Explore the vast range of titles available.

Activity recognition modelling using smartphone Inertial Measurement Units (IMUs) is an underutilized resource defining and assessing work efficiency for a wide range of natural resource management tasks. This study focused on the initial development and validation of a smartphone-based activity recognition system for excavator-based mastication equipment working in Ponderosa pine ( Pinus ponderosa ) plantations in North Idaho, USA. During mastication treatments, sensor data from smartphone gyroscopes, accelerometers, and sound pressure meters (decibel meters) were collected at three sampling frequencies (10, 20, and 50 hertz (Hz)). These data were then separated into 9 time domain features using 4 sliding window widths (1, 5, 7.5 and 10 seconds) and two levels of window overlap (50% and 90%). Random forest machine learning algorithms were trained and evaluated for 40 combinations of model parameters to determine the best combination of parameters. 5 work elements ( masticate , clear , move , travel , and delay ) were classified with the performance metrics for individual elements of the best model (50 Hz, 10 second window, 90% window overlap) falling within the following ranges: area under the curve (AUC) (95.0% - 99.9%); sensitivity (74.9% - 95.6%); specificity (90.8% - 99.9%); precision (81.1% - 98.3%); F1-score (81.9% - 96.9%); balanced accuracy (87.4% - 97.7%). Smartphone sensors effectively characterized individual work elements of mechanical fuel treatments. This study is the first example of developing a smartphone-based activity recognition model for ground-based forest equipment. The continued development and dissemination of smartphone-based activity recognition models may assist land managers and operators with ubiquitous, manufacturer-independent systems for continuous and automated time study and production analysis for mechanized forest operations.

Real-time positioning on mobile devices using global navigation satellite system (GNSS) technology paired with radio frequency (RF) transmission (GNSS-RF) may help to improve safety on logging operations by increasing situational awareness. However, GNSS positional accuracy for ground workers in motion may be reduced by multipath error, satellite signal obstruction, or other factors. Radio propagation of GNSS locations may also be impacted due to line-of-sight (LOS) obstruction in remote, forested areas. The objective of this study was to characterize the effects of forest stand characteristics, topography, and other LOS obstructions on the GNSS accuracy and radio signal propagation quality of multiple Raveon Atlas PT GNSS-RF transponders functioning as a network in a range of forest conditions. Because most previous research with GNSS in forestry has focused on stationary units, we chose to analyze units in motion by evaluating the time-to-signal accuracy of geofence crossings in 21 randomly-selected stands on the University of Idaho Experimental Forest. Specifically, we studied the effects of forest stand characteristics, topography, and LOS obstructions on (1) the odds of missed GNSS-RF signals, (2) the root mean squared error (RMSE) of Atlas PTs, and (3) the time-to-signal accuracy of safety geofence crossings in forested environments. Mixed-effects models used to analyze the data showed that stand characteristics, topography, and obstructions in the LOS affected the odds of missed radio signals while stand variables alone affected RMSE. Both stand characteristics and topography affected the accuracy of geofence alerts.

Real-time data- and location-sharing using mesh networking radios paired with smartphones may improve situational awareness and safety in remote environments lacking communications infrastructure. Despite being increasingly used for wildland fire and public safety applications, there has been little formal evaluation of the network connectivity of these devices. The objectives of this study were to 1) characterize the connectivity of mesh networks in variable forest and topographic conditions; 2) evaluate the abilities of lidar and satellite remote sensing data to predict connectivity; and 3) assess the relative importance of the predictive metrics. A large field experiment was conducted to test the connectivity of a network of one mobile and five stationary goTenna Pro mesh radios on 24 Public Land Survey System sections approximately 260 ha in area in northern Idaho. Dirichlet regression was used to predict connectivity using 1) both lidar- and satellite-derived metrics (LIDSAT); 2) lidar-derived metrics only (LID); and 3) satellite-derived metrics only (SAT). On average the full network was connected only 32.6% of the time (range: 0% to 90.5%) and the mobile goTenna was disconnected from all other devices 18.2% of the time (range: 0% to 44.5%). RMSE for the six connectivity levels ranged from 0.101 to 0.314 for the LIDSAT model, from 0.103 to 0.310 for the LID model, and from 0.121 to 0.313 for the SAT model. Vegetation-related metrics affected connectivity more than topography. Developed models may be used to predict the connectivity of real-time mesh networks over large spatial extents using remote sensing data in order to forecast how well similar networks are expected to perform for wildland firefighting, forestry, and public safety applications. However, safety professionals should be aware of the impacts of vegetation on connectivity.

Analysis of high-resolution inertial sensor and global navigation satellite system (GNSS) data collected by mobile and wearable devices is a relatively new methodology in forestry and safety research that provides opportunities for modeling work activities in greater detail than traditional time study analysis. The objective of this study was to evaluate whether smartwatch-based activity recognition models could quantify the activities of rigging crew workers setting and disconnecting log chokers on cable logging operations. Four productive cycle elements ( travel to log , set choker , travel away , clear ) were timed for choker setters and four productive cycle elements ( travel to log , unhook , travel away , clear ) were timed for chasers working at five logging sites in North Idaho. Each worker wore a smartwatch that recorded accelerometer data at 25 Hz. Random forest machine learning was used to develop predictive models that classified the different cycle elements based on features extracted from the smartwatch acceleration data using 15 sliding window sizes (1 to 15 s) and five window overlap levels (0%, 25%, 50%, 75%, and 90%). Models were compared using multiclass area under the Receiver Operating Characteristic (ROC) curve, or AUC. The best choker setter model was created using a 3-s window with 90% overlap and had sensitivity values ranging from 76.95% to 83.59% and precision values ranging from 41.42% to 97.08%. The best chaser model was created using a 1-s window with 90% overlap and had sensitivity values ranging from 71.95% to 82.75% and precision values ranging from 14.74% to 99.16%. These results have demonstrated the feasibility of quantifying forestry work activities using smartwatch-based activity recognition models, a basic step needed to develop real-time safety notifications associated with high-risk job functions and to advance subsequent, comparative analysis of health and safety metrics across stand, site, and work conditions.

Fuel reduction in forests is a high management priority in the western United States and mechanical mastication treatments are implemented common to achieve that goal. However, quantifying post-treatment fuel loading for use in fire behavior modeling to forecast treatment effectiveness is difficult due to the high cost and labor requirements of field sampling methods and high variability in resultant fuel loading within stands after treatment. We evaluated whether pre-treatment LiDAR-derived stand forest characteristics at 20 m × 20 m resolution could be used to predict post-treatment surface fuel loading following mastication. Plot-based destructive sampling was performed immediately following mastication at three stands in the Nez Perce Clearwater National Forest, Idaho, USA, to correlate post-treatment surface fuel loads and characteristics with pre-treatment LiDAR-derived forest metrics, specifically trees per hectare (TPH) and stand density index (SDI). Surface fuel loads measured in the stand post-treatment were consistent with those reported in previous studies. A significant relationship was found between the pre-treatment SDI and total resultant fuel loading (p = 0.0477), though not between TPH and fuel loading (p = 0.0527). SDI may more accurately predict post-treatment fuel loads by accounting for both tree number per unit area and stem size, while trees per hectare alone does not account for variations of tree size and subsequent volume within a stand. Relatively large root-mean-square errors associated with the random forest models for SDI (36%) and TPH (46%) suggest that increased sampling intensity and modified methods that better account for fine spatial variability in fuels resulting from within-stand conditions, treatment prescriptions and machine operators may be needed. Use of LiDAR to predict fuel loading after mastication is a useful approach for managers to understand the efficacy of fuel reduction treatments by providing information that may be helpful for determining areas where treatments can be most beneficial.

Logging is one of the most hazardous occupations in the United States. Real-time positioning that uses global navigation satellite system (GNSS) technology paired with radio frequency transmission (GNSS-RF) has the potential to reduce fatal and non-fatal accidents on logging operations through the use of geofences that define safe work areas. Until recently, most geofences have been static boundaries. The aim of this study was to evaluate factors affecting mobile geofence accuracy in order to determine whether virtual safety zones around moving ground workers or equipment are a viable option for improving situational awareness on active timber sales. We evaluated the effects of walking pace, transmission interval, geofence radius, and intersection angle on geofence alert delay using a replicated field experiment. Simulation was then used to validate field results and calculate the proportion of GNSS error bearings resulting in early alerts. The interaction of geofence radius and intersection angle affected safety geofence alert delay in the field experiment. The most inaccurate alerts were negative, representing early warning. The magnitude of this effect was largest at the greatest intersection angles. Simulation analysis supported these field results and also showed that larger GNSS error corresponded to greater variability in alert delay. Increasing intersection angle resulted in a larger proportion of directional GNSS error that triggered incorrect, early warnings. Because the accuracy of geofence alerts varied greatly depending on GNSS error and angle of approach, geofencing for occupational safety is most appropriate for general situational awareness unless real-time correction methods to improve accuracy or higher quality GNSS-RF transponders are used.

Logging continues to rank among the most lethal occupations in the United States. Though the hazards associated with fatalities are well-documented and safe distances from hazards is a common theme in safety education, positional relationships between workers and hazards have not been quantified previously. Using GNSS-RF (Global Navigation Satellite System-Radio Frequency) transponders that allow real-time monitoring of personnel, we collected positioning data for rigging crew workers and three common cable logging hazards: a log loader, skyline carriage, and snag. We summarized distances between all ground workers and each hazard on three active operations and estimated the proportion of time crew occupied higher-risk areas, as represented by geofences. We then assessed the extent to which positioning error associated with different stand conditions affected perceived worker safety status by applying error sampled in a separate, controlled field experiment to the operational data. Root mean squared error was estimated at 11.08 m in mature stands and 3.37 m in clearcuts. Simulated error expected for mature stands altered safety status in six of nine treatment combinations, whereas error expected for clearcuts affected only one. Our results show that canopy-associated GNSS error affects real-time geofence safety applications when using single-constellation American Global Positioning System transponders.

In this paper, we provide an overview of positioning systems for moving resources in forest and fire management and review the related literature. Emphasis is placed on the accuracy and range of different localization and location-sharing methods, particularly in forested environments and in the absence of conventional cellular or internet connectivity. We then conduct a second review of literature and concepts related to several emerging, broad themes in data science, including the terms location-based services (LBS), geofences, wearable technology, activity recognition, mesh networking, the Internet of Things (IoT), and big data. Our objective in this second review is to inform how these broader concepts, with implications for networking and analytics, may help to advance natural resource management and science in the future. Based on methods, themes, and concepts that arose in our systematic reviews, we then augmented the paper with additional literature from wildlife and fisheries management, as well as concepts from video object detection, relative positioning, and inventory-tracking that are also used as forms of localization. Based on our reviews of positioning technologies and emerging data science themes, we present a hierarchical model for collecting and sharing data in forest and fire management, and more broadly in the field of natural resources. The model reflects tradeoffs in range and bandwidth when recording, processing, and communicating large quantities of data in time and space to support resource management, science, and public safety in remote areas. In the hierarchical approach, wearable devices and other sensors typically transmit data at short distances using Bluetooth, Bluetooth Low Energy (BLE), or ANT wireless, and smartphones and tablets serve as intermediate data collection and processing hubs for information that can be subsequently transmitted using radio networking systems or satellite communication. Data with greater spatial and temporal complexity is typically processed incrementally at lower tiers, then fused and summarized at higher levels of incident command or resource management. Lastly, we outline several priority areas for future research to advance big data analytics in natural resources.

Real-time data- and location-sharing using mesh networking radios paired with smartphones may improve situational awareness and safety in remote environments lacking communications infrastructure. Despite being increasingly used for wildland fire and public safety applications, there has been little formal evaluation of the network connectivity of these devices. The objectives of this study were to 1) characterize the connectivity of mesh networks in variable forest and topographic conditions; 2) evaluate the abilities of lidar and satellite remote sensing data to predict connectivity; and 3) assess the relative importance of the predictive metrics. A large field experiment was conducted to test the connectivity of a network of one mobile and five stationary goTenna Pro mesh radios on 24 Public Land Survey System sections approximately 260 ha in area in northern Idaho. Dirichlet regression was used to predict connectivity using 1) both lidar- and satellite-derived metrics (LIDSAT); 2) lidar-derived metrics only (LID); and 3) satellite-derived metrics only (SAT). On average the full network was connected only 32.6% of the time (range: 0% to 90.5%) and the mobile goTenna was disconnected from all other devices 18.2% of the time (range: 0% to 44.5%). RMSE for the six connectivity levels ranged from 0.101 to 0.314 for the LIDSAT model, from 0.103 to 0.310 for the LID model, and from 0.121 to 0.313 for the SAT model. Vegetation-related metrics affected connectivity more than topography. Developed models may be used to predict the connectivity of real-time mesh networks over large spatial extents using remote sensing data in order to forecast how well similar networks are expected to perform for wildland firefighting, forestry, and public safety applications. However, safety professionals should be aware of the impacts of vegetation on connectivity.