Browsing by Author "Oh, Hyondong"

Now showing 1 - 9 of 9
  • Thumbnail Image
    ItemOpen Access
    Airborne behaviour monitoring using Gaussian processes with map information
    (Institution of Engineering and Technology, 2013-07-31T00:00:00Z) Oh, Hyondong; Shin, Hyosang; Kim, Seungkeun; Tsourdos, Antonios; White, Brian A.
    This paper proposes an airborne behaviour monitoring methodology of ground vehicles based on a statistical learning approach with domain knowledge given by road map information. To monitor and track the moving ground target using UAVs aboard a moving target indicator, an interactive multiple model (IMM) filter is firstly applied. {\color{red}The IMM filter consists of an on-road moving mode using a road-constrained filter and an off-road moving mode using a conventional filter.} Mode probability is also calculated from the IMM filter, and it provides deviation of the vehicle from the road. Then, a novel hybrid algorithm for anomalous behaviour recognition is developed using a Gaussian process regression on velocity profile along the one-dimensionalised position of the vehicle, as well as the deviation of the vehicle. To verify the feasibility and benefits of the proposed approach, a numerical simulation is performed using realistic car trajectory data in a city traffic.
  • Thumbnail Image
    ItemOpen Access
    Communication-aware convoy following guidance for UAVs in a complex urban environment
    (IEEE, 2016-06-30) Oh, Hyondong; Shin, Hyosang; Kim, Seungkeun; Ladosz, Pawel; Chen, Wen
    This paper proposes a communication-aware trajectory planning approach for UAVs to relay data/information (e.g. live surveillance feed) between a ground control station and friendly ground vehicles (a convoy) moving in a complex urban area. UAVs are controlled to stay: i) within the communication-feasible area (having a direct line-of-sight to the moving convoy and within the maximum communication range) and ii) as close as possible to the convoy to have better communication quality, while satisfying their kinematic and dynamic constraints. Numerical simulations and a proof-of-concept indoor flight test have been performed to validate the benefit and feasibility of the proposed algorithm.
  • Thumbnail Image
    ItemOpen Access
    Communication-aware trajectory planning for unmanned aerial vehicles in urban environments
    (AIAA, 2018-07-16) Oh, Hyondong; Shin, Hyosang; Kim, Seungkeun; Chen, Wen-Hua
    Introduction: Maintaining communication among mobile agents in a networked team is challenging due to limited bandwidth, maximum communication range, transmission power, and physical obscuration or occlusion in the mission environment. With the advent of lightweight, robust, and autonomous platforms as well as wireless networking technologies, it becomes feasible to use small unmanned aerial vehicles (UAVs) as communication relay nodes under limited satellite communication environments [1]. This communication relay UAV could allow a ground operator/system to have a sufficient data link to effectively see beyond the communication range and over the horizon/buildings where traditional methods fail. The relay UAV can also be used to transmit/share critical information efficiently from an operator to an end user or between vehicles.
  • Thumbnail Image
    ItemOpen Access
    Distributed estimation of stochastic multiagent systems for cooperative control with a virtual network
    (IEEE, 2022-10-19) Song, Yeongho; Lee, Hojin; Kwon, Cheolhyeon; Shin, Hyosang; Oh, Hyondong
    This article proposes a distributed estimation algorithm that uses local information about the neighbors through sensing or communication to design an estimation-based cooperative control of the stochastic multiagent system (MAS). The proposed distributed estimation algorithm solely relies on local sensing information rather than exchanging estimated state information from other agents, as is commonly required in conventional distributed estimation methods, reducing communication overhead. Furthermore, the proposed method allows interactions between all agents, including non-neighboring agents, by establishing a virtual fully connected network with the MAS state information independently estimated by each agent. The stability of the proposed distributed estimation algorithm is theoretically verified. Numerical simulations demonstrate the enhanced performance of the estimation-based linear and nonlinear control. In particular, using the virtual fully connected network concept in the MAS with the sensing/communication range, the flock configuration can be tightly controlled within the desired boundary, which cannot be achieved through the conventional flocking methods.
  • Thumbnail Image
    ItemOpen Access
    Receding horizon-based infotaxis with random sampling for source search and estimation in complex environments
    (IEEE, 2022-06-21) Park, Minkyu; Ladosz, Pawel; Kim, Jongyun; Oh, Hyondong
    This paper proposes a receding horizon-based information-theoretic source search and estimation strategy for a mobile sensor in an urban environment in which an invisible harmful substance is released into the atmosphere. The mobile sensor estimates the source term including its location and release rate by using sensor observations based on Bayesian inference. The sampling-based sequential Monte Carlo method, particle filter, is employed to estimate the source term state in a highly nonlinear and stochastic system. Infotaxis, the information-theoretic gradient-free search strategy is modified to find the optimal search path that maximizes the reduction of the entropy of the source term distribution. In particular, receding horizon Infotaxis is introduced to avoid falling into the local optima and to find more successful information gathering paths in obstacle-rich urban environments. Besides, a random sampling method is introduced to reduce the computational load of the receding horizon Infotaxis for real-time computation. The random sampling method samples the predicted future measurements based on current estimation of the source term and computes the optimal search path using sampled measurements rather than considering all possible future measurements. To demonstrate the benefit of the proposed approach, comprehensive numerical simulations are performed for various conditions. The proposed algorithm increases the success rate by about 30% and reduces the mean search time by about 40% compared with the existing information-theoretic search strategy.
  • Thumbnail Image
    ItemOpen Access
    Road-map-assisted standoff tracking of moving ground vehicle using nonlinear model predictive control
    (IEEE, 2015-04-30) Oh, Hyondong; Kim, Seungkeun; Tsourdos, Antonios
    This paper presents road-map-assisted standoff tracking of a ground vehicle using nonlinear model predictive control. In model predictive control, since the prediction of target movement plays an important role in tracking performance, this paper focuses on utilizing road-map information to enhance the estimation accuracy. For this, a practical road approximation algorithm is first proposed using constant curvature segments, and then nonlinear road-constrained Kalman filtering is followed. To address nonlinearity from road constraints and provide good estimation performance, both an extended Kalman filter and unscented Kalman filter are implemented along with the state-vector fusion technique for cooperative unmanned aerial vehicles. Lastly, nonlinear model predictive control standoff tracking guidance is given. To verify the feasibility and benefits of the proposed approach, numerical simulations are performed using realistic car trajectory data in city traffic.
  • Thumbnail Image
    ItemOpen Access
    Towards autonomous surveillance and tracking by multiple UAVs
    (Cranfield University, 2013-11) Oh, Hyondong; Tsourdos, Antonios
    This research investigates the use of small and low-cost UAVs (Unmanned Aerial Vehicles) for autonomous aerial surveillance, which aims to identify and continuously track suspicious vehicles and disguised threats in the ground traffic. Since typical ground traffic in an urban environment is quite dense and involves numerous vehicles, achieving this surveillance capability by a single mobile plat¬form is unlikely to be feasible in many aspects. In particular, due to physical constraints, it might be difficult for one UAV to cover large areas simultaneously, which is often critical to mission success in a rapidly changing environment. Be¬sides, in order to obtain accurate information of ground traffic, a single UAV platform will need to rely on sensors which are expensive yet vulnerable to the failure of the platform or sensing block by obstacles. Using multiple UAVs with relatively cheap aboard sensors with information fusion techniques enhancing sensing accuracy could resolve above issues of a single platform without signifi¬cantly increasing an operational cost. Therefore, this thesis deals with the surveillance application of multiple air¬borne sensor platforms endowed with an appropriate level of autonomous de¬cision making to support human operators. A group of UAVs become multiple mobile sensor platforms, and tasks/routes of each UAV need to be efficiently and optimally planned to cooperatively achieve mission objectives. Efficient and sophisticated algorithms for data acquisition/analysis, information fusion, path planning and formation reconfiguration ensuring feasible and safe cooperation, and decision making for cooperative missions are essentially to be developed, in order to take advantage of multiple aerial sensing sources for surveillance. Among various techniques for autonomous surveillance as listed above, this the¬sis seeks to develop and (partly) integrate some of important components: search route planning, behaviour identification/recognition, and moving target tracking, while examining benefits and drawbacks of using multiple UAVs. A particular focus is on multi-sensor management and information fusion in consideration of physical constraints of the platform and strict real-time requirements of the applications in uncertain and dynamic environments. This research investigates the use of small and low-cost UAVs (Unmanned Aerial Vehicles) for autonomous aerial surveillance, which aims to identify and continuously track suspicious vehicles and disguised threats in the ground traffic. Since typical ground traffic in an urban environment is quite dense and involves numerous vehicles, achieving this surveillance capability by a single mobile plat-form is unlikely to be feasible in many aspects. In particular, due to physical constraints, it might be difficult for one UAV to cover large areas simultaneously, which is often critical to mission success in a rapidly changing environment. Be-sides, in order to obtain accurate information of ground traffic, a single UAV platform will need to rely on sensors which are expensive yet vulnerable to the failure of the platform or sensing block by obstacles. Using multiple UAVs with relatively cheap aboard sensors with information fusion techniques enhancing sensing accuracy could resolve above issues of a single platform without signifi-cantly increasing an operational cost. Therefore, this thesis deals with the surveillance application of multiple air-borne sensor platforms endowed with an appropriate level of autonomous de-cision making to support human operators. A group of UAVs become multiple mobile sensor platforms, and tasks/routes of each UAV need to be efficiently and optimally planned to cooperatively achieve mission objectives. Efficient and sophisticated algorithms for data acquisition/analysis, information fusion, path planning and formation reconfiguration ensuring feasible and safe cooperation, and decision making for cooperative missions are essentially to be developed, in order to take advantage of multiple aerial sensing sources for surveillance. Among various techniques for autonomous surveillance as listed above, this the¬sis seeks to develop and (partly) integrate some of important components: search route planning, behaviour identification/recognition, and moving target tracking, while examining benefits and drawbacks of using multiple UAVs. A particular focus is on multi-sensor management and information fusion in consideration of physical constraints of the platform and strict real-time requirements of the applications in uncertain and dynamic environments. This thesis firstly proposes a road-network search planning algorithm by which UAVs are able to efficiently patrol every road identified in the map. A mixed integer linear programming problem (MILP) is formulated to find an optimal so¬lution minimising a total flight time, while accommodating physical constraints of the UAV with the Dubins path. To overcome the computational burden of the MILP, an approximation approach is also proposed. By running Monte Carlo sim¬ulation with the randomly generated maps, an efficient UAV team size and path planning method is examined. Secondly, this thesis proposes a behaviour recog¬nition methodology for ground vehicles moving within road traffic to identify abnormal behaviour. Ground vehicle behaviour is first classified into represen¬tative driving modes, and string pattern matching theory is applied to detect suspicious behaviours in the driving mode history. Moreover, a fuzzy decision making process is developed to systematically exploit all available information obtained from a complex environment considering spatiotemporal environment factors as well as several aspects of behaviours. Lastly, to achieve continuous tracking of detected suspicious vehicles for closer and higher-resolution surveil¬lance data, this thesis proposes several coordinated standoff tracking guidance algorithms using multiple UAVs. The effect of the improved target estimation accuracy on the tracking guidance performance is also examined using roadmap information and sensor fusion techniques. From this thesis, it can be identified that following aspects need to be carefully considered to realise autonomous surveillance using multiple UAVs: i) how many UAVs/sensors would be enough to perform a mission in terms of efficiency, es¬timation accuracy and guidance performance, ii) information gathered by UAVs only is enough, or domain knowledge (local context and past experience) might be additionally required, iii) communication structure between UAVs, and iv) com¬putation time. The proposed autonomous surveillance system utilising multiple UAVs is expected to greatly increase the amount of area that can be continuously monitored, while reducing the number of human operators and their workload required to analyse surveillance data and respond to identified targets.
  • Thumbnail Image
    ItemOpen Access
    Towards monocular vision-based autonomous flight through deep reinforcement learning
    (Elsevier, 2022-03-09) Kim, Minwoo; Kim, Jongyun; Jung, Minjae; Oh, Hyondong
    This paper proposes an obstacle avoidance strategy for small multi-rotor drones with a monocular camera using deep reinforcement learning. The proposed method is composed of two steps: depth estimation and navigation decision making. For the depth estimation step, a pre-trained depth estimation algorithm based on the convolutional neural network is used. On the navigation decision making step, a dueling double deep Q-network is employed with a well-designed reward function. The network is trained using the robot operating system and Gazebo simulation environment. To validate the performance and robustness of the proposed approach, simulations and real experiments have been carried out using a Parrot Bebop2 drone in various complex indoor environments. We demonstrate that the proposed algorithm successfully travels along the narrow corridors with the texture free walls, people, and boxes.
  • Thumbnail Image
    ItemOpen Access
    Using lazy agents to improve the flocking efficiency of multiple UAVs
    (Springer, 2021-10-27) Song, Yeongho; Gu, Myeonggeun; Choi, Joonwon; Oh, Hyondong; Lim, Seunghan; Shin, Hyosang; Tsourdos, Antonios
    A group of agents can form a flock using the augmented Cucker-Smale (C-S) model. The model autonomously aligns them to a common velocity and maintains a relative distance among the agents in a distributed manner by sharing the information among neighbors. This paper introduces the concept of inactiveness to the augmented C-S model for improving the flocking performance. It involves controlling the energy and convergence time required to form a stable flock. Inspired by the natural world where a few lazy (or inactive) workers are helpful to the group performance in social insect colonies. In this study, we analyzed different levels of inactiveness as a degree of control input effectiveness for multiple fixed-wing UAVs in the flocking algorithm. To find the appropriate inactiveness level for each flock member, the particle swarm optimization-based approach is used as the first step, based on the initial condition of the flock. However, as the significant computational burden may cause difficulties in implementing the optimization-based approach in real time, we also propose a heuristic adaptive inactiveness approach, which changes the inactivity level of selected agents adaptively according to their position and heading relative to the flock center. The performance of the proposed approaches using the concept of lazy (or inactive) agents is verified with numerical simulations by comparing them with the conventional flocking algorithm in various scenarios.