DRONES Funded Research 2018
GIS Mapping with Drones
Dr. Esra Ozdenerol
The UAV industry is actively growing and the addition of geospatial knowledge will only help those looking to find work deploying drones. Today anyone can fly a drone and collect some pretty stunning video or even still photography. That data can be very useful for activities such as inspections, construction, earthworks, crop scouting and marketing. However, if the goal is to quantify a landscape condition (e.g., Urban 3D models, GIS and visualization, cadastral and geodetic surveys, appraisals and asset inventory mapping, train mapping, agriculture, environmental modelling), it is necessary to move that data into a Geographic Information System (GIS). Everyone wants to fly drones and they have seen what is possible, but they lack the necessary geospatial knowledge to do it for themselves or analyze the data properly. Additionally, because simpler tools are now available, many folks are now able to process the data but don't really understand what they have and/or don't know how to extract the necessary information. Dr. Ozdenerol wants to offer a course/ training and workforce development opportunity, through DRONES Research Cluster, titled "GIS Mapping with Drones" designed to fill the existing knowledge gap by teaching students, researchers and professionals the entire workflow – how to collect data properly (and legally), process the data correctly and analyze the data spatially.
Real-time Interactions and Navigation of Autonomous Vehicles for Optimized Unmanned Package Delivery
Dr. Junaid Ahmed Khan, Dr. Lan Wang
Traditional manned last mile delivery of packages and letters in a campus like environment or an urban neighborhood with multiple recipients in nearby buildings is a time consuming and costly operation. Once a delivery truck is in the neighborhood, the delivery person may need to contact another person in a building in order to deliver the package to its desired location which can be either at the entrance or inside the building. In case of absence of a person at the entrance, the delivery person needs to gain physical access to the building and obtain necessary guidance such as floor plan and directions to efficiently deliver the package at the desired recipient in a timely manner. Today, with the advancement in autonomous navigation technology for vehicles, last mile delivery to multiple recipients in the same or multiple nearby buildings can be improved where an autonomous vehicle (AV) can replace the human and complete the deliveries in an optimized unmanned manner once the truck arrive in the neighborhood. In this proposed project, Dr. Khan and Dr. Wang focus on the AV's navigation to the buildings and will address the robots' indoor navigation in a subsequent project. They will address the following challenges: (i) how does the AV detect and interact with pedestrians while navigating towards the destination location. (ii) how to design an optimal AV trajectory such that the AV takes a route that minimizes its energy consumption as well as the time and distance between different destination locations? (iii) how does an AV communicate with the building/infrastructure in order to physically access the premises.
Multi-Sensory UAV Approach to Stream Assessments: Sense and Avoidance Development
Dr. Brian Waldron, Scott Schoefernacker
A visual stream assessment (VSA) is part of a Municipal Separate Storm Sewer System (MS4) permit conducted every 5 years collecting information regarding the stream and riparian zone's overall health as well as documenting individual impairments including outfalls, erosion, exposed pipes, trash, fish barriers, etc. Traditional VSA surveys include sending teams into the stream to document impairments using paper forms and digital photographs which must be transcribed and entered into a database at a later time. In 2013, the Center for Applied Earth Science and Engineering Research (CAESER) conducted a 160 mile VSA for Shelby County Public Works. The project employed 15 students to assess streams and collect data digitally on a tablet-based custom application and synced to a geo-database. The effort took four months to complete over the winter months. Previous funding from the FedEx Institute of Technology DRONES (Drones, Robotics, and Navigation Enabled Systems) Innovation Research Cluster enabled the development of two Quadcopter UAVs (V1 and V2) by EDP Robotics. The innovative design of the V1 and V2 consist of 3D printed modular parts that are easily replaced or modified; foldable motor arms; and internal removable trays to ease calibration and maintenance. Graduate and undergraduate students of the DRONES Research Lab developed the current sense and avoidance (S&A) system using an open-source operating system, UP board processor, scripting, and commercially available low-cost sensors (Intel RealSense). Sensors are mounted on the bow, port and starboard of the drone. Waypoints are established pre-flight using Mission Planner and executed using the onboard processor and GPS. When obstacles are encountered the S&A goes through a protocol to sense a clear path around the obstacle without changing elevation. Once around the obstacle, the drone flies to the next predetermined waypoint