The problem of airtosurface trajectory optimization for a lowaltitude skidtoturn vehicle is considered. The objective is for the vehicle to move level at a low altitude for as long as possible and perform a rapid bunt (negative sensedacceleration load) maneuver near the final time in order to attain terminal target conditions. The vehicle is modeled as a point mass in motion over a flat Earth, and the vehicle is controlled using thrust magnitude, angle of attack, and sideslip angle. The trajectory
optimization problem is posed as a twophase optimal control problem using a weighted objective function. The work described in this paper is the first part of a twopart sequence on trajectory optimization and guidance of a skidtoturn vehicle. In both cases, the objective is to minimize the time taken by the vehicle to complete a bunt maneuver subject to the following constraints: dynamic, boundary, state, path, and interiorpoint event constraints. In the first part of this twopart study, the performance of thevehicle is assessed. In particular, the key features of the optimal reference trajectories and controls are provided. The results of this study identify that as greater weight is placed on minimizing the height of the bunt maneuver or as the maximum altitude constraint is raised, the time of the bunt maneuver decreases and the time of the problem solution increases. Also, the results of this study identify that as the allowable crossrange of the vehicle is reduced, the time and height of the bunt maneuver increases and the time of the problem solution decrease
more »
« less
Performance Optimization and Guidance of a LowAltitude SkidtoTurn Vehicle. Part II: Optimal Guidance
The problem of guidance and control of an airtosurface lowaltitude skidtoturn vehicle is considered. The objective is to steer the vehicle to a ground target from an initial state such that the vehicle remains at a constant low altitude for as long as possible and performs a bunt maneuver (negative sensedacceleration load) rapidly at the end of the trajectory in order to attain terminal target conditions. The vehicle is modeled as a point mass in motion over a flat Earth, and the vehicle is controlled using thrust magnitude, angle of attack, and sideslip angle. The guidance and control problem is posed as a decreasinghorizon optimal control problem and is resolved numerically at constant guidance cycles. The work described in this paper is the second part of a twopart sequence on trajectory optimization and guidance of a skidtoturn vehicle. In both cases, the objective is to minimize the time taken by the vehicle to complete a bunt maneuver subject to the following constraints: dynamic, boundary, state, path, and interiorpoint event constraints. In the second part of this twopart study, a numerical guidance law is employed to resolve the optimal control problem on a shrinking horizon. An assessment is made as to the time required to resolve the optimal control problem both in the absence and presence of a time delay, where the time delay is the amount of time required to solve the numerical approximation of the optimal control problem at the start of each guidance cycle. The results of this study identify that in the absence and presence of a time delay, the size of the mesh on subsequent guidance cycles decreases, the time required to solve the reducedhorizon optimal control problem is small compared to the guidance cycle duration time, and the terminal conditions are met with sufficient accuracy. Thus, the results show that the guidance law presented in this paper has the potential to solve optimal control problems in realtime.
more »
« less
 NSFPAR ID:
 10094024
 Date Published:
 Journal Name:
 2019 AIAA Guidance, Navigation, and Control Conference, San Diego
 Format(s):
 Medium: X
 Sponsoring Org:
 National Science Foundation
More Like this


A Monte Carlo analysis of a contingency optimal guidance strategy is conducted. The guidance strategy is applied to a Mars Entry problem in which it is assumed that the surface level atmospheric density is a random variable. First, a nominal guidance strategy is employed such that the optimal control problem is resolved at constant guidance cycles. When the trajectory lies within a particular distance from a path constraint boundary, the nominal guidance strategy is replaced with a contingency guidance strategy, where the contingency guidance strategy attempts to prevent a violation in the the relevant path constraint. The contingency guidance strategy utilizes the reference optimal control problem formulation, but modifies the objective functional to maximize the margin between the path constraint limit and path constraint function value. The ability of the contingency guidance strat egy to prevent violations in the path constraints is assessed via a Monte Carlo simulation.more » « less

A Monte Carlo analysis of a contingency optimal guidance strategy is conducted. The guidance strategy is applied to a Mars Entry problem in which it is assumed that the surface level atmospheric density is a random variable. First, a nominal guidance strategy is employed such that the optimal control problem is resolved at constant guidance cycles. When the trajectory lies within a particular distance from a path constraint boundary, the nominal guidance strategy is replaced with a contingency guidance strategy, where the contingency guidance strategy attempts to prevent a violation in the the relevant path constraint. The contingency guidance strategy utilizes the reference optimal control problem formulation, but modifies the objective functional to maximize the margin between the path constraint limit and path constraint function value. The ability of the contingency guidance strategy to prevent violations in the path constraints is assessed via a Monte Carlo simulation.more » « less

We consider the problem of finitehorizon optimal control of a discrete linear timevarying system subject to a stochastic disturbance and fully observable state. The initial state of the system is drawn from a known Gaussian distribution, and the final state distribution is required to reach a given target Gaussian distribution, while minimizing the expected value of the control effort. We derive the linear optimal control policy by first presenting an efficient solution for the diffusionless case, and we then solve the case with diffusion by reformulating the system as a superposition of diffusionless systems. We show that the resulting solution coincides with a LQG problem with particular terminal cost weight matrix.more » « less

This paper presents a novel learningbased adaptive optimal controller design for linear timedelay systems described by delay differential equations (DDEs). A key strategy is to exploit the value iteration (VI) approach to solve the linear quadratic optimal control problem for timedelay systems. However, previous learningbased control methods are all exclusively devoted to discretetime timedelay systems. In this article, we aim to fill in the gap by developing a learningbased VI approach to solve the infinitedimensional algebraic Riccati equation (ARE) for continuoustime timedelay systems. One nice feature of the proposed VI approach is that an initial admissible controller is not required to start the algorithm. The efficacy of the proposed methodology is demonstrated by the example of autonomous driving.more » « less