Kinematic Car Model

Deprecation warning

This website is now archived. Please check out the new website for Centre for Intelligent Systems which includes both A-Lab Control Systems Research lab and Re:creation XR lab.

However, the Dynamic System Model Database can still be used and may be updated in the future.

Continuously stirred tank reactor system

Model description: 

A schematic of the CSTR plant is shown in the attached image. The process dynamics are described by

$$\begin{align*} \dot{C}_{a} &=\frac{q}{V}(C_{a0}-c_{a})-a_{0}C_{a}e^{-\frac{E}{RT_{a}}} \\ \dot{T}_{a} &=\frac{q}{V}(T_{f}-T_{a})+a_{1}C_{a}e^{-\frac{E}{RT_{a}}}+a_{3}q_{c}\left(1-e^{\frac{a_{2}}{q_{c}}}\right)(T_{cf}-T_{a}), \end{align*}$$

where the variables $C_a$ and $T_a$ are the concentration and temperature of the tank, respectively; the coolant flow rate $q_c$ is the control input and the parameters of the plant are defined in the attached table. Within the tank reactor, two chemicals are mixed and react to produce a product compound $A$ at a concentration $C_a(t)$ with the temperature of the mixture being $T(t)$. The reaction is both irreversible and exothermic.

In the paper, authors assumed that plant parameters $q, C_{a0}, T_f$ and $V$ are at the nominal values given in the attached table. The activation energy $E/R = 1 \times 10^4K$ is assumed to be known. The state variables the input and the output are defined as $x=[x_1,x_2]^{\mathrm T}=[C_a,T_a]^{\mathrm T},u=q_c,y=C_a$. Using this notation, the CSTR plant can be re-expressed as

$$\begin{align*} \dot{x}_{1} &=1-x_{1}-a_{0}x_{1}e^{-\frac{10^4}{{\rm a}_2}} \\ \dot{x}_{2} &=T_{f}-x_{2}+a_{1}x_{1}e^{-\frac{10^4}{{\rm a}_2}}+a_{3}u\left(1-e^{-\frac{a_2}{u}}\right)(T_{cf}-x_{2}) \\ y &= x_{1}, \end{align*}$$

where the unknown constant parameters are $a_0, a_1, a_2$ and $a_3$.

Type: 

Form: 

Time domain: 

Linearity: 

Attachment: 

Publication details: 

TitleAdaptive nonlinear control of continuously stirred tank reactor systems
Publication TypeConference Paper
Year of Publication2001
AuthorsZhang, T., and Guay M.
Conference NameProceedings of the 2001 American Control Conference, 2001.
Date Published06/2001
PublisherIEEE
Conference LocationArlington, VA
ISBN Number0-7803-6495-3
Accession Number7106659
Keywordsadaptive control, asymptotic stability, chemical technology, closed loop systems, feedback, Lyapunov methods, neurocontrollers, nonlinear control systems, process control
AbstractAdaptive nonlinear control is investigated for a class of continuously stirred tank reactor (CSTR) system. The CSTR plant under study belongs to a class of general nonlinear systems, and contains an unknown parameter that enters the model nonlinearly. Using adaptive backstepping and neural network (NN) approximation techniques, an alternative adaptive NN controller is developed that achieves asymptotic output tracking control. Both stability and control performance analysis of the closed-loop adaptive system are based on Lyapunov's stability techniques
DOI10.1109/ACC.2001.945898

Discrete bilinear plant

Model description: 

The plant is

$$y(k)=1.2y(k-1)-0.8y(k-2)+0.2y(k-1)u(k-1)+u(k-1)+0.6u(k-2) + d(k),$$

where $d(k)$ is a disturbance.

Type: 

Form: 

Model order: 

2

Time domain: 

Linearity: 

Publication details: 

TitleAdaptive Bilinear Model Predictive Control
Publication TypeConference Paper
Year of Publication1986
AuthorsYeo, Y.K., and Williams D.C.
Conference NameAmerican Control Conference, 1986
Date Published06/1986
PublisherIEEE
Conference LocationSeattle, WA
Keywordsadaptive control, control system synthesis, Delay, Error correction, Least squares approximation, Mathematical model, parameter estimation, predictive control, Predictive models, Programmable control
AbstractAn adaptive controller for bilinear plants without delay and with stable inverses is defined based upon a bilinear model predictive control law and a classical recursive identification algorithm. For the case with no disturbance both the control error and the identification error converge to zero. For the case with a bounded disturbance, the control error is bounded and the identification converges. For the case with a constant disturbance, the control error often converges to zero and the identification converges.
URLhttp://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=4789155&queryText%3DADAPTIVE+BILINEAR+MODEL+PREDICTIVE+CONTROL

SISO NLTI plant

Model description: 

Consider the following unknown discrete nonlinear dynamic system:

$$\begin{align*} y(k+1)&=p[{\bf q}(k), u(k)]=0.2\cos[0.8(y(k)+y(k-1))] \\ & +0.4\sin[0.8(y(k)+y(k-1))+2u(k)+u(k-1)] \\ &+0.1[9+y(k)+y(k-1)]+\left[{2(u(k)+u(k-1))\over 1+\cos(y(k))}\right] \end{align*}$$

for $k=0,1,2,\ldots$ with $y(k)=0,u(k)=0$, for $k=0,-1,-2,\ldots$, $\Delta t := t(k+1)-t(k)=0.02sec$, for $k=0,1,2,\ldots$.

Type: 

Form: 

Model order: 

2

Time domain: 

Linearity: 

Publication details: 

TitleRobust nonlinear adaptive control using neural networks
Publication TypeConference Paper
Year of Publication2001
AuthorsAdetona, O., Sathananthan S., and Keel L.H.
Conference NameProceedings of the 2001 American Control Conference, 2001
Date Published06/2001
PublisherIEEE
Conference LocationArlington, VA
ISBN Number0-7803-6495-3
Accession Number7092721
Keywordsadaptive control, asymptotic stability, neurocontrollers, nonlinear control systems, radial basis function networks, robust control
AbstractThis paper provides a robust indirect adaptive control method for non-affine plants. Subject to some mild assumptions, the method can be applied to both minimum and non-minimum phase plants with operating regions of any finite size while avoiding a set of restrictions, at least one of which is imposed by all existing methods. The benefits are achieved under the following assumptions: 1) the operating region is limited to the basin of attraction of an asymptotically stable equilibrium point of the plant; 2) the desired output of the plant is sufficiently slowly varying; and 3) the output of the plant must be sufficiently sensitive to the input signal. It is shown that the adaptive control system will be stable in the presence of unknown bounded modeling errors
DOI10.1109/ACC.2001.946247

2-input 2-output nonlinear system

Model description: 

The suggested tracker scheme is tested with a 2-input 2-output nonlinear system given by:

$$\begin{align*} y_{1} (k) & = 0.21y_{1} (k-1)-0.12y_{2} (k-2) \\ & + 0.3y_{1} (k-1)u_{2} (k-1)-1.6u_{2} (k-1) \\ & + 1.2u_{1} (k-1), \\ y_{2} (k) & = 0.25y_{2} (k-1)-0.1y_{1} (k-2) \\ &- 0.2 y_{2} (k-1)u_{1} (k-1)-2.6u_{1} (k-1) \\ &-1.2u_{2} (k-1). \end{align*}$$

Type: 

Form: 

Time domain: 

Linearity: 

Publication details: 

TitleU-model Based Adaptive Tracking Scheme for Unknown MIMO Bilinear Systems
Publication TypeConference Paper
Year of Publication2006
AuthorsAzhar, A.S.S., Al-Sunni F.M., and Shafiq M.
Conference Name1ST IEEE Conference on Industrial Electronics and Applications, 2006
Date Published05/2006
PublisherIEEE
Conference LocationSingapore
ISBN Number0-7803-9513-1
Accession Number9097014
Keywordsbilinear systems, discrete time systems, linear systems, MIMO systems, neurocontrollers, radial basis function networks
AbstractBilinear systems are attractive candidates for many dynamical processes, since they allow a significantly larger class of behaviour than linear systems, yet retain a rich theory which is closely related to the familiar theory of linear systems. A new technique for the control of unknown MIMO bilinear systems is introduced. The scheme is based on the U-model with identification based on radial basis functions neural networks which is known for mapping any nonlinear function. U-model is a control oriented model used to represent a wide range of non-linear discrete time dynamic plants. The proposed tracking scheme is presented and verified using simulation examples
DOI10.1109/ICIEA.2006.257063

Kinematic Car Model

Model description: 

The attached image depicts the kinematic car in the horizontal plane. Let us suppose that the Ackermann steering assumptions hold true, hence all wheels turn around the same point (denoted by P) which lies on the line of the rear axle. It follows that the kinematics of the car can be fully described by the kinematics of a bicycle fitted in the middle of the car (see attached image. The coordinates of the rear axle midpoint are given by $x$ and $y$. The orientation of the car with respect to the axis of $x$ is denoted by 9. The angle of the front wheel of the bicycle with respect to the longitudinal symmetry axis of the car is denoted by $φ$ . One may consider $φ$ or its time derivative $u_2=\dot{φ}$ as input. The longitudinal velocity of the rear axle midpoint is denoted by $u_1$ if it is a control input (two input case) and by $v_{car}$ if not (one input case). All lengths involved in the kinematic calculations, and in particular $l$, equal to one.

$$\begin{align*} \dot{x} &= u_1 \cos{\theta},\\ \dot{y} &= u_1 \sin{\theta},\\ \dot{\theta} &= u_1 \tan{\varphi},\\ \dot{\varphi} &= u_2. \end{align*}$$

Type: 

Form: 

Model order: 

4

Time domain: 

Linearity: 

Attachment: 

Publication details: 

TitleOn-line time-scaling control of a kinematic car with one input
Publication TypeConference Paper
Year of Publication2007
AuthorsKiss, B., and Szadeczky-Kardoss E.
Conference NameMediterranean Conference on Control & Automation, 2007.
Date Published06/2007
PublisherIEEE
Conference LocationAthens
ISBN Number978-1-4244-1281-5
Accession Number9871515
Keywordsautomobiles, steering systems, tracking, vehicle dynamics
AbstractThis paper reports a time-scaling scheme to realize a tracking controller for the non-differentially flat model of the kinematic car with one input which is the steering angle or the angular velocity of the steering angle. The longitudinal velocity of the car is a measurable external signal and cannot be influenced by the controller. Using an on-line time-scaling, driven by the longitudinal velocity of the car, by a scaling output of the tracking controller, and by their time derivatives up to the second order, one can achieve exponential tracking of any sufficiently smooth reference trajectory, similar to the differentially flat case with two control inputs. The price to pay is the modification of the finite traveling time of the reference trajectory according to the time-scaling.
DOI10.1109/MED.2007.4433947

Pages