Optimal control for multicopters is difficult in part due to the low processing power available, and the instability inherent to multicopters. Deep imitation learning is a method for approximating an expert control policy with a neural network, and has the potential of improving control for multicopters. We investigate the performance and reliability of deep imitation learning with trajectory optimization as the expert policy by first defining a dynamics model for multicopters and applying a trajectory optimization algorithm to it. Our investigation shows that network architecture plays an important role in the characteristics of both the learning process and the resulting control policy, and that in particular trajectory optimization can be leveraged to improve convergence times for imitation learning. Finally, we identify some limitations and future areas of study and development for the technology. / Optimal kontroll för multikoptrar är ett svårt problem delvis på grund av den vanligtvis låga processorkraft som styrdatorn har, samt att multikoptrar är synnerligen instabila system. Djup imitationsinlärning är en metod där en beräkningstung expert approximeras med ett neuralt nätverk, och gör det därigenom möjligt att köra dessa tunga experter som realtidskontroll för multikoptrar. I detta arbete undersöks prestandan och pålitligheten hos djup imitationsinlärning med banoptimering som expert genom att först definiera en dynamisk modell för multikoptrar, sedan applicera en välkänd banoptimeringsmetod på denna modell, och till sist approximera denna expert med imitationsinlärning. Vår undersökning visar att nätverksarkitekturen spelar en avgörande roll för karakteristiken hos både inlärningsprocessens konvergenstid, såväl som den resulterande kontrollpolicyn, och att särskilt banoptimering kan nyttjas för att förbättra konvergenstiden hos imitationsinlärningen. Till sist påpekar vi några begränsningar hos metoden och identifierar särskilt intressanta områden för framtida studier.

The research on autonomous vehicles, and more specifically cooperation between autonomous vehicles, has become a prominent research field during the last cou- ple of decades. One example is the combination of an unmanned aerial vehicle (UAV) together with an unmanned ground vehicle (UGV). The benefits of this are that the two vehicles complement each other, where the UAV provides an aerial view and can reach areas where a ground vehicle can not. Furthermore, since the UAV has a limited range, the UGV can then serve as transport and recharge sta- tion for the UAV. This master thesis studies how model predictive control (MPC) can be used to land a UAV on a moving UGV.  A linear MPC is chosen, since previous work using this has shown promising results. The UAV is chosen to be controlled using commands in pitch, roll and climbing rate. The MPC is designed as a decoupled controller, with a separate horizontal and vertical controller. This allows for a spatial constraint to be im- plemented, which constrains the UAV from entering ground level before arriving above the UGV. It also constrains the UAV from potentially hitting protruding ob- jects on the UGV. The horizontal controller uses a simple planner, which guides the UAV to land on the UGV from behind.  The MPC is evaluated using a additive white Gaussian noise (AWGN) sen- sor error model with zero mean. The scenario used is that the UAV starts 50 m from the UGV, and the UGV starts driving in a given direction turning randomly. The MPC lands successfully in 100 % of the simulations for a wide range of tun- ings. The MPC maintains the same landing statistics with a delay in the sensor information of up to 500 ms. The AWGN could be increased while maintaining successful landings, however with significantly more retakes and longer landing times. Lower AWGN variance only slightly improves performance, suggesting that the MPC is quite robust towards high variance in the state estimation.  The MPC is also compared to a PID controller. The MPC has significantly shorter landing times. The PID has a more oscillatory control signal, however, the PID has a lower variance in landing positions, but a slightly less centered mean on the UGV. The overall results show that an MPC can be used to achieve a flexible controller that can be tuned and reformulated to fit the situation, and performs as good or better compared to a PID controller.  The hardware tests show promising results for the implementation of the MPC. The controller is not tuned and no system identification is done specifi- cally for the physical UAV, suggesting that the controller is robust for varying settings. Even though the UAV never lands on the UGV, the visual behavior and control signal plots suggest that it would be able to land. However, these tests are performed using global navigation satellite system state estimation on a sta- tionary UGV, therefore further tests need to be performed in more challenging scenarios.

The use of drones is becoming increasingly common in the industry due to their efficiency and safety in environments that are difficult to access. The development of industrial drones has created a new need for specialized drones with unique functions. This has led to a growing interest in additive manufacturing as a production method. Additive manufacturing, previously primarily used for prototyping, is now emerging as a viable manufacturing method. This evolution has in turn opened new design methods intended for additive manufacturing, such as topology optimization. The purpose of this project was to redesign a drone to increase its strength, reduce its weight and improve water resistance and appearance. This was achieved using a classic product development process where a concept was developed and refined using simulation tools and a product requirement specification derived from interviews and observations of the existing drone. The product was developed using SolidWorks tools such as Topology Optimization, The Finite Element (FEM) and Computer Aided Design (CAD). The result of the work is a detailed design and a prototype developed using Topology simulations based on the product requirement specification. This project lays the foundation for continued production and development of the drone. The conclusion drawn from the work is that the new product is an improved version in several aspects compared to the previous product.

This bachelors thesis presents a vision based control system for the quadrotor aerial vehicle,Crazy ie 2.0, developed by Bitcraze AB. The main goal of this thesis is to design andimplement an o-board control system based on visual input, in order to control the positionand orientation of the vehicle with respect to a single ducial marker. By integrating a cameraand wireless video transmitter onto the MAV platform, we are able to achieve autonomousnavigation and landing in relatively close proximity to the dedicated target location.The control system was developed in the programming language Python and all processing ofthe vision-data take place on an o-board computer. This thesis describes the methods usedfor developing and implementing the control system and a number of experiments have beencarried out in order to determine the performance of the overall vision control system. Withthe proposed method of using ducial markers for calculating the control demands for thequadrotor, we are able to achieve autonomous targeted landing within a radius of 10centimetres away from the target location. / I detta examensarbete presenteras ett visionsbaserat kontrollsystem for dronaren Crazy ie 2.0som har utvecklats av Bitcraze AB. Malet med detta arbete ar att utforma och implementeraett externt kontrollsystem baserat pa data som inhamtas av en kamera for att reglera fordonetsposition och riktning med avseende pa en markor placerad i synfaltet av kameran. Genom attintegrera kameran tillsammans med en tradlos videosandare pa plattformen, visar vi i dennaavhandling att det ar mojligt att astadkomma autonom navigering och landning i narheten avmarkoren.Kontrollsystemet utvecklades i programmeringsspraket Python och all processering avvisions-datan sker pa en extern dator. Metoderna som anvands for att utvecklakontrollsystemet och som beskrivs i denna rapport har testats under ett ertal experiment somvisar pa hur val systemet kan detektera markoren och hur val de olika ingaendekomponenterna samspelar for att kunna utfora den autonoma styrningen. Genom den metodsom presenteras i den har rapporten for att berakna styrsignalerna till dronaren med hjalp avvisuell data, visar vi att det ar mojligt att astadkomma autonom styrning och landning motmalet inom en radie av 10 centimeter.

This thesis presents a real-time autonomous guidance and control method for a quadrotor in a GPS-denied environment. The quadrotor autonomously seeks a destination while it avoids obstacles whose shape and position are initially unknown. We implement the obstacle avoidance and destination seeking methods using off-the-shelf sensors, including a vision-sensing camera. The vision-sensing camera detects the positions of points on the surface of obstacles. We use this obstacle position data and a potential-field method to generate velocity commands. We present a backstepping controller that uses the velocity commands to generate the quadrotor's control inputs. In indoor experiments, we demonstrate that the guidance and control methods provide the quadrotor with sufficient autonomy to fly point to point, while avoiding obstacles.

The potential of Unmanned Aerial Vehicles (UAVs) for surveillance and military applications is significant — with continued technical advances in the field. The number of incidents where UAVs have intruded into unauthorized areas has increased in recent years and armed drones are commonly used in modern warfare. It is therefore of great interest to investigate methods for UAVs to locate and track intruder drones to prevent and counter surveillance of unauthorized areas and attacks from intruder UAVs. This master’s thesis studied how two autonomous seeker UAVs can be used cooperatively to track and pursue a target UAV. To locate the target UAV, simulated measurements from received Radio Frequency (RF) signals were used by extracting bearing and Received Signal Strength (RSS) data. To track the target and predict its future position, the study employed an Extended Kalman Filter (EKF) on each seeker UAV, which acted together as a Mobile Wireless Sensor Network (MWSN). The thesis explored two formation control methods to keep the seeker UAVs in formation while pursuing the target drone. The formation methods used the predicted position of the target to produce reference positions and/or reference distances for a controller to follow. A Distributed Model Predictive Controller (DMPC) was implemented on the seeker UAVs to pursue the target while maintaining formation and avoiding collisions. The EKF, MPC, and formation methods were first evaluated individually in simulation to assess their performance and for parameter tuning. The respective modules were then combined into the complete system and tuned to achieve improved pursuit and formation in simulation. The results showed that, with the chosen parameters and with a high level of measurement noise, the seeker UAVs were able to pursue the target with a combined average distance error of less than 2 m when the target drone flew in a square pattern with a velocity of 2 m/s. The quality of the pursuit was highly affected by the increase in velocity of the target and the initial positions of the seekers, where a high velocity and a large initial deviation from the reference positions/distances resulted in poorer quality.

The omnicopter is a small unmanned aerial vehicle capable of executing decoupled translational and rotational motion (six degree of freedom, 6DOF, motion). The development of controllers for various 6DOF controllable multirotors has been much more limited than development for quadrotors, which makes selecting a controller for a 6DOF multirotor difficult. The omnicopter is subject to various uncertainties and disturbances from hardware changes, structural dynamics, and airflow, making adaptive controllers particularly interesting to investigate. The goal of this research is to design and evaluate the performance of various position and attitude controller combinations for the omnicopter, specifically focusing on adaptive controllers. Simulations are first used to compare combinations of three position controllers, PID, model reference adaptive control, augmented model reference adaptive control (aMRAC), and four attitude controllers, PI/feedback linearization (PIFL), augmented model reference adaptive control, backstepping, and adaptive backstepping (aBack). For the simulations, the omnicopter is commanded to point at and track a stationary aim point as it travels along a $C^0$ continuous trajectory and a trajectory that is $C^1$ continuous. The controllers are stressed by random disturbances and the addition of an unaccounted for suspended mass. The augmented model reference adaptive controller for position control paired with the adaptive backstepping controller for attitude control is shown to be the best controller combination for tracking various trajectories while subject to disturbances. Based on the simulation results, the PID/PIFL and aMRAC/aBack controllers are selected to be compared during three different flight tests. The first flight test is on a $C^1$ continuous trajectory while the omnicopter is commanded to point at and track a stationary aim point. The second flight test is a hover with an unmodeled added weight, and the third is a circular trajectory with a broken blade. As with the simulation results, the adaptive controller is shown to yield better performance than the nonadaptive controller for all scenarios, particularly for position tracking. With an added weight or a broken propeller, the adaptive attitude controller struggles to return to level flight, but is capable of maintaining steady flight when the nonadaptive controller tends to fail. Finally, while model reference adaptive controllers are shown to be effective, their nonlinearity can make them difficult to tune and certify via standard certification methods, such as gain and phase margin. A method for using time delay margin estimates, a potential certification metric, to tune the adaptive parameter tuning gain matrix is shown to be useful when applied to an augmented MRAC controller for a quadrotor. / Doctor of Philosophy / The omnicopter is a small unmanned aerial vehicle capable of executing decoupled translational and rotational motion. The development of controllers for these types of vehicles has been limited, making controller selection difficult. The omnicopter is subject to variations in hardware and airflow, making adaptive controllers particularly interesting to investigate. The goal of this research is to design and compare the performance of various position and attitude controller combinations for the omnicopter, specifically focusing on adaptive controllers. Simulations are first used to compare combinations of several position and attitude controllers on various trajectories and disturbances. Simulation results showed that a fully adaptive controller combination produced the best trajectory tracking while subject to disturbances. As with the simulation results, flight tests showed the adaptive controller yields better performance than the nonadaptive controller for all scenarios, particularly for position tracking. Finally, while the adaptive position controller was shown to be effective, it is difficult to tune and certify for widespread use. A method for using time delay margin estimates, a potential certification metric, to tune the adaptive controller is shown to be useful when applied to an adaptive controller for a quadrotor.

Unmanned Aerial Vehicle (UAV), also called drones, are used for Search And Rescue (SAR) missions, mainly in the form of a pilot manoeuvring a single drone. However, the increase in labour to cover larger areas quickly would result in a very high cost and time spent per rescue operation. Therefore, there is a need for an easy to use, low-cost, and highly autonomous swarm of drones for SAR missions where the detection and rescue times are kept to a minimum. In this thesis, a Subsumption-based architecture is proposed, which combines multiple behaviours to create more complex behaviours. An investigation of (1) what are the critical aspects of controlling a swarm of drones, (2) how can a combination of different behavioural algorithms increase the performance of a swarm of drones, and (3) what benchmarks are necessary when evaluating the fitness of the behavioural algorithms. The proposed architecture was simulated in AirSim using the SimpleFlight flight controller through experiments that evaluated the individual layers and missions that simulated real-life scenarios. The results validate the modularity and reliability of the architecture, where the architecture has the potential for improvements in future iterations. For the search area of 400×400meters, the swarm consistently produced an average area coverage of at least 99.917% and found all the missing people in all missions, with the slowest average being 563 seconds. Compared to related work, the result produced similar or better times when scaled to the same proportions and higher area coverage. As comparisons of results in SAR missions can be difficult, the introduction of Active time can serve as a benchmark for others in future swarm performance measurements.

The usage of autonomous vehicles is starting to appear in several different domains and the domain of public safety is no exception. Wallenberg Artificial Intelligence, Autonomous Systems and Software Program (WASP) has created a research arena for public safety (WARA-PS) to explore experimental features, usages, and implementation of autonomous vehicles within the domain of public safety. Collaborating in the arena are several companies, universities, and researchers. This thesis examines, in collaboration with Combitech, a company partnered in WARA-PS, how the implementation of autonomous vehicles affects the sociotechnical system of a search and rescue operation during a drifting boat with potential castaways. This is done by creating a case together with domain experts, analyzing the sociotechnical system within the case using cognitive work analysis and then complementing the analyses with the unmanned autonomous vehicles of WARA-PS. This thesis has shown how the WARA-PS vehicles can be implemented in the case of a drifting boat with potential castaways and how the implementation affects the sociotechnical system. Based on the analyses and opinions of domain experts’ future guidelines has been derived to further the work with sociotechnical aspects in WARA-PS. / WARA-PS