Fuzzy Self Tuning of PID Controller for Active Suspension System

A S. Emam

Advances in Powertrains and Automotives OPEN ACCESSPEER-REVIEWED

Fuzzy Self Tuning of PID Controller for Active Suspension System

A S. Emam

Department of Automotive and Tractors Engineering, Faculty of Engineering, Mataria, Helwan University, Cairo, Egypt

Abstract

Suspension system plays an imperative role in retaining the continuous road wheel contact for better road holding. In this paper, fuzzy self-tuning of PID controller is designed to control of active suspension system for quarter car model. A fuzzy self-tuning is used to develop the optimal control gain for PID controller (proportional, integral, and derivative gains) to minimize suspension working space of the sprung mass and its change rate to achieve the best comfort of the driver. The results of active suspension system with fuzzy self-tuning PID controller are presented graphically and comparisons with the PID and passive system. It is found that, the effectiveness of using fuzzy self-tuning appears in the ability to tune the gain parameters of PID controller

Cite this article:

  • A S. Emam. Fuzzy Self Tuning of PID Controller for Active Suspension System. Advances in Powertrains and Automotives. Vol. 1, No. 1, 2015, pp 34-41. http://pubs.sciepub.com/apa/1/1/4
  • Emam, A S.. "Fuzzy Self Tuning of PID Controller for Active Suspension System." Advances in Powertrains and Automotives 1.1 (2015): 34-41.
  • Emam, A. S. (2015). Fuzzy Self Tuning of PID Controller for Active Suspension System. Advances in Powertrains and Automotives, 1(1), 34-41.
  • Emam, A S.. "Fuzzy Self Tuning of PID Controller for Active Suspension System." Advances in Powertrains and Automotives 1, no. 1 (2015): 34-41.

Import into BibTeX Import into EndNote Import into RefMan Import into RefWorks

At a glance: Figures

1. Introduction

Ride comfort and vehicle handling have become two of the important criteria in a passenger vehicle which is directly related to driver fatigue, discomfort, and safety. The purpose of the suspension systems are attempted to isolate vehicle vibration excitations from being transmitted to the drivers and to improve passenger comfort [1, 2]. Conventional suspension system has fix criteria of spring and damping. These two mechanical components have some limitation in isolating vibration especially in various types of road profiles [3]. Also, in order to improvement the passenger comfort, the passive elements (springs and dampers) must be selected in soft part, but in order to increase the qualification of driving the passive elements must be selected in harder part. This discrepancy in selection the type of passive elements indicates that, we cannot attain to both two objectives by using the passive suspension system [4, 5]. Therefore, the suspension system needs to change the system specifications in a dynamic aspect according to the conditions of the road. So, several semi-active, active, and fully active suspension approach have been designed and built to address this problem.

Nowadays, many researchers give more attention in proposed and implemented various semi-active and active vehicle suspension systems both theoretically as well experimentally into real-time application. Hac [6] derived the preview control problem of an active vehicle suspension with a quarter-car model and showed that the suspension performance could be improved by using preview information. Gordon [7] proposed the optimal control of a semi-active vehicle suspension system for quarter-vehicle model. Analysis is based on a simple fifth-order mathematical model of the ride dynamics of an automotive suspension. The semi-active control of vehicle suspension system with magnetorheological damper was presented by Yao et al. [8]. Bouazara and Richard [9] analyzed the effects of vibrations on comfort and road holding capability of road vehicles as observed in the variation of different parameters such as suspension coefficients, road disturbances and the seat position. Du et al. [10] studied the application of MR dampers to semi-active of vehicle suspension. A multi-objective genetic algorithm for assessing the optimality of control algorithms for semi-active vehicle was developed by Crews et al. [11]. The evaluation of experimental performance for an LQ-based semi-active and controller concept was investigated by Unger et al. [12]. Li et al. [13] proposed a multi-objective control method for active vehicle suspension of four-degree-of-freedom half-car model.

Recently, various control strategies have been proposed to control the activeand active suspension system. Linear Quadratic Gaussian (LQG) control, adaptive control, robust control, and non-linear control are developed and proposed to manage the occurring problems [14, 15, 16]. Multi-objective functional such as H2, H, GL2, GH2, GA, etc. control of vehicle suspensions attracts more attention many researchers [17, 18, 19, 20]. Fuzzy logic control (FLC) and PID controller approach has been fruitful research area with semi-active and active suspension system [21, 22, 23, 24].

The main objective of this paper, we present a multi-objective control for the active suspension system for quarter car model by using PID controller. Previous methods are used to determine the gains of PID do not guarantee the best performance of the system, which support the important of fuzzy self-tuning method to obtain the parameters of PID for the best Performance. A fuzzy self-tuning is used to develop the optimal control parameters for PID controller to minimize suspension working space of the sprung mass and its change rate to achieve the best comfort of the driver.

2. Mathematical Model Formulation

2.1. Active Suspension System

This section is devoted to the mathematical modeling of proposed model. Figure 1 shows the two-degrees-of-freedom system that represents the quarter-vehicle active suspension model. It consists of an upper mass , representing the body mass (sprung mass), as well as a lower mass , representing the wheel mass (un-sprung mass), and its associated parts. The vertical motion of the system is described by the displacements , and , while the excitation due to road disturbance is . The suspension spring constant is , damping coefficient is , and the tyre spring constant is (the tyre damping is neglected) . The data employed here for the quarter-vehicle system are listed in Table 1 [25].

Table 1. Quarter vehicle suspension parameters [25]

By applying Newton’s second law the equations of motion for the sprung and unsprung masses of the quarter-car suspension model are given by [26]:

(1)
(2)

: is the control force from the hydraulic actuator.

The equations of motion can be summarized by using the following notation:

The proposed active suspension system can be represented in the state space notation:

Where, the matrices , , and as:

The output equation y can be calculated by using the matrices to:

Where, .

For the active suspension, the actuator force depends on the time histories of the relative displacement x across it:

Where, is also referred to as the suspension working space (SWS). In this paper, it is assumed that only the suspension working space could be measured and used by the controllers.

2.2. Road Excitation Profile

Road profile is considered by single bump, chosen to be very similar to the real-world road profiles. The road bump was described in reference [27] as:

Where, a is the half the bump amplitude, , d is the width of bump, and V is the vehicle velocity. In this study, a=0.035 m, d=0.8m, and V= 0.856 m/s, as in reference [28].

3. Fuzzy Self Tuning of PID Design Procedure

Tuning of a PID controller is to tune or choose numerical values for the PID coefficients are proportional (Kp), integral (Ki) and derivative (Kd) to achieve an optimized value of the desired response. The proposed design procedure includes two steps:

1. Find the optimal gains of PID which are controlled the system.

2. Design Fuzzy logic control (FLC) part, which has self tuning capabilities for the obtained PID.

3.1. Optimal PID Controller

Ziegler and Nichols method to tune the coefficients of a PID controller became the focus of research and become better understood, but cannot guarantee to be always effective. For this reason, this paper investigates the design of self tuning for a PID controller. The transfers function of the PID controller given as [29]:

(3)

Where, is proportional gain, is the integral gain, is the derivative gain.

Figure 2 shows the block diagram of PID controller, and Figure 3 illustrates the matlab/simulink model of active suspension system with PID controller.

Figure 3. Simulink of active suspension system with PID controller

The proportional, integral and derivative terms are summed to calculate the output of the PID controller. The final output defined by u(t) and it given by [29]:

Where, is the error present in the controller, t is the time or instantaneous time, and x is the variable of integration, taken from zero to present one.

3.2. Designs Steps of Fuzzy Self Tuning for the PID Controller

In this section the Fuzzy Self Tuning for the PID Controller is designed. The PID fuzzy controller can be decomposed into the equivalent proportional control, integral control and the derivative control components. The designs steps of fuzzy self tuning can be summarized as follows [30]:

Write the PID controller by the following equation:

This equation can be written as:

Where, , and Kp1, KI1 , KD1 are the gain outputs from fuzzy controller.

•  Put the input membership functions of e and as linguistic labels. The inputs are {Negative Big, , Negative medium, Negative small, Zero, Positive small, Positive medium, Positive Big}, and are referred to in the rules bases as {NB,NM,NS,ZE,PS,PM,PB}, and the linguistic labels of the outputs are {Negative Big, Negative medium, Negative small, Zero, Positive small, Positive medium, Positive Big} and referred to in the rules bases as {NB,NM,NS,ZE,PS,PM,PB }. There are 49 rules in the rule base used as presented in Table 2.

•  Calculate the values of normalizing gain by dividing the boundary values of the universe of discourse of the input member ship functions of e and and by the maximum of the original values of e and , respectively. This is to normalize them within the boundaries of the membership universe of discourse.

•  Apply the defuzzification process as final step to converts the fuzzy output to crisp value to be used as a non-fuzzy control action.

•  Choose the most popular defuzzification method that is the center of gravity or center of area which is formulated as follows:

where u(ui) is the membership grad (weight) of the element ui which is the output of the rule.

This paper proposed two inputs-three outputs self tuning of a PID controller. The controller design used the error and rate of error as inputs to the fuzzy self tuning, and the gains as outputs. The FLC is adding to the PID controller to adjust the parameters of the PID controller on-line according to the change of the signals error and change of the error. The controller proposed also contain scaling gains inputs as shown in Figure 4. Figure 5 illustrates the matlab/simulink model of active suspension system with fuzzy self tuning controller. Figure 6 shows the input and output relations in the form of Cartesian rule surfaces.

Figure 5. Simulink of active suspension system with fuzzy self tuning PID controller

4. Results and Discussion

Figure 7-Figure 10 present the selected results of active suspension system for self tuning fuzzy PID comparison with the normal PID and passive suspension model. Figure 7 shows, the suspension working space for self tuning fuzzy PID controller with various times. The obtained results were compared with passive model and normal PID controller. It is clear that, the active suspension system with self tuning fuzzy PID gives lower value of suspension working space. Also, the suspension working space for self tuning fuzzy PID controller reaches steady state faster than normal PID controller, which faster than passive suspension.

Figure 8 presents, the body acceleration for self tuning fuzzy PID controller compared with passive and normal PID controller. It is observed that, the active suspension system with self tuning fuzzy PID gives lower value of vertical peak as compared with passive suspension and normal PID controller. This result indicated to improve the driver comfort and better potential.

Figure 6. Input and output relations in the form of Cartesian rule surfaces for P, I, and D parameters. (a) P parameter; (b) I parameter; (c) D parameter
Figure 7. Suspension working space (SWS) for self tuning fuzzy PID controller compared with passive and normal PID controller
Figure 8. Vehicle body acceleration (for self tuning fuzzy PID controller compared with passive and normal PID controller

Figure 9 shows, the dynamic tyre load for self tuning fuzzy PID controller compared with passive and normal PID controller. It is clear that, the active suspension system with self tuning fuzzy PID gives lower value of dynamic tyre load as compared with passive suspension and normal PID controller. This result indicated the self tuning fuzzy PID controller has lower values of maximum over shoots, which is directly related to driver safety.

Figure 9. Dynamic tyre load (DTL) for self tuning fuzzy PID controller compared with passive and normal PID controller

Figure 10 presents, the wheel displacement of self tuning fuzzy PID controller compared with passive and normal PID controller. The obtained result indicates that, the active suspension system with self tuning fuzzy PID gives lower values of maximum over shoots and lower. Also, It is clear that wheel displacement with self tuning fuzzy PID controller reaches steady state faster than normal PID controller.

Figure 10. Wheel displacement for self tuning fuzzy PID controller compared with passive and normal PID controller
Figure 11. The generated control force of fuzzy self tuning controller

Figure 11 shows the generated control force for fuzzy self tuning controller. The results obtained by self tuning fuzzy PID controller were compared to those passive and normal PID controller in terms of Roots Mean Square (RMS) values are illustrates in Table 3. It can be observe that, the fuzzy self tuning pid controller gives better results as compared to pid controller and passive suspension.

Table 3. QUARTER VEHICLE RIDE COMFORT PARAMETERS

5. Conclusion

In this paper, a fuzzy self-tuning is designed to develop the optimal control gain for PID controller for a quarter car model of a passenger car to improve the ride comfort and road holding ability. The following points have been concluded:

•  The results and plot indicate that self tuning fuzzy PID controller are less oscillatory, and have lower values of maximum over shoots, compared with passive and normal PID controller. This is directly related to improve driver fatigue, discomfort, and safety.

•  It is found that, self tuning fuzzy PID controller reaches steady state faster than normal PID controller, which faster than passive suspension for all the dynamic performance.

•  Also, it can be observed that, the value of suspension working space has been reduced by sprung mass displacement has been reduced by 30.6 % which shows the improvement in ride comfort, sprung mass acceleration reduced by 11.8 % and dynamic tire load reduced by 18.3 %.

References

[1]  W. Abbas, O. B. Abouelatta, M. El-Azab, M. Elsaidy, A. A. Megahed, and others, “Optimization of biodynamic seated human models using genetic algorithms,” Engineering, vol. 2, pp. 710-719, 2010.
In article      View Article
 
[2]  W. Abbas, O. B. Abouelatta, M. S. El-Azab, and A. A. Megahed, “Application of Genetic Algorithms to the Optimal Design of Vehicle's Driver-Seat Suspension Model,” in Proceedings of the World Congress on Engineering, 2010, vol. II, June 30 - July 2, 2010, London, U.K.
In article      
 
[3]  R. ROSLI, M. MAILAH, and G. PRIYANDOKO, “Active Suspension System for Passenger Vehicle using Active Force Control with Iterative Learning Algorithm,” WSEAS Trans. Syst. CONTROL, vol. 9, pp. 120-129., 2014.
In article      
 
[4]  W. Abbas, A. Emam, S. Badran, M. Shebl, and O. Abouelatta, “Optimal Seat and Suspension Design for a Half-Car with Driver Model Using Genetic Algorithm,” Intell. Control Autom., vol. 4, pp. 199, 2013.
In article      View Article
 
[5]  H. Hashemipour, “Nonlinear Optimal Control of Vehicle Active Suspension Considering Actuator Dynamics.” International Journal of Machine Learning and Computing, vol. 2, pp. 355-359, 2014.
In article      
 
[6]  A. HAĆ, “Optimal linear preview control of active vehicle suspension,” Veh. Syst. Dyn., vol. 21, pp. 167-195, 1992.
In article      View Article
 
[7]  T. J. Gordon, “Non-linear optimal control of a semi-active vehicle suspension system,” Chaos Solitons Fractals, vol. 5, pp. 1603-1617, 1995.
In article      View Article
 
[8]  G. Z. Yao, F. F. Yap, G. Chen, Wh. Li, and S. H. Yeo, “MR damper and its application for semi-active control of vehicle suspension system,” Mechatronics, vol. 12, pp. 963-973, 2002.
In article      View Article
 
[9]  M. Bouazara and M. J. Richard, “An optimal design method to control the vibrations of suspensions for passenger cars,” in International Mechanical Engineering Congress and Exposition: The Winter Annual Meeting of ASME, Atlanta, DSC, 1996, vol. 58, pp. 61-68.
In article      
 
[10]  H. Du, K. Yim Sze, and J. Lam, “Semi-active H∞ control of vehicle suspension with magneto-rheological dampers,” J. Sound Vib., vol. 283, pp. 981-996, 2005.
In article      View Article
 
[11]  J. H. Crews, M. G. Mattson, and G. D. Buckner, “Multi-objective control optimization for semi-active vehicle suspensions,” J. Sound Vib., vol. 330, pp. 5502-5516, 2011.
In article      View Article
 
[12]  A. Unger, F. Schimmack, B. Lohmann, and R. Schwarz, “Application of LQ-based semi-active suspension control in a vehicle,” Control Eng. Pract., vol. 21, pp. 1841-1850, 2013.
In article      View Article
 
[13]  P. Li, J. Lam, and K. C. Cheung, “Multi-objective control for active vehicle suspension with wheelbase preview,” J. Sound Vib., vol. 333, pp. 5269-5282, 2014.
In article      View Article
 
[14]  T. E. Duncan, L. Guo, and B. Pasik-Duncan, “Adaptive continuous-time linear quadratic Gaussian control,” Autom. Control IEEE Trans. On, vol. 44, pp. 1653-1662, 1999.
In article      View Article
 
[15]  S. Chantranuwathana and H. Peng, “Adaptive robust force control for vehicle active suspensions,” Int. J. Adapt. Control Signal Process., vol. 18, pp. 83-102, 2004.
In article      View Article
 
[16]  H. Pan, W. Sun, H. Gao, T. Hayat, and F. Alsaadi, “Nonlinear tracking control based on extended state observer for vehicle active suspensions with performance constraints,” Mechatronics, vol. In Press, Corrected Proof, Available online 27 August 2014, 2014.
In article      
 
[17]  J. Wang and D. A. Wilson, “Mixed GL2/H2/GH2 control with pole placement and its application to vehicle suspension systems,” Int. J. Control, vol. 74, pp. 1353-1369, 2001.
In article      View Article
 
[18]  W. Abbas, O. B. Abouelatta, M. El-Azab, M. El-Saidy, and A. A. Megahed, “Optimal Seat Suspension Design Using Genetic Algorithms,” J. Mech. Eng. Autom., vol. 1, pp. 44-52, 2011.
In article      
 
[19]  L. Zuo and S. A. Nayfeh, “Structured H2 optimization of vehicle suspensions based on multi-wheel models,” Veh. Syst. Dyn., vol. 40, pp. 351-371, 2003.
In article      View Article
 
[20]  M. A. Eltantawie, “Decentralized neuro-fuzzy control for half car with semi-active suspension system,” Int. J. Automot. Technol., vol. 13, no. 3, pp. 423-431, 2012.
In article      View Article
 
[21]  M. M. M. Salem and A. A. Aly, “Fuzzy control of a quarter-car suspension system,” World Acad. Sci. Eng. Technol., vol. 53, pp. 258-263, 2009.
In article      
 
[22]  J. Fengi, J. Li, F. Yui, “GA-based PID and fuzzy logic control for active vehicle suspension system,” International Journal of Automotive Technology, vol. 4, pp. 181-191, 2003.
In article      
 
[23]  D. Lee, J. Allan, H. A. Thompson, and S. Bennett, “PID control for a distributed system with a smart actuator,” Control Eng. Pract., vol. 9, pp. 1235-1244, 2001.
In article      View Article
 
[24]  I. Eski and Ş. Yıldırım, “Vibration control of vehicle active suspension system using a new robust neural network control system,” Simul. Model. Pract. Theory, vol. 17, pp. 778-793, 2009.
In article      View Article
 
[25]  S. Türkay and H. Akçay, “A study of random vibration characteristics of the quarter-car model,” J. Sound Vib., vol. 282, pp. 111-124, 2005.
In article      View Article
 
[26]  J.-S. Chiou and M.-T. Liu, “Using fuzzy logic controller and evolutionary genetic algorithm for automotive active suspension system,” Int. J. Automot. Technol., vol. 10, pp. 703-710, 2009.
In article      View Article
 
[27]  S.-B. Choi and W.-K. Kim, “Vibration control of a semi-active suspension featuring electrorheological fluid dampers,” J. Sound Vib., vol. 234, pp. 537-546, 2000.
In article      View Article
 
[28]  H. Chen, Z.-Y. Liu, and P.-Y. Sun, “Application of constrained H∞ control to active suspension systems on half-car models,” J. Dyn. Syst. Meas. Control, vol. 127, pp. 345-354, 2005.
In article      View Article
 
[29]  Lin, Y-J., Lu, Y-Q., Padovan, J., Fuzzy logic control of vehicle suspension systems, International Journal of Vehicle Design, 14, 457- 470, 1993.
In article      
 
[30]  M. M. Algreer and Y. R. Kuraz, “Design Fuzzy Self Tuning of PID Controller for Chopper-Fed DC Motor Drive,” Al-Rafidain Eng., vol. 16, pp. 54-66, 2008.
In article      
 
  • CiteULikeCiteULike
  • MendeleyMendeley
  • StumbleUponStumbleUpon
  • Add to DeliciousDelicious
  • FacebookFacebook
  • TwitterTwitter
  • LinkedInLinkedIn