round of communication, T reply1 is the waiting time of Node B in the first round of communication, T round2 is the total time required for Node B to receive and send pulses in the second round of communication, and T reply2 is the waiting time of Node A in the second round of communication.
2.3. Design of Software Process. The software was designed to acquire the distances between the base station and the mobile  Journal of Sensors   3 Journal of Sensors station and to calculate the positioning coordinates, key control, and data management, as shown in Figure 3. The GUI enables investigators to read device information, calculate data statistics, and export data to facilitate data processing and analysis. The general flow of the software is shown in Figure 4, which is divided into three stages: preparation, measurement, and upload. During the measurement, the tree number and the corresponding position information can be automatically completed. were measured (a circular plot with a radius of 11.28 m is usually used in forest surveys in South China; therefore, we set the radius as 12 m). The surfaces were mainly covered by leaf litter in plots 1 to 6 and denser weeds in plots 7 to 10 ( Table 2).

Measurement Process Design.
To test the positioning accuracy, the reference data measured using theodolite with a laser rangefinder were compared with the data measured by the device. All measurements used the local east, north, up (ENU) coordinate system. The measurement operation for theodolite with a laser rangefinder and the device are shown in Figures 5 and 6, respectively. Poles B and C of the device pointing to the north and east were the positive y-   Journal of Sensors and x-axes, respectively. Pole A was the positive z-axis. The device measurement process was as follows.
(1) The base station was placed in the center of the circle plot. According to the gyroscope data, the station was adjusted to the horizontal position and then pole B was moved to the north according to the compass data (2) The plot ID was set on the mobile station (3) We placed the mobile station at the measuring position of the trees (1.3 m above the ground), pressed the record button, and recorded the position data. To ensure that the measured value is close to the reference value, the survey crew can press the average key for multiple samplings. Finally, the average value is automatically calculated. The actual operation is shown in Figure 7  S3, S4, S5, S10, S13 13.8 10 46 S5, S10, S13 12.1     Figure 8.
Taking OC as the positive x-axis, OB as the positive y -axis, and OA as the positive z-axis, the coordinates of points A-E can be obtained: Assuming that the P n coordinates are ðX n , Y n , Z n Þ, the distance between P n and each anchor point is as follows:  Journal of Sensors In Equations (3)- (7), any four equations can be used to determine the P n coordinates.
As shown in Figure 9(a), in practical applications, the UWB wireless signal will be obstructed and experience interference by the surrounding environment, such as the trunks, branches, and leaves, resulting in signal attenuation [50,51] and slightly longer communication time; therefore, the AP n , BP n , CP n , DP n , and EP n values are slightly greater than the actual values and do not intersect at a common point. To achieve accurate positioning and mitigate the influence of signal attenuation on ranging, some scholars have proposed using trilateration and quadrilateral localization for corresponding positioning depending on the received signal strength indication (RSSI) [52,53].

Journal of Sensors
According to Equations (11), (12), and (13), the bias, root mean square error (RMSE), and straight-line distance error between the estimation point and the reference point on the x-and y-axes were calculated separately to evaluate the accuracy of the method: where d i is the ith measured value, D ir is the ith reference value, X i and Y i are the ith estimates on the x-axis and y -axis, respectively, X ir and Y ir are the ith reference values on the x-axis and y-axis, respectively, and n is the total number of measurements.

Efficiency Evaluation.
To evaluate the efficiency of measurement, the times required for traditional and device measurement were recorded. To ensure that the two methods worked in the same order, the trees were numbered prior to measurement. Using the theodolite with a laser rangefinder measurement requires two people in a group: one stands in the middle of the circle to operate the theodolite with laser rangefinder correctly and then records the distances and angles from the center of the   10 Journal of Sensors circle plot (the angles to due north) and the other holds a surveyor's pole at the tree position. After the measurement is completed, the data recorded on a paper form is entered into the computer. When using the device for measurements, one person performs ranging and recording integration.  Table 3). The mean Ed was 36.13 cm, the maximum Ed was 82.73 cm, the minimum Ed was 1.03 cm, and the standard deviation was 16.67 cm ( Table 4). The Ed range of the estimated points and the reference points in the xoy plane was 0 cm to 83 cm ( Figure 10). Systematic error was observed between the estimated values and the reference values of each sample plot, although the total error was small ( Figure 11). The mean Ed values in the sample plots from 1 to 6 and 7 to 10 were 36.55 cm and 36.78 cm, respectively, indicating that the slope of the measured sample plot had no significant influence on the location.

Efficiency of Position Measurement.
The measurement times of the two methods are shown in Table 5. The results showed that the mean measurement times were 22.37 and 69.76 s using the device and traditional methods, respectively. The measurement efficiency can be increased more than threefold using the proposed method.

Discussion
Tree position measurement is an important task in forest resource surveys. In recent years, a large number of tree location devices and methods have been reported. The Haglöf Postex® Laser, which was similar to the developed device, used an ultrasonic solution for position measurements, but ultrasound had a short measurement range and was not suitable for measurements of large sample plots. In addition, the device was not cheap [11]. Oveland et al. [18] proposed a positioning scheme under relatively rough conditions using a moving terrestrial laser scanner and GNSS. However, due to the serious attenuation of GNSS signals under the crown, this scheme was not suitable for accurate tree positioning; moreover, the devices were also expensive. Gollob et al. [30] located tree trunk positions using a terrestrial laser scanner; however, the data processing was complex and the device was hard to carry and thus was unsuitable for widespread application in forestry. Fan et al. [33]   11 Journal of Sensors cloud processing. Our device only used the RSSI algorithm and pentagonal localization algorithm, which could be run on STC15 (8-bit) microprocessors (price under $1) and had much lower complexity in both time and space.
In this study, we designed a tree positioning device (cost of approximately $250) using UWB sensors. To test the estimation accuracy, we chose 10 circular plots with a diameter of 24 m to carry out our experiment. The results showed that the x-and y-axes had biases of -8.05 to 4.46 cm and -1.22 cm to 9.34 cm, respectively. The RMSEs of the x-and y-axes were 20.79 to 37.33 cm and 21.09 to 34.80 cm, respectively ( Table 3). The mean error (Ed) and standard deviation of the distance between the estimated points and the reference points were 36.13 and 16.67 cm, respectively (Table 4). In terms of measurement efficiency, the traditional method required two people to complete the work and the data need to be recorded manually. However, the proposed device could be operated by one person and equipped with corresponding host computer software that integrated data measurement, recording, and uploading. The mean times required by the traditional method and the proposed device were 69.76 s and 22.37 s, respectively ( Table 5). The device could locate trees accurately, and the measuring efficiency could be improved by more than three times.
Although the proposed device accurately located trees, considerable improvements remain to be made. For example, the maximum distance between the two anchor points was 2 m, which precludes the device's deployment in sample plots with high stand density. Note that using the device to measure the tree position may be affected by the plot environment. Compared with the slope, the stand density was more influential (Figure 12 and Table 2). In future research, we will ensure the positioning precision to reduce the size of the base station and conduct deeper research in a variety of complex sample environments to identify methods that can improve the positioning accuracy and reduce the impact of the plot environment on ranging.

Conclusions
This paper reports a novel device that uses UWB sensor technology to locate trees. It has a delicate mechanical structure and is based on trilateration and quadrilateral localization. A pentagonal localization algorithm was constructed to precisely locate trees. Through the host computer software, it performs data measurement, recording, and uploading integration, which increases the measurement efficiency more than three times compared to the traditional method.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare no conflict of interest.