Current Search: Xu, Yunjun (x)
View All Items
(41 - 50 of 50)
Pages
- Title
- Interactive Perception in Robotics.
- Creator
-
Baghbahari Baghdadabad, Masoud, Behal, Aman, Haralambous, Michael, Lin, Mingjie, Sukthankar, Gita, Xu, Yunjun, University of Central Florida
- Abstract / Description
-
Interactive perception is a significant and unique characteristic of embodied agents. An agent can discover plenty of knowledge through active interaction with its surrounding environment. Recently, deep learning structures introduced new possibilities to interactive perception in robotics. The advantage of deep learning is in acquiring self-organizing features from gathered data; however,it is computationally impractical to implement in real-time interaction applications. Moreover, it can be...
Show moreInteractive perception is a significant and unique characteristic of embodied agents. An agent can discover plenty of knowledge through active interaction with its surrounding environment. Recently, deep learning structures introduced new possibilities to interactive perception in robotics. The advantage of deep learning is in acquiring self-organizing features from gathered data; however,it is computationally impractical to implement in real-time interaction applications. Moreover, it can be difficult to attach a physical interpretation. An alternative suggested framework in such cases is integrated perception-action.In this dissertation, we propose two integrated interactive perception-action algorithms for real-time automated grasping of novel objects using pure tactile sensing. While visual sensing andprocessing is necessary for gross reaching movements, it can slow down the grasping process if it is the only sensing modality utilized. To overcome this issue, humans primarily utilize tactile perceptiononce the hand is in contact with the object. Inspired by this, we first propose an algorithm to define similar ability for a robot by formulating the required grasping steps.Next, we develop the algorithm to achieve force closure constraint via suggesting a human-like behavior for the robot to interactively identify the object. During this process, the robot adjuststhe hand through an interactive exploration of the object's local surface normal vector. After the robot finds the surface normal vector, it then tries to find the object edges to have a graspable finalrendezvous with the object. Such achievement is very important in order to find the objects edges for rectangular objects before fully grasping the object. We implement the proposed approacheson an assistive robot to demonstrate the performance of interactive perception-action strategies to accomplish grasping task in an automatic manner.
Show less - Date Issued
- 2019
- Identifier
- CFE0007780, ucf:52361
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0007780
- Title
- Color-Ratio Based Strawberry Plant Localization and Nutrition Deficiency Detection.
- Creator
-
Kong, Xiangling, Xu, Yunjun, Elgohary, Tarek, Fu, Qiushi, Wu, Dazhong, Wang, Liqiang, University of Central Florida
- Abstract / Description
-
In recent years, precision agriculture has become popular anticipating to partially meet the needs of an ever-growing population with limited resources. Plant localization and nutrient de?ciency detection are two important tasks in precision agriculture. In this dissertation, these two tasks are studied by using a new color-ratio(C-R) index technique. Firstly, a low cost and light scene invariant approach is proposed to detect green and yellow leaves based on the color-ratio (C-R) indices. A...
Show moreIn recent years, precision agriculture has become popular anticipating to partially meet the needs of an ever-growing population with limited resources. Plant localization and nutrient de?ciency detection are two important tasks in precision agriculture. In this dissertation, these two tasks are studied by using a new color-ratio(C-R) index technique. Firstly, a low cost and light scene invariant approach is proposed to detect green and yellow leaves based on the color-ratio (C-R) indices. A plant localization approach is then developed using the relative pixel relationships of adjacent plants. Secondly, the Sobel operator and morphology techniques are applied to segment the target strawberry leaf from a ?eld image. The characterized color for a speci?c nutrient de?ciency is detected by the C-R indices. The pattern of the detected color on the leaf is then examined to determine the speci?c nutrient de?ciency. The proposed approaches are validated in a commercial strawberry farm.
Show less - Date Issued
- 2019
- Identifier
- CFE0007666, ucf:52482
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0007666
- Title
- Processing, Characterization and Performance of Carbon Nanopaper Based Multifunctional Nanocomposites.
- Creator
-
Liang, Fei, Gou, Jihua, Su, Ming, Fang, Jiyu, Orlovskaya, Nina, Xu, Yunjun, University of Central Florida
- Abstract / Description
-
Carbon nanofibers (CNFs) used as nano-scale reinforcement have been extensively studied since they are capable of improving the physical and mechanical properties of conventional fiber reinforced polymer composites. However, the properties of CNFs are far away from being fully utilized in the composites due to processing challenges including the dispersion of CNFs and the viscosity increase of polymer matrix. To overcome these issues, a unique approach was developed by making carbon nanopaper...
Show moreCarbon nanofibers (CNFs) used as nano-scale reinforcement have been extensively studied since they are capable of improving the physical and mechanical properties of conventional fiber reinforced polymer composites. However, the properties of CNFs are far away from being fully utilized in the composites due to processing challenges including the dispersion of CNFs and the viscosity increase of polymer matrix. To overcome these issues, a unique approach was developed by making carbon nanopaper sheet through the filtration of well-dispersed carbon nanofibers under controlled processing conditions, and integrating carbon nanopaper sheets into composite laminates using autoclave process and resin transfer molding (RTM). This research aims to fundamentally study the processing-structure-property-performance relationship of carbon nanopaper-based nanocomposites multifunctional applications: a) Vibrational damping. Carbon nanofibers with extremely high aspect ratios and low density present an ideal candidate as vibrational damping material; specifically, the large specific area and aspect ratio of carbon nanofibers promote significant interfacial friction between carbon nanofiber and polymer matrix, causing higher energy dissipation in the matrix. Polymer composites with the reinforcement of carbon nanofibers in the form of a paper sheet have shown significant vibration damping improvement with a damping ratio increase of 300% in the nanocomposites. b) Wear resistance. In response to the observed increase in toughness of the nanocomposites, tribological properties of the nanocomposite coated with carbon nanofiber/ceramic particles hybrid paper have been studied. Due to high strength and toughness, carbon nanofibers can act as microcrack reducer; additionally, the composites coated with such hybrid nanopaper of carbon nanofiber and ceramic particles shown an improvement of reducing coefficient of friction (COF) and wear rate. c) High electrical conductivity. A highly conductive coating material was developed and applied on the surface of the composites for the electromagnetic interference shielding and lightning strike protection. To increase the conductivity of the carbon nanofiber paper, carbon nanofibers were modified with nickel nanostrands. d) Electrical actuation of SMP composites. Compared with other methods of SMP actuation, the use of electricity to induce the shape-memory effect of SMP is desirable due to the controllability and effectiveness. The electrical conductivity of carbon fiber reinforced SMP composites can be significantly improved by incorporating CNFs and CNF paper into them. A vision-based system was designed to control the deflection angle of SMP composites to desired values. The funding support from National Science Foundation and FAA Center of Excellence for Commercial Space Transportation (FAA COE CST) is acknowledged.
Show less - Date Issued
- 2012
- Identifier
- CFE0004569, ucf:49194
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0004569
- Title
- Virtual Motion Camouflage Based Nonlinear Constrained Optimal Trajectory Design Method.
- Creator
-
Basset, Gareth, Xu, Yunjun, Kassab, Alain, Lin, Kuo-Chi, Cho, Hyoung, Qu, Zhihua, University of Central Florida
- Abstract / Description
-
Nonlinear constrained optimal trajectory control is an important and fundamental area of research that continues to advance in numerous fields. Many attempts have been made to present new methods that can solve for optimal trajectories more efficiently or to improve the overall performance of existing techniques. This research presents a recently developed bio-inspired method called the Virtual Motion Camouflage (VMC) method that offers a means of quickly finding, within a defined but varying...
Show moreNonlinear constrained optimal trajectory control is an important and fundamental area of research that continues to advance in numerous fields. Many attempts have been made to present new methods that can solve for optimal trajectories more efficiently or to improve the overall performance of existing techniques. This research presents a recently developed bio-inspired method called the Virtual Motion Camouflage (VMC) method that offers a means of quickly finding, within a defined but varying search space, the optimal trajectory that is equal or close to the optimal solution.The research starts with the polynomial-based VMC method, which works within a search space that is defined by a selected and fixed polynomial type virtual prey motion. Next will be presented a means of improving the solution's optimality by using a sequential based form of VMC, where the search space is adjusted by adjusting the polynomial prey trajectory after a solution is obtained. After the search space is adjusted, an optimization is performed in the new search space to find a solution closer to the global space optimal solution, and further adjustments are made as desired. Finally, a B-spline augmented VMC method is presented, in which a B-spline curve represents the prey motion and will allow the search space to be optimized together with the solution trajectory.It is shown that (1) the polynomial based VMC method will significantly reduce the overall problem dimension, which in practice will significantly reduce the computational cost associated with solving nonlinear constrained optimal trajectory problems; (2) the sequential VMC method will improve the solution optimality by sequentially refining certain parameters, such as the prey motion; and (3) the B-spline augmented VMC method will improve the solution optimality without sacrificing the CPU time much as compared with the polynomial based approach. Several simulation scenarios, including the Breakwell problem, the phantom track problem, the minimum-time mobile robot obstacle avoidance problem, and the Snell's river problem are simulated to demonstrate the capabilities of the various forms of the VMC algorithm. The capabilities of the B-spline augmented VMC method are also shown in a hardware demonstration using a mobile robot obstacle avoidance testbed.
Show less - Date Issued
- 2012
- Identifier
- CFE0004298, ucf:49493
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0004298
- Title
- Kinematical Modelling and Its Analytical Inverse Kinematic Solution for the Handling Mechanism of an Agricultural Robot.
- Creator
-
Defterli, Sinem, Xu, Yunjun, Lin, Kuo-Chi, Zheng, Qipeng, Song, Sang-Eun, University of Central Florida
- Abstract / Description
-
Early detection of the crop diseases helps to prevent failure in the amount and the quality of the production. In agricultural robotics, the idea of a disease detection robot is a fresh and an innovative hot-button topic. The exclusion of the diseased parts from the strawberry plants for further analyses is one of the main tasks of a recently developed strawberry robot. To this purpose, the handling mechanism in the robot needs to achieve an accurate manipulation task to reach the target....
Show moreEarly detection of the crop diseases helps to prevent failure in the amount and the quality of the production. In agricultural robotics, the idea of a disease detection robot is a fresh and an innovative hot-button topic. The exclusion of the diseased parts from the strawberry plants for further analyses is one of the main tasks of a recently developed strawberry robot. To this purpose, the handling mechanism in the robot needs to achieve an accurate manipulation task to reach the target. Reaching, cutting and storing the diseased leaf are challenging and delicate processes during the procedure of the handling mechanism operation in the field.The manipulation task of the mechanism is succeeded when the inverse kinematic relations from workspace to joint space are defined properly. The inverse kinematic analysis is usually subjected to the restrictions due to the limitations in mechanical design of the mechanism, hardware components and operation environment of the robots as well as the morphology of the target. This study proposes a set of analytical algorithms to solve the inverse kinematics problem of the handling mechanism under certain constraints. First, proposed analytical approach is based on the calculation of the joint variables by solving only the 3D position information of the target since the output from image processing algorithms of vision subsystem in the ground robot is only the location of the diseased point. The position of target point is the only output from vision subsystem and this data will be given as an input to the proposed algorithms. Second, the mechanism has certain restrictions on its geometrical construction and the joint actuators' capacity. Hence, these restrictions limit the range of joint variables to be solved. Due to sudden and unpredicted nature of field conditions, the quickness of handling mechanism inverse kinematics solution's execution has a vital effect on the success of the picking task of the robot. Another essential factor is to use the battery life of the robot effectively, by minimizing energy consumption. Therefore, the effectiveness of the proposed algorithm is decided by comparing the developed performance indices of consumed energy and CPU time cost via numerical solution namely, a nonlinear constrained optimization method under same restrictions of inverse kinematics problem. Performance of both algorithms is observed by the simulations in MATLAB(&)#174; and laboratory set-up experiments.
Show less - Date Issued
- 2016
- Identifier
- CFE0006291, ucf:51588
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0006291
- Title
- Six Degree of Freedom Dynamic Modeling of a High Altitude Airship and Its Trajectory Optimization Using Direct Collocation Method.
- Creator
-
Pierre-Louis, Pradens, Xu, Yunjun, Lin, Kuo-Chi, Das, Tuhin, University of Central Florida
- Abstract / Description
-
The long duration airborne feature of airships makes them an attractive solution for many military and civil applications such as long-endurance surveillance, reconnaissance, environment monitoring, communication utilities, and energy harvesting. To achieve a minimum energy periodic motion in the air, an optimal trajectory problem is solved using basic direct collocation methods. In the direct approach, the optimal control problem is converted into a nonlinear programming (NLP). Pseudo...
Show moreThe long duration airborne feature of airships makes them an attractive solution for many military and civil applications such as long-endurance surveillance, reconnaissance, environment monitoring, communication utilities, and energy harvesting. To achieve a minimum energy periodic motion in the air, an optimal trajectory problem is solved using basic direct collocation methods. In the direct approach, the optimal control problem is converted into a nonlinear programming (NLP). Pseudo-inverse and several discretization methods such as Trapezoidal and Hermite-Simpson are used to obtain a numerical approximated solution by discretizing the states and controls into a set of equal nodes. These nodes are approximated by a cubic polynomial function which makes it easier for the optimization to converge while ensuring the problem constraints and the equations of motion are satisfied at the collocation points for a defined trajectory. In this study, direct collocation method provides the ability to obtain an approximation solution of the minimum energy expenditure of a very complex dynamic problem using Matlab fmincon optimization algorithm without using Himiltonian function with Lagrange multipliers. The minimal energy trajectory of the airship is discussed and results are presented.
Show less - Date Issued
- 2017
- Identifier
- CFE0006779, ucf:51822
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0006779
- Title
- Bio-inspired, Varying Manifold Based Method with Enhanced Initial Guess Strategies for Single Vehicle's Optimal Trajectory Planning.
- Creator
-
Li, Ni, Xu, Yunjun, Lin, Kuo-Chi, Bai, Yuanli, Behal, Aman, University of Central Florida
- Abstract / Description
-
Trajectory planning is important in many applications involving unmanned aerial vehicles, underwater vehicles, spacecraft, and industrial manipulators. It is still a challenging task to rapidly find an optimal trajectory while taking into account dynamic and environmental constraints. In this dissertation, a unified, varying manifold based optimal trajectory planning method inspired by several predator-prey relationships is investigated to tackle this challenging problem. Biological species,...
Show moreTrajectory planning is important in many applications involving unmanned aerial vehicles, underwater vehicles, spacecraft, and industrial manipulators. It is still a challenging task to rapidly find an optimal trajectory while taking into account dynamic and environmental constraints. In this dissertation, a unified, varying manifold based optimal trajectory planning method inspired by several predator-prey relationships is investigated to tackle this challenging problem. Biological species, such as hoverflies, ants, and bats, have developed many efficient hunting strategies. It is hypothesized that these types of predators only move along paths in a carefully selected manifold based on the prey's motion in some of their hunting activities. Inspired by these studies, the predator-prey relationships are organized into a unified form and incorporated into the trajectory optimization formulation, which can reduce the computational cost in solving nonlinear constrained optimal trajectory planning problems. Specifically, three motion strategies are studied in this dissertation: motion camouflage, constant absolute target direction, and local pursuit. Necessary conditions based on the speed and obstacle avoidance constraints are derived. Strategies to tune initial guesses are proposed based on these necessary conditions to enhance the convergence rate and reduce the computational cost of the motion camouflage inspired strategy. The following simulations have been conducted to show the advantages of the proposed methods: a supersonic aircraft minimum-time-to-climb problem, a ground robot obstacle avoidance problem, and a micro air vehicle minimum time trajectory problem. The results show that the proposed methods can find the optimal solution with higher success rate and faster convergent speed as compared with some other popular methods. Among these three motion strategies, the method based on the local pursuit strategy has a relatively higher success rate when compared to the other two.In addition, the optimal trajectory planning method is embedded into a receding horizon framework with unknown parameters updated in each planning horizon using an Extended Kalman Filter.
Show less - Date Issued
- 2013
- Identifier
- CFE0005023, ucf:49986
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0005023
- Title
- bio-inspired attitude control of micro air vehicles using rich information from airflow sensors.
- Creator
-
Shen, He, Xu, Yunjun, Lin, Kuo-Chi, Kauffman, Jeffrey, An, Linan, University of Central Florida
- Abstract / Description
-
Biological phenomena found in nature can be learned and customized to obtain innovative engineering solutions. In recent years, biologists found that birds and bats use their mechanoreceptors to sense the airflow information and use this information directly to achieve their agile flight performance. Inspired by this phenomenon, an attitude control system for micro air vehicles using rich amount of airflow sensor information is proposed, designed and tested. The dissertation discusses our...
Show moreBiological phenomena found in nature can be learned and customized to obtain innovative engineering solutions. In recent years, biologists found that birds and bats use their mechanoreceptors to sense the airflow information and use this information directly to achieve their agile flight performance. Inspired by this phenomenon, an attitude control system for micro air vehicles using rich amount of airflow sensor information is proposed, designed and tested. The dissertation discusses our research findings on this topic. First, we quantified the errors between the calculated and measured lift and moment profiles using a limited number of micro pressure sensors over a straight wing. Then, we designed a robust pitching controller using 20 micro pressure sensors and tested the closed-loop performance in a simulated environment. Additionally, a straight wing was designed for the pressure sensor based pitching control with twelve pressure sensors, which was then tested in our low-speed wind tunnel. The closed-loop pitching control system can track the commanded angle of attack with a rising time around two seconds and an overshoot around 10%. Third, we extended the idea to the three-axis attitude control scenarios, where both of the pressure and shear stress information are considered in the simulation. Finally, a fault tolerant controller with a guaranteed asymptotically stability is proposed to deal with sensor failures and calculation errors. The results show that the proposed fault tolerant controller is robust, adaptive, and can guarantee an asymptotically stable performance even in case that 50% of the airflow sensors fail in flight.
Show less - Date Issued
- 2014
- Identifier
- CFE0005711, ucf:50150
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0005711
- Title
- Lyapunov-Based Robust and Adaptive Control Design for nonlinear Uncertain Systems.
- Creator
-
Zhang, Kun, Behal, Aman, Haralambous, Michael, Xu, Yunjun, Boloni, Ladislau, Marzocca, Piergiovanni, University of Central Florida
- Abstract / Description
-
The control of systems with uncertain nonlinear dynamics is an important field of control scienceattracting decades of focus. In this dissertation, four different control strategies are presentedusing sliding mode control, adaptive control, dynamic compensation, and neural network for a nonlinear aeroelastic system with bounded uncertainties and external disturbance. In Chapter 2, partial state feedback adaptive control designs are proposed for two different aeroelastic systems operating in...
Show moreThe control of systems with uncertain nonlinear dynamics is an important field of control scienceattracting decades of focus. In this dissertation, four different control strategies are presentedusing sliding mode control, adaptive control, dynamic compensation, and neural network for a nonlinear aeroelastic system with bounded uncertainties and external disturbance. In Chapter 2, partial state feedback adaptive control designs are proposed for two different aeroelastic systems operating in unsteady flow. In Chapter 3, a continuous robust control design is proposed for a class of single input and single output system with uncertainties. An aeroelastic system with a trailingedge flap as its control input will be considered as the plant for demonstration of effectiveness of the controller. The controller is proved to be robust by both athematical proof and simulation results. In Chapter 3, a robust output feedback control strategy is discussed for the vibration suppression of an aeroelastic system operating in an unsteady incompressible flowfield. The aeroelastic system is actuated using a combination of leading-edge (LE) and trailing-edge (TE) flaps in the presence of different kinds of gust disturbances. In Chapter 5, a neural-network based model-free controller is designed for an aeroelastic system operating at supersonic speed. The controller is shown to be able to effectively asymptotically stabilize the system via both a Lyapunov-based stability proof and numerical simulation results.
Show less - Date Issued
- 2015
- Identifier
- CFE0005748, ucf:50110
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0005748
- Title
- Bio-Inspired Visual Servo Control of a Picking Mechanism in an Agricultural Ground Robot.
- Creator
-
Defterli, Sinem, Xu, Yunjun, Kauffman, Jeffrey L., Lin, Kuo-Chi, Song, Sang-Eun, Zheng, Qipeng, University of Central Florida
- Abstract / Description
-
For a recently constructed disease detection agricultural ground robot, the segregation of unhealthy leaves fromstrawberry plants is a major task of the robot's manipulation subsystem in field operations. In this dissertation, the motion planning of a custom-designedpicking mechanism in the ground robot's subsystem is studied in two sections. First, a set of analytical, suboptimal semi-analyticaland numerical algorithms are studied to solve the inverse kinematics problem of the handling...
Show moreFor a recently constructed disease detection agricultural ground robot, the segregation of unhealthy leaves fromstrawberry plants is a major task of the robot's manipulation subsystem in field operations. In this dissertation, the motion planning of a custom-designedpicking mechanism in the ground robot's subsystem is studied in two sections. First, a set of analytical, suboptimal semi-analyticaland numerical algorithms are studied to solve the inverse kinematics problem of the handling mechanism in firmcircumstances. These premeditated approaches are built on the computation of the joint variables by an identified 3Dposition data of the target leaf only. The outcomes of the three solution algorithms are evaluated in terms of the performanceindexes of energy change and the CPU time cost. The resultant postures of the mechanism for different target pointlocations are observed both in simulations and the hardware experiments with each IK solution. Secondly, after the manipulation task of the mechanism via the proposed inverse kinematicalgorithms is performed, some compensation may be needed due to the sudden and unpredicted deviation of the targetposition under field conditions.For the purpose of finding optimal joint values under certain constraints, a trajectory optimization problem in image-based visual servoing method via the camera-in-handconfiguration is initiated when the end-effector is in the close proximity of the target leaf. In this part of the study, a bio-inspired trajectory optimization problem in image-basedvisual servoing method is constructed based on the mathematical model derived from the prey-predatorrelationships in nature. In this biological phenomenon, the predator constructs its path in a certain subspace whilecatching the prey. When this motion strategy is applied to trajectory optimization problems, it causes a significantreduce in the computation cost since it finds the optimum solution in a certain manifold. The performance of the introducedbio-inspired trajectory optimization in visual servoing is validated with the hardware experiments both in laboratory settings and in fieldconditions.
Show less - Date Issued
- 2018
- Identifier
- CFE0007170, ucf:52247
- Format
- Document (PDF)
- PURL
- http://purl.flvc.org/ucf/fd/CFE0007170
