Coupled modeling using PRZM/RICEWQ and SWAT for the North Tiaoxi Watershed.

Watershed exposure caused by the use of pesticide in farmland has become a major environmental concern. Currently, there are two major approaches to quantify the watershed exposure: monitoring and modeling. Watershed monitoring is expensive, and short-term monitoring is difficult to be used to address potential long-term exposure variability. Model simulation is widely used because not only can it save time and efforts, but it can also simulate the environmental transport process of pesticide over a long time frame to better understand temporal variability. Research on application of commonly used pesticide exposure assessment models such as PRZM, RICEWQ on watershed scale has found that those models need to be coupled together with waterbody models to assess pesticide exposure at the watershed level, and they are applied on a single crop in targeted area within a watershed, failing to consider the diversity of regional and watershed cropping conditions. To address pesticide exposure assessment in different waterbodies after application on multiple crops within a watershed, this study coupled PRZM, RICEWQ, and SWAT models simultaneously in North Tiaoxi watershed. PRZM model and RICEWQ model were used to simulate the exposure of pesticides in dryland and rice paddies separately, and the pesticide masses through runoff, overflow, spray drift, and other routes simulated by the above two models were set as the input of SWAT model which could simulate hydrology and pollutant transport at watershed scale. Pesticide use, cropping, hydrology, and watershed data were collected, and parameterized for exposure modeling of carbaryl in the North Tiaoxi River after uses on orchard, corn, and rice within the watershed. Model predictions showed high degree of agreement between the simulated results and the field monitoring data. The coupled PRZM, RICEWQ, and SWAT model could simulate reasonably well pesticide exposures in waterbodies with applications on multiple crops within a watershed.

Fast Signals of Opportunity Fingerprint Database Maintenance with Autonomous Unmanned Ground Vehicle for Indoor Positioning.

Indoor positioning technology based on Received Signal Strength Indicator (RSSI) fingerprints is a potential navigation solution, which has the advantages of simple implementation, low cost and high precision. However, as the radio frequency signals can be easily affected by the environmental change during its transmission, it is quite necessary to build location fingerprint database in advance and update it frequently, thereby guaranteeing the positioning accuracy. At present, the fingerprint database building methods mainly include point collection and line acquisition, both of which are usually labor-intensive and time consuming, especially in a large map area. This paper proposes a fast and efficient location fingerprint database construction and updating method based on a self-developed Unmanned Ground Vehicle (UGV) platform NAVIS, called Automatic Robot Line Collection. A smartphone was installed on NAVIS for collecting indoor Received Signal Strength Indicator (RSSI) fingerprints of Signals of Opportunity (SOP), such as Bluetooth and Wi-Fi. Meanwhile, indoor map was created by 2D LiDAR-based Simultaneous Localization and Mapping (SLAM) technology. The UGV automatically traverse the unknown indoor environment due to a pre-designed full-coverage path planning algorithm. Then, SOP sensors collect location fingerprints and generates grid map during the process of environment-traversing. Finally, location fingerprint database is built or updated by Kriging interpolation. Field tests were carried out to verify the effectiveness and efficiency of our proposed method. The results showed that, compared with the traditional point collection and line collection schemes, the root mean square error of the fingerprinting-based positioning results were reduced by 35.9% and 25.0% in static tests and 30.0% and 21.3% respectively in dynamic tests. Moreover, our UGV can traverse the indoor environment autonomously without human-labor on data acquisition, the efficiency of the automatic robot line collection scheme is 2.65 times and 1.72 times that of the traditional point collection and the traditional line acquisition, respectively.