Giovanni Fusina

FPGA Implementation of Genetic Algorithm for UAV Real-Time Path Planning

The main objective of an Unmanned-Aerial-Vehicle (UAV) is to provide an operator with services from its payload. Currently, to get these UAV services, one extra human operator is required to navigate the UAV. Many techniques have been investigated to increase UAV navigation autonomy at the Path Planning level. The most challenging aspect of this task is the re-planning requirement, which comes from the fact that UAVs are called upon to fly in unknown environments. One technique that out performs the others in path planning is the Genetic Algorithm (GA) method because of its capacity to explore the solution space while preserving the best solutions already found. However, because the GA tends to be slow due to its iterative process that involves many candidate solutions, the approach has not been actively pursued for real time systems. This paper presents the research that we have done to improve the GA computation time in order to obtain a path planning generator that can recompile a path in real-time, as unforeseen events are met by the UAV. The paper details how we achieved parallelism with a Field Programmable Gate Array (FPGA) implementation of the GA. Our FPGA implementation not only results in an excellent autonomous path planner, but it also provides the design foundations of a hardware chip that could be added to an UAV platform.

Modeling of Urban Wind Field Effects on Unmanned Rotorcraft Flight

Surveillance applications of unmanned aerial vehicles within urban areas is made difficult by turbulent winds generated by buildings. A methodology is proposed by which urban wind data is evaluated, selected, and applied during flight simulation. The urban environment is represented by a combination of discrete single buildings and canyons each easily amenable to computational fluid dynamics. These simulations are used to generate a database of building wakes for various wind conditions and building configurations typical to the North American urban environment. A selection algorithm is used that determines if the current aircraft position is influenced by the building wake and, if so, calculates the resulting effect on the aircraft. Results are presented for a Yamaha R-50 R/C helicopter in both forward and vertical flight. Variations in aircraft attitude by as much as 10 are observed when considering building wake effects generated by a 8 kt wind. It is also demonstrated that this approach is able to identify locations within a given wake at which aircraft flight is most affected.

Unmanned Aerial Vehicle formation flying using Linear Model Predictive Control

A team of three Unmanned Aerial Vehicles (UAVs) accomplishes a line abreast, triangular and cross formation based on high-level Linear Model Predictive Control (LMPC). All flight tests respect Reynolds rules of flocking, where the UAVs avoid collisions with nearby flockmates, attempt to match velocity of other team members and attempt to stay close to other flockmates. A linear system identification model is at the base of the error dynamics describing the formation control algorithm. The main contribution of this paper lies in the use of LMPC to implement multiple formations on UAVs in simulation and using the Qball-X4 quadrotor.

Urban Wake-Field Generation Using Large-Eddy Simulation for Application to Quadrotor Flight

A method is presented for using large-eddy simulation to generate urban wake fields for use in studying the effects on the autonomous flight performance of a small quadrotor. The flowfield is solved around a single square building using OpenFOAM and stored in a database accessed by a MATLAB/Simulink flight simulator. Four flight missions are evaluated to compare the difference in performance between wake fields generated by Reynolds-averaged Navier–Stokes and large-eddy simulation solutions. The results of the holding position in a constant freestream wind show both methods produce similar results and can hold position in all three directions within approximately ±1.5 body lengths. When the quadrotor is in or on the boundary of the building wake, the maximum deviation volumes, as calculated when using a Reynolds-averaged Navier–Stokes or large-eddy simulation air wake, can differ by two orders of magnitude. Additionally, the large-eddy simulation air wakes can cause skewed deviations by as much as five to...

Validation of swarms of robots: theory and experimental results

In this paper we present a formulation of a swarm algorithm for some predefined formations and we show that using the control laws derived, the whole system is stable. Furthermore, the algorithm for the formation is implemented in microcontrollers, which effectively makes the experiment a system of systems. The only input for the algorithm is the position of the robot and the relative positions of neighbours within two metres of the robot. The results of experiments presented show that the algorithm presents the same general behaviour as predicted in simulations. However, due to noise in measurements and commands, the behaviour is slightly different.

Tasking system capabilities modeling using the complexity profile

The present defence and security environment has been increasingly recognized to be represented by complex systems. The complexity profile is a method for characterizing the complexity of a system. According to a modified version of Ashbys Law of Requisite Variety, a system designed to perform a complex task should have a complexity profile that matches the complexity profile of the task. This modified version of Ashbys Law is called the Law of Requisite Multiscale Variety (Bar-Yam, “Multiscale Complexity/Entropy”, Advances in Complex Systems, vol. 7, 2004). Thus far there have not been quantitative characterizations of real systems in the scientific literature that illustrate the application of this law. In this paper we illustrate how the Law of Requisite Multiscale Variety can be applied to a simple robot tasking problem so that the viability of the task the robots undertake can be determined solely by complexity arguments, and not by extensive simulation or live experimentation. The system was chosen for its simplicity in illustration of the method and its generality of representing a broad class of problems.