Stephen Haviland

Electric Multirotor UAV Propulsion System Sizing for Performance Prediction and Design Optimization

One of the more daunting tasks of designing a multirotor unmanned aerial vehicle (UAV) is the selection of a propulsion system that will provide desired performance. Rigorous methods for selecting these drive components, that is, the motors, propellers, and batteries for electric UAVs are not readily available. Currently, many UAV designs are based on legacy selections or limited and at times incorrect manufacturer data. These design methods are either simplistic or lacking in analysis and validation of component selection. Proper propulsion system design should address the mission requirements for which the vehicle is being designed. A proper design methodology is the best chance that the designer has to create a new vehicle that will be mission-capable. This paper attempts to satisfy the need for more thorough method of propulsion component selection. The paper is written also to document the popular online drive system analysis tool due to numerous requests. This tool is one example implementation of the methodologies described by this paper.

Vision-based closed-loop tracking using micro air vehicles

This paper describes the target detection and tracking architecture used by the Georgia Tech Aerial Robotics team for the American Helicopter Society (AHS) Micro Aerial Vehicle (MAV) challenge. The vision system described enables vision-aided navigation with additional abilities such as target detection and tracking all performed onboard the vehicles computer. The author suggests a robust target tracking method that does not solely depend on the image obtained from a camera, but also utilizes the other sensor outputs and runs a target location estimator. The machine learning based target identification method uses Haar-like classifiers to extract the target candidate points. The raw measurements are plugged into multiple Extended Kalman Filters (EKFs). The statistical test (Z-test) is used to bound the measurement, and solve the corresponding problem. Using Multiple EKFs allows us not only to optimally estimate the target location, but also to use the information as one of the criteria to evaluate the tracking performance. The MAV utilizes performance-based criteria that determine whether or not to initiate a maneuver such as hover or land over/on the target. The performance criteria are closed in the loop which allows the system to determine at any time whether or not to continue with the maneuver. For Vision-aided Inertial Navigation System (VINS), a corner Harris algorithm finds the feature points, and we track them using the statistical knowledge. The feature point locations are integrated in Bierman Thornton extended Kalman Filter (BTEKF) with Inertial Measurement Unit (IMU) and sonar sensor outputs to generate vehicle states: position, velocity, attitude, accelerometer and gyroscope biases. A 6-degrees-of-freedom quadrotor flight simulator is developed to test the suggested method. This paper provides the simulation results of the vision-based maneuvers: hovering over the target, and landing on the target. In addition to the simulation results, flight tests have been conducted to show and validate the system performance. The 500 gram Georgia Tech Quadrotor (GtQ)-Mini, was used for the flight tests. All processing is done onboard the vehicle and it is able to operate without human interaction. Both of the simulation and flight test results show the effectiveness of the suggested method. This system and vehicle were used for the AHS 2015 MAV Student Challenge where the GPS-denied closed-loop target search is required. The vehicle successfully found the ground target, and landed on the desired location. This paper shares the data obtained from the competition.

Dynamic Modeling and Analysis of a VTOL Freewing Concept

The freewing aircraft is an airplane with the wing placed on a hinged bearing about the pitch axis that allows for free rotation with respect to the fuselage body. The freewing only differs from a conventional airplane in that the wing is free to rotate. The 70 year old concept, through various studies, has been shown that by placing the wing on a freeto-rotate hinge, it effectively reduces the inertia of the wing making it easier for the wing to adjust to turbulence. The work described in this paper makes two large contributions to past freewing work. The first is the motors are placed on the wing which allows for vertical takeoff and landing (VTOL). The second is a different dynamic analysis approach that uses multi-body dynamics and uses hinge constraints as part of the state vector which allow for an easier way of solving the equations of motion. Data is shown for how the longitudinal modes change when varying design parameters. Simulation data is also presented highlighting the VTOL aspect of the vehicle.

Particle filter for closed-loop detection and tracking from occluded airborne images

This paper describes a novel vision-based target tracking and landing method that uses aerial images from an on-board camera. The proposed method explicitly deals with occlusions that often occur during these maneuvers. Normalized cross correlation (NCC) is used to locate an image patch in a reference image with a measure of certainty. The key insight is that over the course of the vehicle approach, there is a transition between the target being contained in the camera images, and camera images being contained in the target image. When a vehicle is at high altitude, the NCC of the target over an entire camera image is computed. When at low altitude, the reverse operation is performed: the NCC of the camera images is computed over the target image. Additionally, at both high and low altitude, we find interesting region using contour trees, and the NCC of the template with the region is calculated. This way, we can recognize a target even when it is only partly in view. A particle filter is used to fuse highly multi-modal measurements from the three techniques. Each particle chooses its update measurement using a roulette wheel selection with the size of the slice being proportional to the measurements NCC and, therefore, converges to a location that has a greater NCC and numerous positive hits. The particle filter allows estimation of target position and velocity states, which are used to determine criteria for safe landing. We evaluate our system with an image-in-the-loop simulation and closed-loop flight tests with a quadrotor.

Collaborative Autonomy for Mapping, Search, and Pursuit

This paper describes the development and flight tests of a multi-aircraft collaborative architecture, focused on decentralized autonomous decision making to solve a scenario-driven challenge problem. The architecture includes two search coverage algorithms, a hide location detection technique, behavior estimation, and a target manipulation algorithm. The architecture was implemented on a pair of Yamaha RMAX helicopters outfitted with modular avionics, as well as an associated set of simulation tools. Simulation and flight test results for single- and multiple aircraft scenarios are presented. Further work suggested includes identification and development of more sophisticated evader models and pursuit algorithms.

Light-weight quadrotor with on-board absolute vision-aided navigation

This paper presents a novel small unmanned aerial system which can operate independently and navigate using visual sensors. The basis of the system is a 600 gram quadrotor platform with sonar, inertial sensors and a camera. All processing is performed onboard the vehicle on a lightweight and powerful computer. Vehicle navigation is provided by a two part system. A Bierman Thornton extended Kalman filter (BTEKF) generates vehicle state and image feature estimates in a visual simultaneous localization and mapping (SLAM) formulation. Absolute position updates are provided to the BTEKF by a particle filter map-matching technique. Simulation and flight test results demonstrate the capabilities of the system to both navigate in an unknown environment with minimal drift while also showing the ability to eliminate drift when a priori known visual information is visible.