Rozprawy doktorskie na temat „UAV collision avoidance”

Includes abstract.
Includes bibliographical references.
An increased level of autonomy is required for future Unmanned Aerial Vehicle (UAV) missions. One of the technologies required for this to occur is an adequate sense and avoid system to enable the UAV to detect threat aircraft and take evasive action if required. This thesis investigates a collision avoidance system to satisfy a significant portion of the requirements for sense and avoid. It was hypothesised that a recently published method of UAV guidance, Specific Acceleration Matching (SAM) Control, could address the shortcomings of the current implementations. Additionally, a novel algorithm, the Linear 3D Velocity Guidance Control Algorithm (3DVGC) was developed to address the particular requirements of UAV collision avoidance.

ITC/USA 2008 Conference Proceedings / The Forty-Fourth Annual International Telemetering Conference and Technical Exhibition / October 27-30, 2008 / Town and Country Resort & Convention Center, San Diego, California
UUV homing and docking and UAV collision avoidance are two seemingly separate research topics for different applications. Upon close examination, these two are a pair of dual problems, with interesting correspondences and commonality. In this paper, we present the theoretical analysis, signal processing, and the field experiments of these two algorithms in UAV and UUV applications in homing and docking as well as collision avoidance.

For the last few years computer vision due to its low exploitation cost and great capabilities has been experiencing rapid growth. One of the research fields that benefits from it the most is the aircrafts positioning and collision avoidance. Light cameras with low energy consumption are an ideal solution for UAVs (Unmanned Aerial Vehicles) navigation systems. With the new Swedish law – unique to Europe, that allows for civil usage of UAVs that fly on altitudes up to 120 meters, the need for reliable and cheap positioning systems became even more dire. In this thesis two possible solutions for positioning problem and one for collision avoidance were proposed and analyzed. Possibility of tracking the vehicles position both from ground and from air was exploited. Camera setup for successful positioning and collision avoidance systems was defined and preliminary results for of the systems performance were presented.
Vision systems are employed more and more often in navigation of ground and air robots. Their greatest advantages are: low cost compared to other sensors, ability to capture large portion of the environment very quickly on one image frame, and their light weight, which is a great advantage for air drone navigation systems. In the thesis the problem of UAV (Unmanned Aerial Vehicle) is considered. Two different issues are tackled. First is determining the vehicles position using one down-facing or two front-facing cameras, and the other is sparse obstacle detection. Additionally, in the thesis, the camera calibration process and camera set up for navigation is discussed. Error causes and types are analyzed.

Unmanned aircraft systems (UAS) represent the future of modern aviation. Over the past 10 years their use abroad by the military has become commonplace for surveillance and combat. Unfortunately, their use at home has been far more restrictive. Due to safety and regulatory concerns, UAS are prohibited from flying in the National Airspace System without special authorization from the FAA. One main reason for this is the lack of an on-board pilot to "see and avoid" other air traffic and thereby maintain the safety of the skies. Development of a comparable capability, known as "Sense and Avoid" (SAA), has therefore become a major area of focus. This research focuses on the SAA problem as it applies specifically to small UAS. Given the size, weight, and power constraints on these aircraft, current approaches fail to provide a viable option. To aid in the development of a SAA system for small UAS, various simulation and hardware tools are discussed. The modifications to the MAGICC Lab's simulation environment to provide support for multiple agents is outlined. The use of C-MEX s-Functions to improve simulation performance and code portability is also presented. For hardware tests, two RC airframes were constructed and retrofitted with autopilots to allow autonomous flight. The development of a program to interface with the ground control software and run the collision avoidance algorithms is discussed as well. Intruder sensing is accomplished using a low-power, low-resolution radar for detection and an Extended Kalman Filter (EKF) for tracking. The radar provides good measurements for range and closing speed, but bearing measurements are poor due to the low-resolution. A novel method for improving the bearing approximation using the raw radar returns is developed and tested. A four-state EKF used to track the intruder's position and trajectory is derived and used to provide estimates to the collision avoidance planner. Simulation results and results from flight tests using a simulated radar are both presented. To effectively plan collision avoidance paths a tree-branching path planner is developed. Techniques for predicting the intruder position and creating safe, collision-free paths using the estimates provided by the EKF are presented. A method for calculating the cost of flying each path is developed to allow the selection of the best candidate path. As multiple duplicate paths can be created using the branching planner, a strategy to remove these paths and greatly increase computation speed is discussed. Both simulation and hardware results are presented for validation.

Approximately 20 years have passed now since the NTSB issued its original recommendation to expedite development, certification and production of low-cost proximity warning and conflict detection systems for general aviation [1]. While some systems are in place (TCAS [2]), ¡¨see-and-avoid¡¨ remains the primary means of separation between light aircrafts sharing the national airspace. The requirement for a collision avoidance or sense-and-avoid capability onboard unmanned aircraft has been identified by leading government, industry and regulatory bodies as one of the most significant challenges facing the routine operation of unmanned aerial systems (UAS) in the national airspace system (NAS) [3, 4]. In this thesis, we propose and develop a novel image-based collision avoidance system to detect and avoid an upcoming conflict scenario (with an intruder) without first estimating or filtering range. The proposed collision avoidance system (CAS) uses relative bearing ƒÛ and angular-area subtended ƒê , estimated from an image, to form a test statistic AS C . This test statistic is used in a thresholding technique to decide if a conflict scenario is imminent. If deemed necessary, the system will command the aircraft to perform a manoeuvre based on ƒÛ and constrained by the CAS sensor field-of-view. Through the use of a simulation environment where the UAS is mathematically modelled and a flight controller developed, we show that using Monte Carlo simulations a probability of a Mid Air Collision (MAC) MAC RR or a Near Mid Air Collision (NMAC) RiskRatio can be estimated. We also show the performance gain this system has over a simplified version (bearings-only ƒÛ ). This performance gain is demonstrated in the form of a standard operating characteristic curve. Finally, it is shown that the proposed CAS performs at a level comparable to current manned aviations equivalent level of safety (ELOS) expectations for Class E airspace. In some cases, the CAS may be oversensitive in manoeuvring the owncraft when not necessary, but this constitutes a more conservative and therefore safer, flying procedures in most instances.

This thesis work presents a preliminary design of a detection and avoidance system for an Unmanned Aerial Vehicle (UAV). The code realized is capable to sense cooperative aircraft surrounding the UAV and to avoid them in a complete autonomous way. The detection of the aircraft is made by the elaboration of the information that they share in the medium. On the other hand, the generation of the trajectories to avoid the collisions is treated using a geometrical approach and dealing with different scenarios: no-moving obstacle, moving obstacle, multiple obstacles and terrain collision avoidance. With the so called point mass model, the controls to obtain the generated trajectory are found accordantly with the main literature. The proposed solution can be used for almost any kind of UAV that shall operate in civil air space.

This thesis investigates the crucial problem of collision avoidance for autonomous vehicles.  An anti-collision system for an unmanned aerial vehicle (UAV) is studied in particular. The purpose of this system is to make sure that the own vehicle avoids collision with other aircraft in mid-air. The sensor used to track any possible threat is for a UAV limited basically to a digital video camera. This sensor can only measure the direction to an intruding vehicle, not the range, and is therefore denoted an angle-only sensor. To estimate the position and velocity of the intruder a tracking system, based on an extended Kalman filter, is used. State estimates supplied by this system are very uncertain due to the difficulties of angle-only tracking. Probabilistic methods are therefore required for risk calculation. The risk assessment module is one of the essential parts of the collision avoidance system and has the purpose of continuously evaluating the risk for collision. To do this in a probabilistic way, it is necessary to assume a probability distribution for the tracking system output. A common approach is to assume normality, more out of habit than on actual grounds. This thesis investigates the normality assumption, and it is found that the tracking output rapidly converge towards a good normal distribution approximation. The thesis furthermore investigates the actual risk assessment module to find out how the collision risk should be determined. The traditional way to do this is to focus on a critical time point (time of closest point of approach, time of maximum collision risk etc.). A recently proposed alternative is to evaluate the risk over a horizon of time. The difference between these two concepts is evaluated. An approximate computational method for integrated risk, suitable for real-time implementations, is also validated. It is shown that the risk seen over a horizon of time is much more robust to estimation accuracy than the risk from a critical time point. The integrated risk also gives a more intuitively correct result, which makes it possible to implement the risk assessment module with a direct connection to specified aviation safety rules.

The thesis aims to design and implement an Android application with ability to control the autopilot of the Skydog aircraft model using the wireless telemetry. The application shall receive data from an aircraft model gathered from various installed sensors. These data shall be then processed and corresponding instructions for autopilot shall be sent back. When collision with terrain or obstacle is detected, the application shall send instructions to autopilot to avoid such collision. RRT algorithm is used to find collision-free flight trajectory. Database of known obstacles and digital terrain model are provided to application in formats XML and GeoTIFF respectively.

Uninhabited aerial vehicles (UAVs) are a cutting-edge technology that is at the forefront of aviation/aerospace research and development worldwide. Many consider their current military and defence applications as just a token of their enormous potential. Unlocking and fully exploiting this potential will see UAVs in a multitude of civilian applications and routinely operating alongside piloted aircraft. The key to realising the full potential of UAVs lies in addressing a host of regulatory, public relation, and technological challenges never encountered be- fore. Aircraft collision avoidance is considered to be one of the most important issues to be addressed, given its safety critical nature. The collision avoidance problem can be roughly organised into three areas: 1) Sense; 2) Detect; and 3) Avoid. Sensing is concerned with obtaining accurate and reliable information about other aircraft in the air; detection involves identifying potential collision threats based on available information; avoidance deals with the formulation and execution of appropriate manoeuvres to maintain safe separation. This thesis tackles the detection aspect of collision avoidance, via the development of a target detection algorithm that is capable of real-time operation onboard a UAV platform. One of the key challenges of the detection problem is the need to provide early warning. This translates to detecting potential threats whilst they are still far away, when their presence is likely to be obscured and hidden by noise. Another important consideration is the choice of sensors to capture target information, which has implications for the design and practical implementation of the detection algorithm. The main contributions of the thesis are: 1) the proposal of a dim target detection algorithm combining image morphology and hidden Markov model (HMM) filtering approaches; 2) the novel use of relative entropy rate (RER) concepts for HMM filter design; 3) the characterisation of algorithm detection performance based on simulated data as well as real in-flight target image data; and 4) the demonstration of the proposed algorithm's capacity for real-time target detection. We also consider the extension of HMM filtering techniques and the application of RER concepts for target heading angle estimation. In this thesis we propose a computer-vision based detection solution, due to the commercial-off-the-shelf (COTS) availability of camera hardware and the hardware's relatively low cost, power, and size requirements. The proposed target detection algorithm adopts a two-stage processing paradigm that begins with an image enhancement pre-processing stage followed by a track-before-detect (TBD) temporal processing stage that has been shown to be effective in dim target detection. We compare the performance of two candidate morphological filters for the image pre-processing stage, and propose a multiple hidden Markov model (MHMM) filter for the TBD temporal processing stage. The role of the morphological pre-processing stage is to exploit the spatial features of potential collision threats, while the MHMM filter serves to exploit the temporal characteristics or dynamics. The problem of optimising our proposed MHMM filter has been examined in detail. Our investigation has produced a novel design process for the MHMM filter that exploits information theory and entropy related concepts. The filter design process is posed as a mini-max optimisation problem based on a joint RER cost criterion. We provide proof that this joint RER cost criterion provides a bound on the conditional mean estimate (CME) performance of our MHMM filter, and this in turn establishes a strong theoretical basis connecting our filter design process to filter performance. Through this connection we can intelligently compare and optimise candidate filter models at the design stage, rather than having to resort to time consuming Monte Carlo simulations to gauge the relative performance of candidate designs. Moreover, the underlying entropy concepts are not constrained to any particular model type. This suggests that the RER concepts established here may be generalised to provide a useful design criterion for multiple model filtering approaches outside the class of HMM filters. In this thesis we also evaluate the performance of our proposed target detection algorithm under realistic operation conditions, and give consideration to the practical deployment of the detection algorithm onboard a UAV platform. Two fixed-wing UAVs were engaged to recreate various collision-course scenarios to capture highly realistic vision (from an onboard camera perspective) of the moments leading up to a collision. Based on this collected data, our proposed detection approach was able to detect targets out to distances ranging from about 400m to 900m. These distances, (with some assumptions about closing speeds and aircraft trajectories) translate to an advanced warning ahead of impact that approaches the 12.5 second response time recommended for human pilots. Furthermore, readily available graphic processing unit (GPU) based hardware is exploited for its parallel computing capabilities to demonstrate the practical feasibility of the proposed target detection algorithm. A prototype hardware-in- the-loop system has been found to be capable of achieving data processing rates sufficient for real-time operation. There is also scope for further improvement in performance through code optimisations. Overall, our proposed image-based target detection algorithm offers UAVs a cost-effective real-time target detection capability that is a step forward in ad- dressing the collision avoidance issue that is currently one of the most significant obstacles preventing widespread civilian applications of uninhabited aircraft. We also highlight that the algorithm development process has led to the discovery of a powerful multiple HMM filtering approach and a novel RER-based multiple filter design process. The utility of our multiple HMM filtering approach and RER concepts, however, extend beyond the target detection problem. This is demonstrated by our application of HMM filters and RER concepts to a heading angle estimation problem.

This thesis explain today and future technologies of air traffic collision avoidance systems and of its utilizability by unmanned aircraft. The first part describes the UAVs and their categorization. Next part deals with the legislative requirements for their operation. The main part deals with TCAS, ADS-B, FLARM and others that are now used in civil aviation as a key technology to avoid a collision. The last part describes the UAV systems with a focus on the actual sensors used in systems for unmanned aerial vehicles for collision avoidance. The whole work deals with issues of development of collision avoidance systems and summarizes the current technology with a view to its possible application in the near future.

Unmanned Aerial Vehicles (UAVs) can navigate in indoor environments and through environments that are hazardous or hard to reach for humans. This makes them suitable for use in search and rescue missions and by emergency response and law enforcement to increase situational awareness. However, even for an experienced UAV tele-operator controlling the UAV in these situations without colliding into obstacles is a demanding and difficult task. This thesis presents a human-UAV interface along with a collision avoidance method, both optimized for a human tele-operator. The objective is to simplify the task of navigating a UAV in indoor environments. Evaluation of the system is done by testing it against a number of use cases and a user study. The results of this thesis is a collision avoidance method that is successful in protecting the UAV from obstacles while at the same time acknowledges the operator’s intentions.
Obemannad luftfarkoster (UAV:er) kan navigera i inomhusmiljöer och genom miljöer som är farliga eller svåra att nå för människor. Detta gör dem lämpliga för användning i sök- och räddningsuppdrag och av akutmottagning och rättsväsende genom ökad situationsmedvetenhet. Dock är det även för en erfaren UAV-teleoperatör krävande och svårt att kontrollera en UAV i dessa situationer utan att kollidera med hinder. Denna avhandling presenterar ett människa-UAV-gränssnitt tillsammans med en kollisionsundvikande metod, båda optimerade för en mänsklig teleoperatör. Målet är att förenkla uppgiften att navigera en UAV i inomhusmiljöer. Utvärdering av systemet görs genom att testa det mot ett antal användningsfall och en användarstudie. Resultatet av denna avhandling är en kollisionsundvikande metod som lyckas skydda UAV från hinder och samtidigt tar hänsyn till operatörens avsikter.

Operation of Unmanned Aerial Vehicles (UAVs) in civil airspace is restricted by the aviation authorities which require full compliance with regulations that apply for manned aircraft. This thesis proposes control algorithms for a collision avoidance system that can be used as an advisory system or a guidance system for UAVs that are flying in civil airspace under visual flight rules. An effective collision avoidance system for the UAV should be able to perform the different functionalities of the pilot in manned aircraft. Thus, it should be able to determine, generate, and perform safe avoidance manoeuvres. However, the capability to generate resolution advisories is crucial for the advisory systems. A decision making system for collision avoidance is developed based on the rules of the air. The proposed architecture of the decision making system is engineered to be implementable in both manned aircraft and UAVs to perform different tasks ranging from collision detection to a safe avoidance manoeuvre initiation. Avoidance manoeuvres that are compliant with the rules of the air are proposed based on pilot suggestions for a subset of possible collision scenarios. The avoidance manoeuvre generation algorithm is augmented with pilot experience by using fuzzy logic technique to model pilot actions in generating the avoidance manoeuvres. Hence, the generated avoidance manoeuvres mimic the avoidance manoeuvres of manned aircraft. The proposed avoidance manoeuvres are parameterized using a geometric approach. An optimal collision avoidance algorithm is developed for real-time local trajectory planning. Essentially, a finite-horizon optimal control problem is periodically solved in real-time hence updating the aircraft trajectory to avoid obstacles and track a predefined trajectory. The optimal control problem is formulated in output space, and parameterised by using B-splines. Then the optimal designed outputs are mapped into control inputs of the system by using the inverse dynamics of a fixed wing aircraft.

The thesis aims to investigate and develop innovative tools to provide autonomous flight capability to a fixed-wing unmanned aircraft. Particularly it contributes to research on path optimization, tra jectory tracking and collision avoidance with two algorithms designed respectively for path planning and navigation. The complete system generates the shortest path from start to target avoiding known obstacles represented on a map, then drives the aircraft to track the optimum path avoiding unpredicted ob jects sensed in flight. The path planning algorithm, named Kinematic A*, is developed on the basis of graph search algorithms like A* or Theta* and is meant to bridge the gap between path-search logics of these methods and aircraft kinematic constraints. On the other hand the navigation algorithm faces concurring tasks of tra jectory tracking and collision avoidance with Nonlinear Model Predictive Control. When A* is applied to path planning of unmanned aircrafts any aircraft kinematics is taken into account, then practicability of the path is not guaranteed. Kinematic A* (KA*) generates feasible paths through graph-search logics and basic vehicle characteristics. It includes a simple aircraft kinematic-model to evaluate moving cost between nodes of tridimensional graphs. Movements are constrained with minimum turning radius and maximum rate of climb. Furtermore, separation from obstacles is imposed, defining a volume around the path free from obstacles (tube-type boundaries). Navigation is safe when the tracking error does not exceed this volume. The path-tracking task aims to link kinematic information related to desired aircraft positions with dynamic behaviors to generate commands that minimize the error between reference and real tra jectory. On the other hand avoid obstacles in flight is one of the most challenging tasks for autonomous aircrafts and many elements must be taken into account in order to implement an effective collision avoidance maneuver. Second part of the thesis describes a Nonlinear Model Predictive Control (NMPC) application to cope with collision avoidance and path tracking tasks. First contribution is the development of a navigation system able to match concurring problems: track the optimal path provided with KA* and avoid unpredicted obstacles detected with sensors. Second Contribution is the Sense & Avoid (S&A) technique exploiting spherical camera and visual servoing control logics.

The increasing demand to integrate unmanned aircraft systems (UAS) into the national airspace is motivated by the rapid growth of the UAS industry, especially small UAS weighing less than 55 pounds. Their use however has been limited by the Federal Aviation Administration regulations due to collision risk they pose, safety and regulatory concerns. Therefore, before civil aviation authorities can approve routine UAS flight operations, UAS must be equipped with sense-and-avoid technology comparable to the see-and-avoid requirements for manned aircraft. The sense-and-avoid problem includes several important aspects including regulatory and system-level requirements, design specifications and performance standards, intruder detecting and tracking, collision risk assessment, and finally path planning and collision avoidance. In this dissertation, our primary focus is on developing an collision detection, risk assessment and avoidance framework that is computationally affordable and suitable to run on-board small UAS. To begin with, we address the minimum sensing range for the sense-and-avoid (SAA) system. We present an approximate close form analytical solution to compute the minimum sensing range to safely avoid an imminent collision. The approach is then demonstrated using a radar sensor prototype that achieves the required minimum sensing range. In the area of collision risk assessment and collision prediction, we present two approaches to estimate the collision risk of an encounter scenario. The first is a deterministic approach similar to those been developed for Traffic Alert and Collision Avoidance (TCAS) in manned aviation. We extend the approach to account for uncertainties of state estimates by deriving an analytic expression to propagate the error variance using Taylor series approximation. To address unanticipated intruders maneuvers, we propose an innovative probabilistic approach to quantify likely intruder trajectories and estimate the probability of collision risk using the uncorrelated encounter model (UEM) developed by MIT Lincoln Laboratory. We evaluate the proposed approach using Monte Carlo simulations and compare the performance with linearly extrapolated collision detection logic. For the path planning and collision avoidance part, we present multiple reactive path planning algorithms. We first propose a collision avoidance algorithm based on a simulated chain that responds to a virtual force field produced by encountering intruders. The key feature of the proposed approach is to model the future motion of both the intruder and the ownship using a chain of waypoints that are equally spaced in time. This timing information is used to continuously re-plan paths that minimize the probability of collision. Second, we present an innovative collision avoidance logic using an ownship centered coordinate system. The technique builds a graph in the local-level frame and uses the Dijkstra's algorithm to find the least cost path. An advantage of this approach is that collision avoidance is inherently a local phenomenon and can be more naturally represented in the local coordinates than the global coordinates. Finally, we propose a two step path planner for ground-based SAA systems. In the first step, an initial suboptimal path is generated using A* search. In the second step, using the A* solution as an initial condition, a chain of unit masses connected by springs and dampers evolves in a simulated force field. The chain is described by a set of ordinary differential equations that is driven by virtual forces to find the steady-state equilibrium. The simulation results show that the proposed approach produces collision-free plans while minimizing the path length. To move towards a deployable system, we apply collision detection and avoidance techniques to a variety of simulation and sensor modalities including camera, radar and ADS-B along with suitable tracking schemes.

This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Thesis: S.M., Massachusetts Institute of Technology, Department of Aeronautics and Astronautics, 2019
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 115-118).
Small Unmanned Aircraft Systems (sUAS) have proliferated over the last decade. While these platforms offer many benefits to society, they pose a dangerous mid-air collision hazard. In order to safely integrate into airspace shared by other users, sUAS must be able to avoid collisions with manned aircraft. To better understand sUAS flight behavior and inform Collision Avoidance (CA) systems for sUAS, over 600 active UAS platforms were reviewed. The mean climb rate capability was 720 feet per minute (fpm) for all reviewed sUAS, which suggests that CA systems currently used by manned aircraft (which require 2,500 fpm climb capability) would be inappropriate for implementation on sUAS. Novel CA systems are therefore required. Next, to assess the feasibility of CA system equipage on sUAS, the Size, Weight, Power, and Cost (SWaP-C) of equipment necessary for CA systems were studied.
It was determined that a complete CA system utilizing cooperative surveillance could weigh less than 70 grams and require less than 2 W of average input power. Because cooperative surveillance broadcasts from sUAS could overload the spectrum currently used to share aviation information, signal transmissions were simulated for a population of sUAS broadcasting alongside current users. While transmitting sUAS would quickly degrade performance on the busy 1090 MHz channel, the 978 MHz channel could potentially support about 1 transmitting sUAS per square kilometer if sUAS broadcast ADS-B signals at only 80 mW. Finally, close encounters between sUAS and manned aircraft were simulated in the Mode C Veil environment to evaluate threat resolution options used by different CA systems. Manned aircraft using existing CA systems to avoid sUAS would achieve extremely high levels of safety (risk ratios below 0.05) but would experience high rates of alerts.
Furthermore, sUAS are so small that manned aircraft without CA systems would be unlikely to visually see and avoid them. Novel CA systems were modeled on sUAS and were able to avoid manned aircraft with currently-accepted levels of safety (risk ratios below 0.18) even with limited or no vertical maneuvering by using horizontal escape maneuvers (i.e. turns). Alerting rates for horizontal maneuvers were high but may be acceptable for use on sUAS. The new sUAS CA systems cooperated well with existing systems for manned aircraft and resulted in extremely low collision risk (risk ratios below 0.02) in encounters where manned aircraft and sUAS both took action to avoid collisions. Results therefore indicate that sUAS could utilize existing cooperative surveillance systems and prototype CA policies to mitigate close encounters with manned aircraft in Mode C Veils at safety levels that are currently accepted among manned aircraft.
by John Logan Deaton.
S.M.
S.M. Massachusetts Institute of Technology, Department of Aeronautics and Astronautics

For unmanned aircraft systems to gain full access to the National Airspace System (NAS), they must have the capability to detect and avoid other aircraft. This research focuses on the development of a detect-and-avoid (DAA) system for small unmanned aircraft systems. To safely avoid another aircraft, an unmanned aircraft must detect the intruder aircraft with ample time and distance. Two analytical methods for finding the minimum detection range needed are described. The first method, time-based geometric velocity vectors (TGVV), includes the bank-angle dynamics of the ownship while the second, geometric velocity vectors (GVV), assumes an instantaneous bank-angle maneuver. The solution using the first method must be found numerically, while the second has a closed-form analytical solution. These methods are compared to two existing methods. Results show the time-based geometric velocity vectors approach is precise, and the geometric velocity vectors approach is a good approximation under many conditions. The DAA problem requires the use of a robust target detection and tracking algorithm for tracking multiple maneuvering aircraft in the presence of noisy, cluttered, and missed measurements. Additionally these algorithms needs to be able to detect overtaking intruders, which has been resolved by using multiple radar sensors around the aircraft. To achieve these goals the formulation of a nonlinear extension to R-RANSAC has been performed, known as extended recursive-RANSAC (ER-RANSAC). The primary modifications needed for this ER-RANSAC implementation include the use of an EKF, nonlinear inlier functions, and the Gauss-Newton method for model hypothesis and generation. A fully functional DAA system includes target detection and tracking, collision detection, and collision avoidance. In this research we demonstrate the integration of each of the DAA-system subcomponents into fully functional simulation and hardware implementations using a ground-based radar setup. This integration resulted in various modifications of the radar DSP, collision detection, and collision avoidance algorithms, to improve the performance of the fully integrated DAA system. Using these subcomponents we present flight results of a complete ground-based radar DAA system, using actual radar hardware.

Autonomy is an essential feature of any robotic system. Aerial robots, commonly known as Unmanned Aerial Vehicles (UAVs), are being integrated into airspace and various trials towards achieving higher levels of autonomy are in progress. When multiple UAVs share the same airspace, safety from inter-UAV conflict is of utmost importance. Collision avoidance is an unavoidable feature of any UAV, and diverse methods addressing this problem are available in the literature. This thesis presents avoidance maps, a collision avoidance algorithm for fixed-wing UAVs. Avoidance of fixed-wing UAVs is challenging because of their inability to hover in contrast to their rotary-wing counterpart. Further, physical constraints like minimum turn radius make the process less flexible. The proposed avoidance map partitions the control input space of the UAVs into those leading to collision (red region) and avoidance (green region). Here, the control input used is constant lateral acceleration. Various versions of this are developed, which improves its computational cost. The algorithm could be implemented for cooperative, non-cooperative, and multiple UAVs and is demonstrated by suitable examples. In the next part of this thesis, precision UAV collision avoidance is discussed. This method is characterized by a gradual reduction of applied lateral acceleration during the avoidance process. Precision-control based avoidance optimizes the energy expenditure of the UAVs. The UAVs get away from their initial course while maneuvering. They are brought back to the initial direction of motion using Dubins curves, which joins two points via the shortest distance. The return to the course is achieved by Dubins path, where the necessary maneuvers are chosen from the avoidance map. An avoidance map can be used for realistic systems also. This utility is demonstrated by simulations using guidance models and six-degree-of-freedom UAV models. The avoidance map is further extended to few versions in the subsequent chapter. A time-graded version is introduced first, which classifies the collision region based on time to collision. This enables the use of several maneuvers from the collision region of the map as well. Next, asynchronous avoidance is introduced, which makes the avoidance process flexible for UAVs. The asynchronous avoidance maps compute avoidance maneuvers with a predetermined time delay for either of the UAVs. This results in one of the UAVs remaining on course for the desired time delay before maneuvering to avoid. Avoidance map is extended for constrained environments like corridors or geo-fences where the control input is the UAV heading angle. The application of avoidance maps for virtual intersections and lane changing for UAV virtual skyways are also discussed in this work. The last part of the thesis formulates collision avoidance of UAVs using game theory. This applies to both fixed-wing and rotor-craft categories and is based on the solution concept of correlated equilibrium. UAVs are considered to be intelligent players and the conflict resolution process is formulated as a game. The decision-making framework, which is termed CONCORD, works independently of the kind of avoidance algorithm used. The framework is found suitable for cooperative, non-cooperative, and multiple UAVs. It is shown that the proposed framework fairly resolves conflicts among UAVs and guarantees safety. A brief discussion on UAV integration to airspace and concord integration to such UAV traffic management system concludes this work.

碩士
淡江大學
航空太空工程學系碩士班
107
This thesis mainly studies the design of the guidance law and collision avoidance control for UAV formation flight. The leader-follower approach was adopted for the guidance law development. The error between the follower’s position and the desired position was used for the generation of the guidance law. The collision avoidance controls involve the collision avoidance for leader and follower, and for follower and follower. The collision avoidance control for leader and follower is based on a geometric collision-cone approach using the line-of-sight vector, the relative speed and the distance between leader and follower as conditions for generation of the control action. The collision avoidance law for the follower and follower is mainly based on the distance between the followers. Gain scheduling was also incorporated for the formation control to avoid large transient during mode transition between collision avoidance control and formation control. Computer simulations, including straight flight, level turn and regrouping were conducted to demonstrate the success of the UAV formation flight and collision avoidance control.

碩士
國立交通大學
電信工程研究所
107
In this thesis, we propose a reinforcement learning approach of collision avoidance and investigate optimal trajectory planning for unmanned aerial vehicle (UAV) communication networks. Specifically, each UAV takes charge of delivering objects in the forward path and collecting data from ground IoT devices in the backward path. We adopt reinforcement learning for assisting UAVs to learn collision avoidance without knowing the trajectories of other UAVs in advance. In addition, for each UAV, we use optimization theory to find out a shortest backward path that assures data collection from all associated IoT devices. To obtain an optimal visiting order for IoT devices, we formulate and solve a no-return traveling salesman problem with neighborhoods (TSPN). Given a visiting order, we formulate and solve a sequence of convex optimization problems to obtain segments of an optimal backward path. We use analytical results and simulation results to justify the usage of the proposed approach. Simulation results show that the proposed approach is superior to a number of alternative approaches.

The Unmanned Aerial Vehicles (UAV) and its applications are growing for both civilian and military purposes. The operability of an UAV proved that some tasks and operations can be done easily and at a good cost-efficiency ratio. Nowadays, an UAV can perform autonomous tasks, by using waypoint mission navigation using a GPS sensor. These autonomous tasks are also called missions. It is very useful to certain UAV applications, such as meteorology, vigilance systems, agriculture, environment mapping and search and rescue operations. One of the biggest problems that an UAV faces is the possibility of collision with other objects in the flight area. This can cause damage to surrounding area structures, humans or the UAV itself. To avoid this, an algorithm was developed and implemented in order to prevent UAV collision with other objects. “Sense and Avoid” algorithm was developed as a system for UAVs to avoid objects in collision course. This algorithm uses a laser distance sensor called LiDAR (Light Detection and Ranging), to detect objects facing the UAV in mid-flights. This light sensor is connected to an on-board hardware, Pixhawk’s flight controller, which interfaces its communications with another hardware: Raspberry Pi. Communications between Ground Control Station or RC controller are made via Wi-Fi telemetry or Radio telemetry. “Sense and Avoid” algorithm has two different modes: “Brake” and “Avoid and Continue”. These modes operate in different controlling methods. “Brake” mode is used to prevent UAV collisions with objects when controlled by a human operator that is using a RC controller. “Avoid and Continue” mode works on UAV’s autonomous modes, avoiding collision with objects in sight and proceeding with the ongoing mission. In this dissertation, some tests were made in order to evaluate the “Sense and Avoid” algorithm’s overall performance. These tests were done in two different environments: A 3D simulated environment and a real outdoor environment. Both modes worked successfully on a simulated 3D environment, and “Brake” mode on a real outdoor, proving its concepts.
Os veículos aéreos não tripulados (UAV) e as suas aplicações estão cada vez mais a ser utilizadas para fins civis e militares. A operacionalidade de um UAV provou que algumas tarefas e operações podem ser feitas facilmente e com uma boa relação de custo-benefício. Hoje em dia, um UAV pode executar tarefas autonomamente, usando navegação por waypoints e um sensor de GPS. Essas tarefas autónomas também são designadas de missões. As missões autónomas poderão ser usadas para diversos propósitos, tais como na meteorologia, sistemas de vigilância, agricultura, mapeamento de áreas e operações de busca e salvamento. Um dos maiores problemas que um UAV enfrenta é a possibilidade de colisão com outros objetos na área, podendo causar danos às estruturas envolventes, aos seres humanos ou ao próprio UAV. Para evitar tais ocorrências, foi desenvolvido e implementado um algoritmo para evitar a colisão de um UAV com outros objetos. O algoritmo "Sense and Avoid" foi desenvolvido como um sistema para UAVs de modo a evitar objetos em rota de colisão. Este algoritmo utiliza um sensor de distância a laser chamado LiDAR (Light Detection and Ranging), para detetar objetos que estão em frente do UAV. Este sensor é ligado a um hardware de bordo, a controladora de voo Pixhawk, que realiza as suas comunicações com outro hardware complementar: o Raspberry Pi. As comunicações entre a estação de controlo ou o operador de comando RC são feitas via telemetria Wi-Fi ou telemetria por rádio. O algoritmo "Sense and Avoid" tem dois modos diferentes: o modo "Brake" e modo "Avoid and Continue". Estes modos operam em diferentes métodos de controlo do UAV. O modo "Brake" é usado para evitar colisões com objetos quando controlado via controlador RC por um operador humano. O modo "Avoid and Continue" funciona nos modos de voo autónomos do UAV, evitando colisões com objetos à vista e prosseguindo com a missão em curso. Nesta dissertação, alguns testes foram realizados para avaliar o desempenho geral do algoritmo "Sense and Avoid". Estes testes foram realizados em dois ambientes diferentes: um ambiente de simulação em 3D e um ambiente ao ar livre. Ambos os modos obtiveram funcionaram com sucesso no ambiente de simulação 3D e o mode “Brake” no ambiente real, provando os seus conceitos.

This thesis addresses the problems of collision avoidance and coalition formation of multiple UAVs in high density traffic environments, proposes simple and efficient algorithms as solutions, and discusses their applications in multiple UAV missions. First, the problem of collision avoidance among UAVs is considered and deconfliction algorithms are proposed. The efficacy of the proposed algorithms is tested using simulations involving random flights in high density traffic. Further, the proposed collision avoidance algorithms are implemented using realistic six degree of freedom UAV models. The studies in this thesis show that implementation of the proposed collision avoidance algorithms leads to a safer and efficient operational airspace occupied by multiple UAVs. Next, coalition formation in a search and prosecute mission involving a large number of UAVs and targets is considered. This problem is shown to be NP-hard and a sub-optimal but polynomial time coalition formation strategy is proposed. Simulations are carried out to show that this coalition formation algorithm works well. The coalition formation algorithm is then extended to handle situations where the UAVs have limited communication ranges. Finally, this thesis considers some multiple UAV missions that require the application of collision avoidance and coalition formation techniques. The problem of multiple UAV rendezvous is tackled by using (i) a consensus among the UAVs to attain rendezvous and (ii) the collision avoidance algorithm previously developed for safety. The thesis also considers a search and prosecute mission where the UAVs also have to avoid collisions among one another. In summary, the main contributions of this thesis include (a) novel collision avoidance algorithms, which are conceptually simple and easy to implement, for resolving path conflicts – both planar and three dimensional – in a high density traffic airspace with UAVs in free flight and (b) efficient coalition formation algorithms for search and prosecute task with large number of UAVs and targets where UAVs have limited communication ranges and targets are maneuvering. Simulations to evaluate the performance of algorithms based on these concepts to carry out realistic tasks by UAV swarms are also given.

This thesis addresses the problems of collision avoidance and coalition formation of multiple UAVs in high density traffic environments, proposes simple and efficient algorithms as solutions, and discusses their applications in multiple UAV missions. First, the problem of collision avoidance among UAVs is considered and deconfliction algorithms are proposed. The efficacy of the proposed algorithms is tested using simulations involving random flights in high density traffic. Further, the proposed collision avoidance algorithms are implemented using realistic six degree of freedom UAV models. The studies in this thesis show that implementation of the proposed collision avoidance algorithms leads to a safer and efficient operational airspace occupied by multiple UAVs. Next, coalition formation in a search and prosecute mission involving a large number of UAVs and targets is considered. This problem is shown to be NP-hard and a sub-optimal but polynomial time coalition formation strategy is proposed. Simulations are carried out to show that this coalition formation algorithm works well. The coalition formation algorithm is then extended to handle situations where the UAVs have limited communication ranges. Finally, this thesis considers some multiple UAV missions that require the application of collision avoidance and coalition formation techniques. The problem of multiple UAV rendezvous is tackled by using (i) a consensus among the UAVs to attain rendezvous and (ii) the collision avoidance algorithm previously developed for safety. The thesis also considers a search and prosecute mission where the UAVs also have to avoid collisions among one another. In summary, the main contributions of this thesis include (a) novel collision avoidance algorithms, which are conceptually simple and easy to implement, for resolving path conflicts – both planar and three dimensional – in a high density traffic airspace with UAVs in free flight and (b) efficient coalition formation algorithms for search and prosecute task with large number of UAVs and targets where UAVs have limited communication ranges and targets are maneuvering. Simulations to evaluate the performance of algorithms based on these concepts to carry out realistic tasks by UAV swarms are also given.

Estágio realizado na Critical Software, S. A. e orientado pelo Eng.º Marco António Manso Ventura e Nuno Alexandre Duro dos Santos
Tese de mestrado integrado. Engenharia Informátca e Computação. Faculdade de Engenharia. Universidade do Porto. 2008

Estágio realizado na Critical Software, S. A. e orientado pelo Eng.º Marco António Manso Ventura e Nuno Alexandre Duro dos Santos
Tese de mestrado integrado. Engenharia Informátca e Computação. Faculdade de Engenharia. Universidade do Porto. 2008

Unmanned Aerial Vehicles will have a safe access to the Civil Airspace only when they will be able to avoid collisions even with non cooperative flying obstacles. Thus, they need to replace the capability of human eye to detect potential mid-air collisions with other airframes and the pilot experience to find an adequate avoidance trajectory. This thesis deals with development and test of a fully autonomous system devoted to avoidance of non cooperative intruders. In particular, it focuses on sensors, and processing logics and hardware, required on the unmanned system to acquire situational awareness. The study was carried out in collaboration with the Italian Aerospace Research Center within a research project named TECVOL, funded in the frame of National Aerospace Research Program. The performed activities covered all the steps in the development process from the analysis of requirements deriving from the application, to the real time implementation of designed logics. Designed prototype system is based on a multi-sensor architecture with a Ka-band pulsed radar as the main sensor, and four electro-optical cameras as aiding sensors. Proper logics and algorithms for real time sensor fusion have been developed, tested in off-line simulations, and later implemented on embedded systems to enable technology flight demonstration. Numerical results and flight data have shown the potential of the developed system. Also on the basis of the international scenario, this technology demonstration has gained a significant scientific value.