Expensive multiobjective optimization for robotics

Tesch, Matthew; Schneider, Jeff; Choset, Howie

doi:10.1109/icra.2013.6630691

Cited by 38 publications

(28 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, we propose an algorithmic solution for the generation of Pareto optimal MPC designs, which is based on a hypervolume criterion described in [2]. In contrast to generic implementation [3], the proposed algorithm speeds up the solution of expensive MOO problems by (i) parallelizing the evaluations and (ii) decoupling the objective function and constraint models, so that objective functions are not evaluated for infeasible designs. This is the first application of a systematic MOO method for codesign of an MPC algorithm and underlying computational platform.…”

Section: Performance Vs Time Energy and Spacementioning

confidence: 99%

Multi-objective Co-design for model predictive control with an FPGA

Khusainov

Kerrigan

Constantinides

2016

2016 European Control Conference (ECC)

View full text Add to dashboard Cite

Abstract-In order to achieve the best possible performance of a model predictive controller (MPC) for a given set of resources, the software algorithm and computational platform have to be designed simultaneously. Moreover, in practical applications the controller design problem has a multi-objective nature: performance is traded off against computational hardware resource usage, namely time, energy and space. This paper proposes formulating an MPC design problem as a multiobjective optimization (MOO) problem in order to explore the design trade-offs in a systematic way.Since the design objectives in the resulting MOO problem are expensive to evaluate, i.e. evaluation requires time consuming simulations, most of the classical and evolutionary MOO algorithms cannot be employed for this class of design problems. For this reason a practical MOO algorithm that can deal with expensive-to-evaluate functions is presented. The algorithm is based on Kriging and the hypervolume criterion that was recently proposed in the expensive optimization literature. A numerical example for a fast gradient-based controller design shows that the proposed approach can efficiently explore optimal performance-resource trade-offs. I. PERFORMANCE VS. TIME, ENERGY AND SPACEModel predictive control (MPC) has already proven to be an efficient and reliable solution for a wide-range of applications, most of which are characterized by relatively slow dynamics. The necessity of solving an optimization problem at each sampling instance has traditionally been the main bottleneck that prevented more significant expansion of MPC, especially in relation to fast dynamic plants. Recent developments in communication technologies and hardware processing systems have sped up online optimization solvers and hence increased the potential application scope of MPC.Most MPC and underlying optimization algorithms are designed at a high level of abstraction without regard to the intended hardware platform. This decoupled approach usually leads to inefficient utilization of available resources [1]. In contrast, the co-design approach implies simultaneous design of both software and hardware components in order to achieve the best possible performance for a given set of available resources. However, improvement of the closed-loop performance cannot be considered as the only design objective. In practice, there is a trade-off between *The work leading to these results has received funding from the People Programme (Marie Curie Actions) of the European Unions Seventh Framework Programme (FP7/2007(FP7/ -2013 e.kerrigan@imperial.ac.uk performance and computational hardware resource usage. Resources that are required to perform computations include time, energy and space, which are also the functions of the MPC algorithm and hardware platform. Instead of looking for one optimal design (which often does not exist due to conflicts between objectives) engineers might make a decision based on the whole series of Pareto optimal designs, i.e. designs that cannot be improved ...

show abstract

Section: Performance Vs Time Energy and Spacementioning

confidence: 99%

Multi-objective Co-design for model predictive control with an FPGA

Khusainov

Kerrigan

Constantinides

2016

2016 European Control Conference (ECC)

View full text Add to dashboard Cite

show abstract

“…Nevertheless, related work is available in the control of a snake robot, to stabilize the motion of the robot's end modules as a gait progresses [8], [9]. Particularly, in [9], optimization techniques are used to the problem of moving a snake robot (an expensive system) optimizing simultaneously the gait performance as well as the ability to maintain motion of a specific module in certain way i.e.…”

Section: Introductionmentioning

confidence: 99%

Where to place cameras on a snake robot: Focus on camera trajectory and motion blur

Mutlu

Melo

Vespignani

et al. 2015

2015 IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR)

View full text Add to dashboard Cite

Abstract-Visual information is heavily used in robotics, in particular for SLAM applications. Visual SLAM algorithms depend on robust feature extraction and reliable state estimation. Quality of the visual information highly depends on how that information is captured. The nature of snake robots' locomotion presents considerable challenges on the quality of images captured by an onboard mobile camera. Although placing the camera on the "head" of the snake robot has advantages when the robot is stationary since the body can be used as a manipulator observing for the environment, how to place the camera in order to capture more useful images for navigation during locomotion is not clear. In this paper, we present a comparative study to discuss implications of the camera location on field coverage and types of image quality for three snake gaits: Rolling, sidewinding and linear progression. Camera pose during locomotion is examined in detail and quality of images are quantified using a motion blur metric which relates camera egomotion to blur. Linear progression is found to be very promising in terms of supplying sharper images. But, there are also other merits that can be exploited in different locomotion types and camera locations.

show abstract

“…However, many real-world applications naturally present multiple criteria to be optimized. For example, in complex robotic systems, we must consider performance criteria such as motion accuracy, speed, robustness to noise or energy-efficiency [9]. Typically, it is impossible to optimize all these desiderata at the same time as they may be conflicting.…”

Section: Introductionmentioning

confidence: 99%

Intelligent Autonomous Systems 12

Groen¹,

Amato²,

Bonarini³

et al. 2013

Advances in Intelligent Systems and Computing

View full text Add to dashboard Cite

Many real-world applications require the optimization of multiple conflicting criteria. For example, in robot locomotion, we are interested is maximizing speed while minimizing energy consumption. Multi-objective Bayesian optimization (MOBO) methods, such as ParEGO [6], ExI [5] and SMS-EGO [8] make use of models to define the next experiment, i.e., select the next parameters for which the objective function is to be evaluated. However, suggesting the next experiment is typically the only use of models in MOBO. In this paper, we propose to further exploit these models to improve the estimation of the final Pareto front and ultimately provide useful tool to the user for further analysis. We demonstrate that a small philosophical difference leads to substantial advantages in the practicality of most MOBO methods "for free".Optimization often requires the definition of a single objective function to be optimized. However, many real-world applications naturally present multiple criteria to be optimized. For example, in complex robotic systems, we must consider performance criteria such as motion accuracy, speed, robustness to noise or energy-efficiency [9]. Typically, it is impossible to optimize all these desiderata at the same time as they may be conflicting. However, it is still desirable to find a trade-off that satisfies as much as possible the different criteria and the necessities of the final user.For practical purposes, the existence of multiple criteria is often side-stepped by designing a single objective that incorporates all criteria, e.g., by defining a weighted sum of the criteria [3]. Alternatively, the optimization of these different objectives can be formalized as a Multi-Objective Optimization (MOO) [2]. In MOO, the goal is to return a Pareto front (PF), which represents the best trade-off possible between the different criteria [7]. From this Pareto front, it is the responsibility of the user to ultimately select the most convenient/promising set of parameters to apply. Intuitively, the goodness of the returned PF can be measured by the accuracy (how close is the proposed PF to the real unknown optimal PF), by the size (having a large set of solutions in the PF is desirable) and by the diversity (having solutions that encompass a wide range of trade-offs).Many model-based MOO methods exist, which extend Bayesian optimization to the multi-objective case, such as ParEGO [6], ExI [5] and SMS-EGO [8]. We here define all these methods as Multi Objective Bayesian Optimization (MOBO) methods. Currently, the main advantage of MOBO methods, compared to model-free MOO methods [4,12], is to reduce the number of experiments/ evaluations of the objective function. However, the models of the objective functions are used exclusively to select the next parameters to evaluate.In this paper, we present a new perspective on MOBO and on the use of models in MOO. We demonstrate that it is possible to exploit the learned models, e.g., for better approximations of the Pareto front or for computing additional s...

show abstract

Expensive multiobjective optimization for robotics

Cited by 38 publications

References 20 publications

Multi-objective Co-design for model predictive control with an FPGA

Multi-objective Co-design for model predictive control with an FPGA

Where to place cameras on a snake robot: Focus on camera trajectory and motion blur

Intelligent Autonomous Systems 12

Contact Info

Product

Resources

About