Yunjun Xu

Computing short-time aircraft maneuvers using direct methods

This paper analyzes the applicability of direct methods to design optimal short-term spatial maneuvers for an unmanned vehicle in a faster than real-time scale. It starts by introducing different basic control schemes, which employ online trajectory generation. Next, it presents and analyzes the results obtained through two recently developed direct transcription (collocation) methods: the Gauss pseudospec-tral method and the Legendre-Gauss-Lobatto pseudosp ectral method. The achieved results are further compared with those found through the Pontryagin’s Maximum (Minimum) Principle, and the paper continues by providing another set of direct method simulations incorporating more realistic boundary conditions. Finally, the results obtained using the third direct method, based on inverse dynamics in the virtual domain, are presented and discussed.

Vision Based Flexible Beam Tip Point Control

Because of the light weight and less wear and tear on components, the flexible beam has been and will continue to be an appealing option for civil and military applications. However, flexibility brings with it unwanted oscillations and severe chattering which may even lead to an unstable system. To tackle these challenges, a two-time scale controller is presented to track a desired tip point signal and at the same time mitigate the tip point vibration. To obtain more precise information of the tip point location and facilitate the easy extension to multiple-flexible-link problems, a camera is used to provide vision feedback in which the delayed vision signal is compensated by the state estimator and predictor. The controller is experimentally verified, and shown to exceed the performance of other tested controllers.

Sequential virtual motion camouflage method for nonlinear constrained optimal trajectory control

Nonlinear constrained optimal trajectory planning is a challenging and fundamental area of research. This paper proposes bio-inspired fast-time approaches for this type of problems based on the inspiration drawn from the natural phenomenon known as the motion camouflage. Two algorithms are proposed: the virtual motion camouflage (VMC) subspace method and the sequential VMC method. As a hybrid approach, the sequential VMC method works through a two-step structure in each iteration. First, the VMC subspace method will solve for an optimal solution over a selected subspace. Second, an algorithm consisting of a linear programming and a line search will vary the subspace so that the next VMC subspace result will be guaranteed not to be worse than that of the current step. The dimension and time complexities of the algorithms will be analyzed, and the optimality of the solution via the sequential VMC approach will be studied. Through the VMC approaches, the state and control variables in the kinematics or dynamics models of vehicles in the selected subspace can be represented by a single degree-of-freedom vector, called the path control parameter vector. The reduction in dimension and no involvement of equality constraints will in practice make the convergence faster and easier, and a much smaller computational cost is expected. Two simulation examples, the Breakwell problem and a minimum time robot obstacle avoidance problem with different numbers of obstacles, are used to demonstrate the capabilities of the algorithms.

Brief paper: Virtual motion camouflage based phantom track generation through cooperative electronic combat air vehicles

As an example of complex cooperative missions, coherent phantom track generation through controlling multiple electronic combat air vehicles is currently an area of great interest to the defense agency for the purpose of deceiving a radar network. However, it has been a challenge to design optimal coherent trajectories for this type of problems due to the high dimensionality and kinematic, dynamics, and geometric constraints. This paper describes how an interesting bio-inspired motion strategy can be used to design real-time trajectories for a (1) feasible constant speed coherent mission, (2) maximum-duration constant speed coherent mission, and (3) optimal trajectory mission for general cases.

Pre and Post Optimality Checking of the Virtual Motion Camouflage based Nonlinear Constrained Subspace Optimal Control

Nonlinear constrained trajectory optimization remains an active field of research. Current popular methods end up with either solving a classical multi-dimensional two-point boundary value problem or a high dimensional nonlinear programming problem. Inspired by the motion camouflage phenomenon of insects like dragonflies, recently the authors proposed a subspace optimization method, the virtual motion camouflage method, in order to reduce the problem dimension. Thus the computational cost experienced in the widely used direct collocation and nonlinear programming methods can be reduced. The solution found through this approach is in the feasible region however the optimality of the solution is not guaranteed automatically in the original full search space. In this paper, two optimality checking ways extended from the Karush-Kuhn-Tucker necessary condition are proposed. The pre-optimality checking is proposed to judge the selection of the virtual prey motion and the reference point, while a post-optimality checking is suggested to test the solution obtained from the virtual motion camouflage method in the original search space. Two numerical examples are used to illustrate the capabilities of the new subspace optimal control algorithm.

Nonlinear Stochastic Control Part II: Ascent Phase Control of Reusable Launch Vehicles

In designing a robust ascent phase control for reusable launch vehicles, uncertainties such as variations in aerodynamics, jet effects, hinge moments, mass property, and navigation processing, etc. have to be considered and normally time and labor intensive Monte Carlo simulations are used in order to achieve a desired tracking performance distribution. In this paper, a systematic stochastic control design method based upon a direct quadrature method of moments proposed in Part I [26] will be applied to the ascending phase attitude control problem. In conjunction with a nonlinear robust control and an offline optimization through nonlinear programming, any order of stationary statistical moments can be directly controlled. Two simulation scenarios of the X-33 ascent phase control have been used to demonstrate the capabilities of the proposed method and the results are validated by Monte Carlo runs.

Electrical actuation and shape recovery control of shape-memory polymer nanocomposites

Shape-memory polymers (SMPs) are one of the most popular smart materials due to their light weight and high elastic deformation capability. The synergistic effect of carbon nanofiber (CNF) and carbon nanofiber paper (CNFP) on the electro-actuation of SMP nanocomposites was studied. The electrical conductivity of SMPs was significantly improved by incorporating CNF and CNFP into them. The dynamic mechanical analysis result reveals good thermal stability of SMP nanocomposites even after they were mixed with CNFs. A vision-based control system is designed to precisely control the shape recovery of SMP composites. Any quasi-state shape between the permanent shape and a temporary shape can be achieved by adjusting the electrical energy input. Experimental results demonstrated that (1) compared with the baseline material, the full recovery time of the SMP nanocomposites was decreased by 1000% to less than 80 s; (2) a good repeatability was shown in the developed vision system in 10 experimental trials and the accuracy of the controlled deflection angle of SMPs was within a 5% error bound.

Vision based flexible beam tip point control

Because of the light weight and less wear and tear on components, the flexible beam/arm has been and will continue to be an appealing option for civil and military applications, such as space-based flexible manipulators. However, flexibility brings with it unwanted oscillations and severe chattering which may even lead to an unstable system. To tackle these challenges, a two-time scale controller is presented to track a desired tip point signal and at the same time mitigate the tip point vibration using direct vision feedback. In particular, an linear quadratic regulator (LQR) controller in the fast mode stabilizes the oscillations of the beam, and a boundary layer augmented sliding mode controller is proposed to track the desired position. To obtain more precise information of the tip point location and facilitate the easy extension to multiple-flexible-link problems, a camera is used to provide vision feedback in which the delayed vision signal is compensated by the state estimator and predictor. The vision data, which provides a direct measurement of the tip point, proves to be a better substitute for the more traditional strain gauge, which can only provide indirect measurements based on the mathematically derived mode shapes of the beam. The controller is experimentally verified, and shown to exceed the performance of other tested controllers.

Recovery torque modeling of carbon fiber reinforced shape memory polymer nanocomposites

Carbon fiber and carbon nanofiber paper (CF&CNFP) can be incorporated into shape memory polymers (SMPs) to increase electrical conductivity and allow high speed electrical actuation with a low power. This paper studies the interactions among the recovery torques of CF&CNFP and SMP and the gravity torque during the shape recovery process. The proposed recovery torque model in a SMP CF&CNFP based structure is validated by experimental data obtained using a recently developed low cost, non-contact measurement testbed.

Evacuation Modeling From the Control Perspective and Corresponding Sequential-Based Optimal Evacuation Guidance

Most of the emergency alert systems to date can only produce alarm sounds and remind people not to use elevators. Without effective guidance information, it is highly possible that people will rush into certain staircases, which may, however, result in stampedes, crushing, and trampling. To increase peoples survivability and mitigate the losses when disasters happen, it is desirable that the future emergency alert system is capable of providing guidance information to people in real time. In this brief, a new scalable evacuation model is proposed from the control perspective. The partial decoupling characteristic of the model leads to a sequential approach in the optimal guidance design so that the computational cost can be significantly reduced as compared with the centralized approach. It is shown that the sequential approach can achieve the same minimum time performance, although the optimal guidance may only be one of the many solutions. Characteristics of the proposed model and the capabilities of the associated optimization methods are illustrated through simulations.