0
Research Papers

# Optimal Control of a Formula One Car on a Three-Dimensional Track—Part 1: Track Modeling and Identification

[+] Author and Article Information
Giacomo Perantoni

Department of Engineering Science,
University of Oxford,
Oxford OX1 3BN, UK
e-mail: giacomo.perantoni@gmail.com

David J. N. Limebeer

Department of Engineering Science,
University of Oxford,
Oxford OX1 3BN, UK
e-mail: david.limebeer@eng.ox.ac.uk

If a × b = c, then R(a × b) = Rc. If $a'=Ra,b'=Rb$ and $c'=Rc$, then $a'×b'=c'$, since the relative orientations and lengths of these vectors have not changed. Consequently (Ra) × (Rb) = Rc = R(a × b).

Throughout this paper we will assume that we are dealing with constant-speed curves and so derivatives with respect to time t and arc length s are equivalent; $(ds/dt)=1$. We will use a dot to denote derivatives with respect to either variable.

Interpolation can be used to enforce this condition if the number of measurement points for the two boundaries is different.

Contrary to accepted convention we use χ to represent the state vector and υ to denote the control vector. This notation is used to avoid confusion with the first translational component x of a position vector, and with the longitudinal velocity u of a vehicle as is used in Part 2 [30].

The forward Euler scheme for finite differences can be used for x0, whereas xf has to be defined using a backward finite-differentiation rule.

The World Geodetic System is a standard for use in cartography and geodesy. It comprises a standard coordinate frame for the Earth, a datum or reference ellipsoid for altitude data and a gravitational equipotential surface (the geoid) that defines nominal sea level. The latest revision is WGS 84 that was last revised in 2004.

Contributed by the Dynamic Systems Division of ASME for publication in the JOURNAL OF DYNAMIC SYSTEMS, MEASUREMENT, AND CONTROL. Manuscript received December 13, 2013; final manuscript received August 7, 2014; published online January 27, 2015. Assoc. Editor: Shankar Coimbatore Subramanian.

J. Dyn. Sys., Meas., Control 137(5), 051018 (May 01, 2015) (11 pages) Paper No: DS-13-1506; doi: 10.1115/1.4028253 History: Received December 13, 2013; Revised August 07, 2014; Online January 27, 2015

## Abstract

The identification of three-dimensional (3D) race track models from noisy measured GPS data is treated as a problem in the differential geometry of curves and surfaces. Curvilinear coordinates are adopted to facilitate the use of the track model in the solution of vehicular optimal control problems. Our proposal is to model race tracks using a generalized Frenet–Serret apparatus, so that the track is specified in terms of three displacement-dependent curvatures and two edge variables. The optimal smoothing of the curvature and edge variables is achieved using numerical optimal control techniques. Track closure is enforced through the boundary conditions associated with the optimal control problem. The Barcelona formula one track is used as an illustrative example.

<>
Your Session has timed out. Please sign back in to continue.

## References

Hendrikx, J. M., Meijlink, T., and Kriens, R. F. C., 1996, “Application of Optimal Control Theory to Inverse Simulation of Car Handling,” Vehicle System Dynamics, 26(6), pp. 449–461.
Casanova, D., 2000, “On Minimum Time Vehicle Manoeuvring: The Theoretical Optimal Lap,” Ph.D. thesis, Cranfield University School of Engineering, Bedfordshire, UK.
Kelly, D. P., 2008, “Lap Time Simulation with Transient Vehicle and Tyre Dynamics,” Ph.D. thesis, Cranfield University School of Engineering, Bedfordshire, UK.
Cossalter, V., Lio, M. D., Lot, R., and Fabbri, L., 1999, “A General Method for the Evaluation of Vehicle Manoeuvrability With Special Emphasis on Motorcycles,” Veh. Syst. Dyn., 31(2), pp. 113–135.
Perantoni, G., and Limebeer, D. J., 2014, “Optimal Control for a Formula One Car With Variable Parameters,” Veh. Syst. Dyn., 52(5), pp. 653–678.
Timings, J. P., and Cole, D. J., 2013, “Minimum Maneuver Time Calculation Using Convex Optimization,” ASME J. Dyn. Syst., Meas., Control, 135(3), p. 031015.
Koenderink, J. J., 1990, Solid Shape (Artificial Intelligence), MIT, Cambridge, MA.
White, J. H., and Bauer, W. R., 1986, “Calculation of the Twist and the Writhe for Representative Models of DNA,” J. Mol. Bio., 189(2), pp. 329–341.
Panyukov, S., and Rabin, Y., 2000, “Fluctuating Filaments: Statistical Mechanics of Helices,” Phys. Rev. E, 62(5 Pt B), pp. 7135–46.
Kessler, D. A., and Rabin, Y., 2003, “Effect of Curvature and Twist on the Conformations of a Fluctuating Ribbon,” J. Chem. Phys., 118(2), pp. 897–904.
Rappaport, S. M., and Rabin, Y., 2007, “Differential Geometry of Polymer Models: Worm-Like Chains, Ribbons and Fourier Knots,” J. Phys. A: Math. Theor., 40(17), pp. 4455–4466.
Behringer, R., van Holt, V., and Dickmanns, D., 1992, “Road and Relative Ego-State Recognition,” Proceedings of the Intelligent Vehicles '92 Symposium, Detroit, MI, June 29–July 7, pp. 385–390.
Dickmanns, E., and Mysliwetz, B., 1992, “Recursive 3-D Road and Relative Ego-State Recognition,” IEEE Trans. Patt. Anal. Mach. Intell., 14(2), pp. 199–213.
Behringer, R., 1995, “Detection of Discontinuities of Road Curvature Change by GLR,” Proceedings of the Intelligent Vehicles '95 Symposium, Detroit, MI, Sept. 25–26, pp. 78–83.
Khosla, D., 2002, “Accurate Estimation of Forward Path Geometry Using Two-Clothoid Road Model,” IEEE Intelligent Vehicle Symposium, Versailles, France, June 17–21, Vol. 1, pp. 154–159.
Loose, H., and Franke, U., 2010. “B-Spline-Based Road Model for 3D Lane Recognition,” 13th International IEEE Conference on Intelligent Transportation Systems (ITSC), Funchal, Madeira Island, Portugal, Sept. 19–22, pp. 91–98.
Cong, S., Shen, S., and Hong, L., 2009, “Road Curvature Estimation System,” U.S. Patent No. 7,626,533.
Shen, T., and Ibrahim, F., 2012, “Interacting Multiple Model Road Curvature Estimation,” 15th International IEEE Conference on Intelligent Transportation Systems (ITSC), Anchorage, AK, Sept. 16–19, pp. 710–715.
Eidehall, A., and Gustafsson, F., 2006 “Obtaining Reference Road Geometry Parameters From Recorded Sensor Data,” IEEE Intelligent Vehicles Symposium, Tokyo, Japan, June 13–15, pp. 256–260.
Mena, J., 2003, “State of the Art on Automatic Road Extraction for GIS Update: A Novel Classification,” Pattern Recognit. Lett., 24(16), pp. 3037–3058.
Lin, X., Zhang, J., Liu, Z., Shen, J., and Duan, M., 2011, “Semi-Automatic Extraction of Road Networks by Least Squares Interlaced Template Matching in Urban Areas,” Int. J. Remote Sens., 32(17), pp. 4943–4959.
Willemsen, P., Kearney, J., and Wang, H., 2003, “Ribbon Networks for Modeling Navigable Paths of Autonomous Agents in Virtual Urban Environments,” IEEE Virtual Reality, Los Angeles, CA, Mar. 22–26, pp. 79–86.
Kreyszig, E., 1991, Differential Geometry, Dover Publications, New York.
Struik, D. J., 1988, Lectures on Classical Differential Geometry, 2nd ed., Dover, New York.
Gear, C. W., 1971, Numerical Initial Value Problems in Ordinary Differential Equations (Prentice-Hall Series in Automatic Computation), Prentice-Hall, Englewood Cliffs, NJ.
Brenan, K. E., Campbell, S. L., and Petzold, L. R., 1996, Numerical Solution of Initial-Value Problems in Differential-Algebraic Equations (SIAM's Classics in Applied Mathematics), SIAM, Philadelphia, PA.
Griffiths, D., and Higham, D., 2010, Numerical Methods for Ordinary Differential Equations: Initial Value Problems, Springer, New York.
Betts, J. T., 2001, Practical Methods for Optimal Control and Estimation Using Nonlinear Programming, 2nd ed., SIAM, Philadelphia, PA.
Darby, C. L., Hager, W. W., and Rao, A. V., 2011, “An Hp-Adaptive Pseudospectral Method for Solving Optimal Control Problems,” Optim. Control Appl. Methods, 32(4), pp. 476–502.
Limebeer, D. J. N., and Perantoni, G., 2013, “Optimal Control of a Formula One Car on a Three-Dimensional Track Part 2: Optimal Control,” ASME J. Dyn. Syst., Meas., Control,(submitted).
Patterson, M. A., and Rao, A. V., 2013, “GPOPS—II: A Matlab Software for Solving Multiple-Phase Optimal Control Problems Using Hp-Adaptive Gaussian Quadrature Collocation Methods and Sparse Nonlinear Programming,” ACM Trans. Math. Soft., 39(3), 41 pages.
Patterson, M. A., and Rao, A. V., 2011, “Exploiting Sparsity in Direct Collocation Pseudospectral Methods for Solving Optimal Control Problems,” J. Spacecr. Rockets, 49(2), pp. 364–377.
Lawrence, J. D., 1972, A Catalog of Special Plane Curves, Dover Publications, New York.

## Figures

Fig. 1

A ribbon R generated by a spine curve C. The spine is described in terms of its torsion and curvature, while the camber of the ribbon comes from the twist angle ν(s).

Fig. 2

Three-dimensional ribbon for the extended Lemniscate of Bernoulli. The solid and dashed lines are the optimized boundaries and centerline, respectively. The dots represent the noisy simulated boundary-point measurements on which the solution is based. The ribbon gray scale represents the z-coordinate of the optimized centerline and is calibrated according to the color bar given.

Fig. 4

Sums of the left-hand and right-hand boundary errors for different weight selections. The color convention is the same as that used in Table 1.

Fig. 3

Parameters for the extended lemniscate ribbon. The theoretically correct values are the red dashed curves, while the optimal estimates are the solid curves. The color convention is the same as that given in Table 1. The top plot shows the relative torsion Ωx, the second the normal curvature Ωy, and the third the geodesic curvature Ωz. The distances between the ribbon boundaries and its spine are shown in the bottom diagram. The theoretical boundaries come from Eq. (40).

Fig. 5

Ribbon-based model of the Circuit de Catalunya. The solid lines are the optimized track boundaries. The track spine's z-coordinate is shown using a gray scale.

Fig. 8

Distances (in meters) to the right-hand (solid black) and left-hand (dashed red) boundaries of the track as measured from the track's spine

Fig. 9

Error estimates for the track boundaries (as calculated in the ribbon tangent plane in meters). The solid black curve corresponds to the right-hand side boundary, while the dashed red curve corresponds to the left-hand boundary.

Fig. 6

Magnified views of corner 10 (left-hand side), and corners 14 and 15 (right-hand side). The measured boundary points are shown as black dots, with the dashed red curve the spine curve estimate.

Fig. 7

Estimated curvature of the Circuit de Catalunya. The plots show, respectively, the relative torsion Ωx, the normal curvature Ωy, and the geodesic curvature Ωz (in radians per meter). The optimal model-based estimates are shown as the solid black curves, while the finite-difference approximations are the red dashed curves.

## Errata

Some tools below are only available to our subscribers or users with an online account.

### Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related Proceedings Articles
Related eBook Content
Topic Collections