UFUASRP


Our mission is to research and develop the UAS technologies that enable its use as a scientific tool. As a result, the research addresses a broad but interrelated set of topics that have an impact on the actual deployment and operation of the UAS. In addition to scientific research, the UF UAS Research Group continues to be a leader in developing safe and effective protocols for deploying UASs in the National Airspace in compliance with FAA regulations. The interdisciplinary research team collaborates in the following areas to support the use of the UAS in a variety of applications.

Ecological Monitoring

Mission Planning

Example Flight Plan

Mission planning and in-flight guidance must account for environmental variables such as wind speed and direction, as well as the movement, geometry, and spectral properties of the target. The ability to generate robust flight plans using waypoint-based guidance systems that account for these factors is an active area of research. The collection of geospatial data from unmanned platforms depends on the reliable execution of a flight plan that ensures complete coverage of the target area as well as robust and blunder-free sensor data. Failure to acquire sufficient data in an otherwise adequate preplanned flight can be caused by a multitude of sources; environmental disturbance or other physical interruption of the flight plan, failure of the guidance and navigation systems such as loss of GPS lock, or through sensor payload malfunctions.

The specific objective of this research is to allow the unmanned system operator to designate a target area by generating a simple polygon over a pre-existing geospatial reference. The mission is then dynamically generated and continuously updates a waypoint-based flight plan incorporating data product specifications such as optical sensor resolution and accuracy. Updates from the algorithm will use sensory feedback data from navigation, imaging, and other integrated sensors to detect anomalous behavior and automatically develop corrective maneuvers.

Airframe Optimization

Aerodynamic Modeling

Improving the design of sUAS for biological and environmental studies continues to be a central research topic, with five generations of airframe design over the past decade. Current efforts are geared toward optimizing the efficiency of the data collection process, primarily by improving aerodynamic efficiency at cruise. Improving the efficiency is challenging due to the constraints of cost, simplicity, robustness, and the specialized features required for ecological applications.

The current airframe version is the Mako, a 2.7 m wingspan airframe specifically designed to be operated from unimproved or wetlands areas. To overcome the inherent challenges involved with operating in these areas, the Mako was designed to be hand-launched and water-landable, eliminating the traditional runway or catapult requirements for UAVs in this size class. Building on experience gained from four previous generations of UAVs, the design goals focused on providing a stable amphibious platform that maximized flight endurance and ground coverage per flight. Providing these capabilities is a waterproof fuselage with a high-wing, v-tail configuration designed to protect the wing and tail surfaces from water immersion and impacts with brush upon landing in unimproved areas. The Mako was designed with a bolt-on three-piece wing and removable v-tail to enable compact transportation, with tool-free fasteners allowing for rapid field assembly and breakdown.

Payload Design

The primary mission of the UAS is the collection of high-resolution, directly georeferenced, visible-spectrum imagery. The sensor suite and payload controller developed for this platform has evolved to meet these goals as technology has improved. The following factors have been identified as critical to the design of the payload.

Photogrammetry

A core advantage of small UASs is their ability to survey inaccessible or dangerous territory with minimal deployment effort. From a technological standpoint, this capability is driven by direct georeferencing. This refers to the ability of the payload to use GPS and other navigation systems to remotely locate and accurately map the target without limited prior information about the survey site. This is a central component in the broader objective of an autonomic data collection system to complete missions without human intervention. Traditional remote sensing methods require ground control, or pre-identified points with known locations, that will be included in the survey. The ground control necessity eliminates the advantages of direct georeferencing.

The clear advantages of direct georeferencing are accompanied by the limitation that little or no confirmation of the accuracy of the remotely sensed data is available. That is, each dataset can report an estimate of the precision of the map generated, but that estimate is based solely on the data collected by the UAV itself. This implies that errors inherent to the UAVs measurement system are not easily detectable, which has important implications where the spatial measurements derived from the data set are used for engineering projects, or where the USACE is liable for the accuracy of the information. As a result, a central goal of the autonomic algorithm is to evaluate whether the data that is being collected will produce the desired output product.

The precision and accuracy of the sensor system must be fully characterized to provide a baseline against which the knowledge system constructed by the intermediate control layer can be evaluated. The accuracy of directly georeferenced products is dependent on the accuracy and reliability of the sensors and processing methods used to generate them. Careful assessment of the measurement system can provide confidence in the accuracy of the output product even when independent verification is not available. Understanding the errors and error bounds of the measurement system is imperative to the adoption of this technology for the USACE.

Indian Prairie Canal/Fisheating Bay

5-cm pre-treatment orthomosaic and post-treatment orthomosaic acquired 6 weeks later

As part of an ongoing USACE invasive aquatic species management program on Lake Okeechobee, numerous mapping flights were made pre- and post-treatment to assess the effectiveness of their species-targeted herbicide treatments. Target species include luziola (Luziola subintegra), water hyacinth (Eichhornia crassipes), and water lettuce (Pistia stratiotes). Continued monitoring flights are ongoing to evaluate long term treatment effects on native plant communities.

In addition to developing operational procedures to effectively monitor environmental changes over time, the Lake Okeechobee missions have also proven valuable in establishing real-world timelines for acquiring and processing UAS-derived geospatial products. This information is now used by UF and USACE project managers to more efficiently plan monitoring programs with realistic estimates instead of lofty academic projections.

Finally, supervised classification techniques were performed on the imagery sets to automatically identify and quantify relative abundances of invasive and indigenous plant species.

Classified pre-treatment orthomosaic

Loxahatchee

2.5-cm orthomosaic of bird islands in Loxahatchee NWR

Managed by the US Fish and Wildlife Service, the Arthur R. Marshall Loxahatchee National Wildlife Refuge is the home to numerous colonies of wading birds and other wildlife. The research focus in this area is on mapping the geographic extent of individual bird islands, automatically determining colony populations, and precisely locating and monitoring individual nesting sites during and across nesting seasons. The high positional accuracy and optical resolution afforded by the Mako sUAS allows biologists to not only identify multiple species from aerial images, but to count individual eggs within each nest site.

L-6 Levee

Google Earth plot showing 2 executed flight paths along the levee path

The USACE-administered L-6 levee project in South Florida was flown as an investigation into optimal flight planning for mapping long, linear features as would be typical for many infrastructure monitoring applications. During these flights, trials were conducted determining flight timing to minimize surface glare on open water, evaluating the effectiveness of a mobile ground station for continuous operation within FAA-mandated visual line-of-sight minimums, and determining minimum required optical resolution/flight altitudes for accurate infrastructure condition assessment.

Picayune Strand

(Clockwise from top left) Road degradation, pump station construction, and invasive vegetation imagery collected over Picayune Strand

Part of the Comprehensive Everglades Restoration Plan (CERP), Picayune Strand is the site of USACE efforts to restore sheet flow to an area previously drained by canals and prepared for housing development in the 1960s. In anticipation of the restored flow and the subsequent change in habitat type, the Corps is also engaged in removing invasive and historically non-native species, including Brazilian pepper (Schinus terebinthifolius) and cabbage palm (Sabal palmetto).

Monitoring flights are engaged in pre-, during, and post-degradation assessments of sample vegetation plots as well as along roads and canals as they are degraded to their historical state. Due to the variety of ground targets and applications for aerial imagery, Picayune Strand has been an ideal proving ground for data collection assessments including geometric resolution, color infrared (CIR) vegetation classification, and flight path efficiency evaluations.

Additional flights were undertaken to monitor the ongoing status of road degradation work as well as construction on a flow control pump station.

Lightning Lab

120m x 60m grid of aerial calibration targets

To assist in sensor and payload calibration and benchmarking, a precisely surveyed grid of 32 aerial targets was established at the International Center for Lightning Research and Testing (ICLRT) at Camp Blanding, Florida. This site is ideal for evaluating potential and existing payloads and provides a very accurate assessment of both optical (geometric) and navigational (position and orientation) fidelity for determining appropriate system parameters for photogrammetric post-processing.

Herbert Hoover Dike

3D terrain model of levee with projected mosaic

The 143-mile-long USACE-administered Herbert Hoover dike around Lake Okeechobee has been an ideal location for developing effective flight planning methods to provide accurate three-dimensional topographic models of linear terrain features. Similar to work done on the L-6 levee, the Herbert Hoover Dike monitoring flights were useful in evaluating stereographic flight planning techniques and determining optical resolution requirements for infrastructure monitoring. The program is currently completing a mission to detect/measure physical characteristics indicative of leaks, seeps, and boils in dikes and levees.

Levee monitoring is currently supported by visual inspections performed by field personnel. The UAV will augment these efforts by providing a permanent record of the levee's condition in unprecedented resolution. The data can also be photogrammetrically processed to produce a 3D geospatial model of the levee and seamless orthomosaic views. This comprehensive data set directly impacts the decision making process by providing crucial information that can be continuously evaluated as conditions change. New analysis can be performed on demand, above and beyond the work performed by field personnel during the initial inspection. Utilizing this imagery, the Corps of Engineers can increase the effectiveness of its monitoring efforts. These deliverables will prove to be an important contribution to the long-term infrastructure monitoring efforts performed throughout the Corps.

Primary Navigation