Analytical results for the multi-objective design of model-predictive control

Bachtiar, Vincent; Manzie, Chris; Moase, William H.; Kerrigan, Eric C.

doi:10.1016/j.conengprac.2016.07.009

Cited by 4 publications

(7 citation statements)

References 38 publications

(58 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The study then demonstrates how the proposed nonlinear MPC-based controller can be calibrated (tuned) offline with a multi-objective outlook, considering both closed-loop performance and required computational capacity to implement the controller. This extends recent work of [16] on linear MPC. Two tuning objectives are considered, the first being the miss-distance to measure control performance whilst the second is the required computational load to implement the controller as a measure of the design cost of the controller.…”

supporting

confidence: 86%

“…The OCP optimizes the cost function J that represents the performance of the plant. The optimization is subject to the predictive model (18c) representing the dynamics of the plant and its constraints (18d) as given in (15) and (16).…”

Section: Model Predictive Missile Autopilot and Guidancementioning

confidence: 99%

“…The non-dimensional Capacity Number [16] can be used to indicate the required resources/computational power of the implementation hardware in multiples relative to the simulation hardware. This metric is given by…”

Section: Required Computational Capacitymentioning

confidence: 99%

“…More recently, combined optimization problems in control have been analyzed with more rigor and depth to reveal certain characteristics and properties of the problem. In these analyses, deterministic optimization algorithms have been used [16,17].…”

mentioning

confidence: 99%

“…The tuning parameters studied are those pertaining to the structure of the MPC, namely the sampling time and number of prediction steps, thus affecting both objectives competitively. The solution method for the resulting multi-objective optimization problem is based on a Lipschitzian optimizer, dubbed Dividing Triangles (DITRI) and developed in [16]. The result of the design approach is the optimal trade-off curve consisting of a set of optimal designs that the practitioner can base the controller design on.…”

mentioning

confidence: 99%

See 4 more Smart Citations

Nonlinear Model-Predictive Integrated Missile Control and Its Multiobjective Tuning

Bachtiar

Manzie

Kerrigan

2017

Journal of Guidance, Control, and Dynamics

Self Cite

View full text Add to dashboard Cite

I. Introduction Advanced missile control designs have been founded on recent developments aimed at a more streamlined engineering design approach to improve performance and cost of design. Central to these developments is the attempt for a full integration of subsystems to exploit any potential synergy. In missile control, much effort has been concentrated on integrating the autopilot and guidance subsystems [1, 2]. Such a unification of the control algorithm has direct benefits for the performance of the missile, primarily by removing the previously existing lag between the commanded and tracked acceleration from guidance and autopilot, respectively [3]. Several strategies have been proposed following this integrated control outlook, including sliding mode control [4, 5, 6], state feedback regulators [7, 8], as well as model-based approaches, for instance, a recently proposed linear quadratic regulation with model bias [9]. Model predictive control (MPC) is recognized for its ability in handling nonlinearities and constraints, therefore it is suited for high-performing agile missiles operated near constraints. This is demonstrated, for example, by the fact that the predominant proportional navigation (PN) in missile guidance is based on linear quadratic regulation [10], which is a special case of model-predictive control where constraints are absent and linear prediction model is used. Furthermore, the performance and feasibility of model-predictive missile autopilot has been studied in [11, 12]. Although promising a superior performance, MPC is typically associated with high levels of computational cost. Therefore, prior to the real-time realization of MPC, it is important to know the required computational capacity to implement MPC, especially in applications involving fast sampling rates such as missile control. A comprehensive offline tuning of MPC that balances both performance and cost is therefore needed before the implementation. In light of this, as a natural tuning method for an integrated autopilot and guidance algorithm, a multi-objective design approach that considers both closed-loop performance and required computational capacity allows for the comprehensive consideration of both algorithm and implementation designs of the system. A coupled approach in control tuning is known to be valuable due to its potential in streamlining the design process and produce system-optimal designs [13]. This allows for a more comprehensive analysis, such as the optimization of multiple underlying tuning/design objectives along with constraints, as introduced and discussed in [14]. This notion

show abstract

supporting

confidence: 86%

Section: Model Predictive Missile Autopilot and Guidancementioning

confidence: 99%

Section: Required Computational Capacitymentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations

Nonlinear Model-Predictive Integrated Missile Control and Its Multiobjective Tuning

Bachtiar

Manzie

Kerrigan

2017

Journal of Guidance, Control, and Dynamics

Self Cite

View full text Add to dashboard Cite

show abstract

Active suspension control strategy for vehicles based on road surface recognition

Yang,

Li,

et al. 2024

Nonlinear Dyn

View full text Add to dashboard Cite

An adaptive model predictive control (AMPC) algorithm based on pavement identi cation was proposed to determine the in uence of different pavement inputs on a vehicle suspension system. First, a vehicle dynamics model was established, and a discrete leveling index-based pavement comfort assessment method was proposed based on the international leveling index to quantify the comfort level by calculating the maximum instantaneous vibration index based on the vertical acceleration of the driver's seat. Second, an augmented Kalman lter algorithm with a forgetting factor is proposed to track the pavement time-varying parameters and estimate pavement leveling. Finally, the control of the quartervehicle active suspension system is transformed into solving the decay of hard constraints, designing the AMPC strategy, parameterizing the cost function of AMPC with the global cost of the performance index as the evaluation function, and using Bayesian optimization to predict the time domain and weight of the cost function to achieve the global performance optimum. While satisfying the dynamic constraints, passenger comfort is improved by attenuating roadway pre-sighting disturbances. Experimental results indicate that the proposed active suspension AMPC algorithm improves suspension smoothness and driver comfort compared with the MPC algorithm and passive suspension. The effectiveness of the proposed control algorithm for different road vibrations was veri ed.

show abstract

Vehicle Path-Tracking Linear-Time-Varying Model Predictive Control Controller Parameter Selection Considering Central Process Unit Computational Load

Wang

Bai

Wang

et al. 2019

Journal of Dynamic Systems, Measurement, and Control

View full text Add to dashboard Cite

Model predictive control (MPC) has drawn a considerable amount of attention in automotive applications during the last decade, partially due to its systematic capacity of treating system constraints. Even though having received broad acknowledgements, there still exist two intrinsic shortcomings on this optimization-based control strategy, namely the extensive online calculation burden and the complex tuning process, which hinder MPC from being applied to a wider extent. To tackle these two drawbacks, different methods were proposed. Nevertheless, the majority of these approaches treat these two issues independently. However, parameter tuning in fact has double-sided effects on both the controller performance and the real-time computational burden. Due to the lack of theoretical tools for globally analyzing the complex conflicts among MPC parameter tuning, controller performance optimization, and computational burden easement, a look-up table-based online parameter selection method is proposed in this paper to help a vehicle track its reference path under both the stability and computational capacity constraints. matlab-carsim conjoint simulations show the effectiveness of the proposed strategy.

show abstract

Analytical results for the multi-objective design of model-predictive control

Cited by 4 publications

References 38 publications

Nonlinear Model-Predictive Integrated Missile Control and Its Multiobjective Tuning

Nonlinear Model-Predictive Integrated Missile Control and Its Multiobjective Tuning

Active suspension control strategy for vehicles based on road surface recognition

Vehicle Path-Tracking Linear-Time-Varying Model Predictive Control Controller Parameter Selection Considering Central Process Unit Computational Load

Contact Info

Product

Resources

About