Дисертації з теми "Avoidance Map"

Le drone orienté grand public est principalement une caméra volante, stabilisée et de bonne qualité. Ceux-ci ont démocratisé la prise de vue aérienne, mais avec leur succès grandissant, la notion de sécurité est devenue prépondérante.Ce travail s'intéresse à l'évitement d'obstacle, tout en conservant un vol fluide pour l'utilisateur.Dans ce contexte technologique, nous utilisons seulement une camera stabilisée, par contrainte de poids et de coût.Pour leur efficacité connue en vision par ordinateur et leur performance avérée dans la résolution de tâches complexes, nous utilisons des réseaux de neurones convolutionnels (CNN). Notre stratégie repose sur un systeme de plusieurs niveaux de complexité dont les premieres étapes sont de mesurer une carte de profondeur depuis la caméra. Cette thèse étudie les capacités d'un CNN à effectuer cette tâche.La carte de profondeur, étant particulièrement liée au flot optique dans le cas d'images stabilisées, nous adaptons un réseau connu pour cette tâche, FlowNet, afin qu'il calcule directement la carte de profondeur à partir de deux images stabilisées. Ce réseau est appelé DepthNet.Cette méthode fonctionne en simulateur avec un entraînement supervisé, mais n'est pas assez robuste pour des vidéos réelles. Nous étudions alors les possibilites d'auto-apprentissage basées sur la reprojection différentiable d'images. Cette technique est particulièrement nouvelle sur les CNNs et nécessite une étude détaillée afin de ne pas dépendre de paramètres heuristiques.Finalement, nous développons un algorithme de fusion de cartes de profondeurs pour utiliser DepthNet sur des vidéos réelles. Plusieurs paires différentes sont données à DepthNet afin d'avoir une grande plage de profondeurs mesurées
Customer unmanned aerial vehicles (UAVs) are mainly flying cameras. They democratized aerial footage, but with thei success came security concerns.This works aims at improving UAVs security with obstacle avoidance, while keeping a smooth flight. In this context, we use only one stabilized camera, because of weight and cost incentives.For their robustness in computer vision and thei capacity to solve complex tasks, we chose to use convolutional neural networks (CNN). Our strategy is based on incrementally learning tasks with increasing complexity which first steps are to construct a depth map from the stabilized camera. This thesis is focused on studying ability of CNNs to train for this task.In the case of stabilized footage, the depth map is closely linked to optical flow. We thus adapt FlowNet, a CNN known for optical flow, to output directly depth from two stabilized frames. This network is called DepthNet.This experiment succeeded with synthetic footage, but is not robust enough to be used directly on real videos. Consequently, we consider self supervised training with real videos, based on differentiably reproject images. This training method for CNNs being rather novel in literature, a thorough study is needed in order not to depend too moch on heuristics.Finally, we developed a depth fusion algorithm to use DepthNet efficiently on real videos. Multiple frame pairs are fed to DepthNet to get a great depth sensing range

This work focuses on the development and implementation of an autonomous path planning and obstacle avoidance algorithm for an autonomous surface vehicle (ASV) in a riverine environment. The algorithm effectively handles trap situations, which occur when the river bends away from the destination. In addition, the algorithm uses real-time sensor feedback to avoid obstacles. A general global route is proposed based on an a priori shoreline map. Then, local paths are calculated considering both the a priori data and measurements received from an obstacle sensor. These paths roughly follow the global path. The algorithm was tested on an ASV equipped with basic navigational sensors and an omnidirectional camera for obstacle detection, and experimentation verified its effectiveness.
Master of Science

Aquesta tesi suggereix usar un compte enrere determinista després de les transmissions exitoses en les xarxes d'àrea local sense fils. Com que les estacions que han transmès amb èxit en el seu darrer intent no poden col·lisionar entre elles en el seu proper intent, el nombre de col·lisions es redueix. Per tant, anomenem el protocol accés múltiple per detecció de portadora amb evitament de col·lisions millorat. El protocol es modela i es simula en diversos escenaris per a la seva avaluació. S'arriba a la conclusió de que el protocol proposat ofereix un rendiment igual o major que l'existent.
This thesis suggests the use of a deterministic backoff after successful transmissions in the MAC protocol of WLANs. Since those stations that successfully transmitted in their last collision attempt cannot collide among them in their next transmission attempt, the number of collisions is reduced. Hence, the protocol is called Carrier Sense Multiple Access with Enhanced Collision Avoidance. The protocol is modelled and simulations are used to assess its performance in a variety of scenarios. It is concluded that the proposed protocol performs always equal or better than the legacy one.

Com o aumento na disponibilidade e considerável redução de custo das tecnologias que permitem a obtenção de dados de objetos móveis, torna-se abundante a oferta de dados de trajetórias. O estudo dessas trajetórias tem por objetivo permitir uma melhor compreensão dos dados, bem como a extração de novo conhecimento utilizando técnicas computacionais. Um tipo de estudo que pode ser realizado é a análise do comportamento das trajetórias de objetos móveis. O foco do estudo de padrões comportamentais em trajetórias tem sido a busca por padrões de aglomeração ou semelhança no deslocamento de entidades no espaço ou espaço-tempo. Este trabalho propõe a formalização de um novo padrão comportamental que indique quando um objeto móvel está evitando determinadas regiões espaciais. Este padrão foi denominado avoidance. A identificação e o estudo de tal comportamento pode ser de interesse de diversas áreas tais como segurança, jogos eletrônicos, comportamento social, entre outros. Inicialmente, este trabalho apresenta uma heurística com base nas observações das ocorrências do evento avoidance. A partir daí, são definidas as formalizações deste novo padrão. É apresentado, também, um algoritmo para identificar automaticamente o padrão comportamental avoidance em trajetórias de objetos móveis. Para avaliar a eficácia deste algoritmo, ao final desta dissertação são apresentados resultados de experimentos realizados em trajetórias coletadas por pedestres e por veículos em diversos locais da cidade de Porto Alegre e Xangri-lá.
With the increasing availability and considerable price reduction of technologies that allow the collection of moving object data, the offering of trajectory data becomes abundant. The study of these trajectories has the objective to allow a better understanding of the data, as well as the extraction of new knowledge using computational techniques. One of the possible studies is the analysis of the behavior of moving objects. The study on behavior patterns in trajectories has focused on agglomeration and similarity of entities moving in space or space and time. This work proposes the formalization of a new behavior pattern that indicates when a moving object is avoiding some determined spatial region, here called avoidance. The study and identification of this behavior may be of interest in many application areas, like security, electronic games, social behavior, and so on. Initially, this work presents an heuristic based on the observations of avoidance events, and based on this heuristic we define the formalization of this new kind of trajectory behavior pattern. This work also presents an algorithm to automatically identify this behavior in trajectories of moving objects. To evaluate the effectiveness of this algorithm, experiments were performed on trajectory datasets, collected by pedestrians in a park, and by vehicles at several places in the cities of Porto Alegre and Xangri-lá.

Thesis (M.A.)--City University of Hong Kong, 2006.
"City University of Hong Kong, School of Law, LW6409 Dissertation" Title from PDF t.p. (viewed on 29 Jun, 2007) Includes bibliographical references.

Denna rapport behandlar prestandaproblem i den nyligen standardiserade Mesh kommunikationsstandarden (IEEE 802.11s). I denna rapport, undersöker och förbättra vi ett förhållande som resulterar i reduktion av genomströmningen i en kedja av noder topologi. IEEE802.11s är mycket lovande med många fördelar för både IoT-systemen och trådlösa nätverk i båda hemmet och arbete.Vi arbetar med frågan om orättvisa när CSMA/CA tillämpas, vilket orsakar genomströmningsreduktion på grund av paketförluster och indikerar svältning. Vi analyserar konsekvenserna av Collision Avoidance (CA) mekanism och föreslår en ersättning för CA som är både rättvist och samtidigt kan upprätthålla undvikande av kollisioner. Vi implementera detta i en simulator och resultatet visar på betydligt högre end-to-end-genomströmning än standard CSMA/CA och inga paketförluster på grund av buffertspill.
This thesis rapport deals with the performance issues of the newly standardized Wireless mesh protocol (IEEE 802.11s). In this thesis, we work on improving the conditions that results in throughput degradation in a chain of nodes topology. The mesh standard is very promising with many advantages for both IoT systems and home wireless networks.We work on the issue of unfairness when CSMA/CA is applied, which causes throughput degradation due to packet loss and indicates starvation. We analyze the implication of the Collision Avoidance (CA) mechanism and propose a replacement for the CA that is both fair and able to maintain collision avoidance. We implement this in a simulator and the result shows significantly higher end-to-end throughput compared to the original CSMA/CA and no packet loss due to buffer overflow.

碩士
國立臺北科技大學
電機工程系碩士班
92
In order to control the automatic guided vehicle (AGV) following a preset target and avoiding obstacle along the process, a fuzzy attraction force regulator and a fuzzy repulsive force regulator are proposed to control the AGV moving around the obstacle until attaining the preset target. A fuzzy attraction force regulator applies the heading orientation of mobile robot and target angle to provide attraction angle to achieve a preset target. A map building scheme utilizing fuzzy logic is applied in a fuzzy repulsive force regulator. The location of obstacle along the route is identified through map building. According to the estimated obstacle location, a repulsive density function and a fuzzy inference method of sonar range are cooperated to determine the obstacle avoiding direction. Due to the fact that sonar model and map building are mixed and incomplete, it will cause noise on the edge of the fuzzy map. We present fuzzy filter and mixed filtering method to improve map building.

碩士
國立成功大學
航空太空工程學系
87
An avionic flight-warning system is one of the key factors in avionics safety. The traditional systems are relied on the position reports from either incoming flight or the radar detection. These are basically in the field of point-by-point computation. This study adopts a volume search approach with the aids from the digital terrain map (DTM). Due to a lot of civil avionic accidents occurred in mountain clash, the research scenario is set to be low attitude, low visibility, and rapid change in the circumstance. The main goal of this study is to design a flight warning system with 3D capability in which the positions of the studied vehicle are derived from GPS with the terrain information from the aerial photographs. The designed system has the process procedures including data extraction, distance computation, and graphic display. The whole process time has to be within an acceptable time interval. The designed system is tested by a simulated helicopter flight with using a true DTM data. Three data extraction approaches are adopted: full, static, and dynamic extractions. The dynamic extraction provides the best result with the total computation time under 0.09 seconds with 2-meter resolution. When the resolution requirement rise up to 8-meter, the computation time is within 0.006 seconds. Three flight conditions are also considered. The air speed is the main factor in the consideration. There are three levels: 0~80, 80~160, and > 160 km/hr. The designed response time is set to be 5 seconds. The system provides different resolutions of DTM according to various velocities.

This thesis discusses occupancy grid map, collision avoidance system and belief theory, and propose some of the latest and the most effective method such as predictive occupancy grid map, risk evaluation model and OGM role in the belief function theory with the approach of decision uncertainty according to the environment perception with the degree of belief in the driving command acceptability. Finally, how the proposed models mitigate or prevent the occurrence of the collision.

Indiana University-Purdue University Indianapolis (IUPUI)
This thesis discusses occupancy grid map, collision avoidance system and belief theory, and propose some of the latest and the most effective method such as predictive occupancy grid map, risk evaluation model and OGM role in the belief function theory with the approach of decision uncertainty according to the environment perception with the degree of belief in the driving command acceptability. Finally, how the proposed models mitigate or prevent the occurrence of the collision.

碩士
長庚大學
機械工程學系
98
The purpose of this thesis is to develop an omni-directional mobile platform for the advance of intelligent vehicles. The system integrates 2-dimensional map construction technologies with obstacle avoidance rules to enable the vehicle to construct geometric model of unknown environment without bumping into obstacles. The system utilizes URG-04LX laser range finder、provided by Hokuyo Inc.、to detect distances between the sensor and surrounding objects. The scanned dada are transmitted through wireless communication to a console PC to construct global map in real time. The map offers a remote operator a clear understanding of the environment surrounding the vehicle. As global map is constructed gradually from the corresponding matching between local maps、the downhill simplex optimization algorithm is used to find the best match. Furthermore、the distance transform are applied to maps to provide gradients for cost value evaluation. Experimental results show that the tasks of 2-D map construction and obstacle avoidance can be achieved without encoder if the environment includes only flat ground with asymmetrical obstacles or walls.

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
國立臺北大學
通訊工程研究所
99
Underwater Sensor Networks (UWSNs) got a lot of attention recently because of Global Climate Change. Communication carrier of UWSNs is acoustic wave, so propagation speed of acoustic signals in underwater environments is much lower than Radio Frequency (RF) propagation in terrestrial environment. Underwater sensors are also affected by ocean current, sensors are moving by push generated from differences of temperature and density. Propagation delay is varying as a result of underwater sensors affected by ocean current. The propagation delay of UWSNs is not fixed so data transmission scheme in UWSNs is more difficult. In this paper, we propose a collision avoidance MAC called Mobi-MAC protocol for mobile UWSNs. The aim of this protocol is designed to overcome varying propagation delay which generates unpredictable data transmission collision. A data transmission scheduling based on TDMA-based MAC that improves network throughput of UWSNs proposed. All different moving patterns are found via researching the moving state between any two sensor nodes. Different moving state gets different relative velocity that is a main factor of propagation delay variation. Then nodes can schedule data transmission according to propagation delay variation, and then reduce the data collision probability by the prime-hopping (PH) data scheduling proposed in this paper. Finally, compare performance with existing MAC protocol for UWSNs via simulations and then results show Mobi-MAC protocol can achieve better than existing MAC protocol in terms of the network throughput.

碩士
逢甲大學
資訊工程學系
102
IEEE 802.11 DCF is widely used and able to avoid the collisions of the hidden terminal problem in a wireless network. However, there may be collisions existing in the DCF MAC protocol. That is, when a node in NAV state does not receive any control signal, a collision may occur after the node ends the NAV period. In this paper, to solve this collision problem, a detailed analysis about the problem is given. Afterwards, according to the analysis, we utilize the “mask-ing effect” and the detection of signal variation to design the MAC collision avoidance protocol. With the received signal strength and the varia-tion of received signal strength of a node, the node can determine whether there are other transmissions occurring in the transmission range. Therefore, a node can determine whether the period of NAV state can be extended to avoid collisions. Simulation results show that our pro-posed MAC protocol can efficiently reduce 60% and 41% collisions than the IEEE 802.11 DCF protocol and OCS-MD. In addition, our pro-posed MAC protocol also shows shorter trans-mission delay and lower retransmissions than the 802.11 DCF and OCS-MD to have better network throughput.

碩士
淡江大學
資訊工程學系資訊網路與通訊碩士班
100
Wireless technology has been developed and used for many years. Due to the nature of water, instead of radio wave, sound wave is used for underwater transmission. However, it will waste a lot of time for propagation at every transmission. Recently, IEEE 802.11 DCF is the most famous MAC protocol but is not suitable for underwater scenarios. We figure out the attenuation of acoustic in underwater and interference problem through our analysis. Therefore, this proposal mainly focuses on how to modify the MAC protocol for underwater acoustic network. It can avoid collisions by power control when sending control packets in our proposed protocol. In simulation, our protocol can outperform than other protocols in network throughput and number of collisions.