Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

A New Fuzzy-Evidential Controller for Stabilization of the Planar Inverted Pendulum System

A New Fuzzy-Evidential Controller for Stabilization of the Planar Inverted Pendulum System

  • Yongchuan Tang, 
  • Deyun Zhou, 
  • Wen Jiang
PLOS
x

Abstract

In order to realize the stability control of the planar inverted pendulum system, which is a typical multi-variable and strong coupling system, a new fuzzy-evidential controller based on fuzzy inference and evidential reasoning is proposed. Firstly, for each axis, a fuzzy nine-point controller for the rod and a fuzzy nine-point controller for the cart are designed. Then, in order to coordinate these two controllers of each axis, a fuzzy-evidential coordinator is proposed. In this new fuzzy-evidential controller, the empirical knowledge for stabilization of the planar inverted pendulum system is expressed by fuzzy rules, while the coordinator of different control variables in each axis is built incorporated with the dynamic basic probability assignment (BPA) in the frame of fuzzy inference. The fuzzy-evidential coordinator makes the output of the control variable smoother, and the control effect of the new controller is better compared with some other work. The experiment in MATLAB shows the effectiveness and merit of the proposed method.

1 Introduction

The planar inverted pendulum system [1, 2] was generalized from the linear inverted pendulum [3, 4]. Both of them are self-unstable, high-order, multi-variable and nonlinear system, while the planar inverted pendulum system is more sophisticated than the linear inverted pendulum because of a higher degree of freedom and a stronger coupling in both axes. The detailed mathematical model and the introduction to the physical structure of the planar inverted pendulum are presented in [5]. The stability of the planar inverted pendulum system in the sense of Lyapunov is proved in [6] based on the mathematical model of the system. The planar inverted pendulum system is an effective benchmark to verify the effectiveness of a control method. Researches on stabilization of the planar inverted pendulum system are of great help to some key technologies in rocket, missile and robotics control methods, for example, the equilibrium of a humanoid robot for its bipedal walking [7]. Some researches on the planar inverted pendulum have been reported in recent years based on various methods including the linear quadratic regulator (LQR) method [8, 9], the neural network control theory [1], the sliding mode control theory [5, 10] and the fuzzy control theory [2, 11, 12].

Fuzzy control theory [13, 14] is based on fuzzy sets [15] and fuzzy inference [16, 17]. So far, fuzzy inference has been extensively used in real applications including risk analysis [1822], controller design [2325], decision-making [26, 27], and so on [2830]. Similar to fuzzy sets theory, Dempster-Shafer theory of evidence or evidence theory [31, 32] is effective in uncertain representation and data fusion [33, 34]. Evidential reasoning [3537] is effective in dealing with problems related to decision-making [38, 39], pattern recognition [40, 41], fault diagnosis and risk analysis [42, 43], human reliability analysis [44], environment protection [45], and so on [4648]. Both fuzzy inference and evidential reasoning are effective ways for empirical knowledge representation and process. There are already some studies incorporate fuzzy sets theory with evidence theory for controller design [23, 49, 50], classification [51], knowledge management [52], fault diagnosis [53], and so on [54]. In [23], Yager and Filev introduce evidence theory into the basic fuzzy system model to handle the probabilistic uncertainty of the consequent in fuzzy rules. With the combination of evidential reasoning and fuzzy inference, Graham [49] fuses the sensory information of the robot’s environment to fulfill the collision avoidance control of robots. An expert system is designed with a fuzzy logic controller in [50], while the output of the expert system for decision making is suggested with a belief level accomplished by evidential reasoning.

The existing researches show that a hybrid intelligent approach can make each conventional intelligent algorithm more effective and flexible for dealing with the problems in real applications [5256]. In this paper, a fuzzy-evidential controller for the planar inverted pendulum system is proposed. This work not only addresses the uncertainty in the consequent of fuzzy rules as it is in [23], but also coordinates the coupling from different fuzzy controllers with the proposed fuzzy-evidential coordinator. The empirical knowledge for stabilization of the planar inverted pendulum system is expressed by fuzzy rules while designing the fuzzy nine-point controller, and a coordinator for different control variables is designed based on the dynamic basic probability assignment (BPA) in the frame of fuzzy inference. The control strategy in the proposed controller doesn’t depend on the mathematical model of the controlled system, which means a shorter developing period and lower cost in real application. The experiment in MATLAB shows that the proposed method is effective. In addition, the proposed fuzzy-evidential controller forces the planar inverted pendulum into its equilibrium state faster than the fuzzy controller [12] and the LQR method [8].

The rest of this paper is organized as follows. In Section 2, some preliminaries are briefly introduced. In Section 3, a fuzzy-evidential controller for the planar inverted pendulum system is presented. In Section 4, the control experiment in MATLAB verifies the effectiveness and the merit of the proposed method. The conclusions are given in Section 5.

2 Preliminaries

Some basic concepts of fuzzy control theory and evidence theory are presented in this section, as well as an introduction to the planar inverted pendulum system.

2.1 Fuzzy control theory

The fuzzy control theory [13, 14] is based on the fuzzy sets [15] and fuzzy inference [16, 17]. Some key steps of designing a fuzzy controller are shown as follows [13, 14].

  • Step 1. Defining the structure of the fuzzy controller.
  • Step 2. Defining the input and output fuzzy sets.
  • Step 3. Defining the membership function of input and output function.
  • Step 4. Defining the fuzzy control rules and the fuzzy inference process.
  • Step 5. Solving the fuzziness with defuzzifier.

A fuzzy set [1315] A in X is a set of ordered pairs, denoted as A = {(x, uA (x))}, where X denotes a collection of objects (points) denoted by x, and uA(x) is the membership grade of x in A, uA: XM is a function from X to the membership space M.

The equation of singleton fuzzifier is shown as follows [57, 58]: (1) where the point x* ∈ X is mapped as a singleton in X, and its grade of membership is 1, while the others are 0. The equation of weighted average defuzzifier is shown as follows [57, 58]: (2) where viV, and V is the output space of a fuzzy controller, M is the membership space, wi is a ratio corresponding to the response characteristic of a system.

2.2 Dempster-Shafer theory of evidence

In this section, some basic concepts of the evidence theory [31, 32] are introduced, including the frame of discernment, basic probability assignment (BPA) and Dempster’s rule of combination.

Frame of discernment.

Evidence theory assumes a finite nonempty set of mutually exclusive events Θ = {θ1,θ2, …,θn}, a power set 2|Θ| is defined as the frame of discernment, shown as follows [31, 32]: (3)

Basic probability assignment.

The basic probability assignment(BPA) function or mass function m is defined as a mapping from the power set of Θ to a number between 0 and 1, which satisfies [31, 32]: (4) where ∅ is an empty set, A is any subsets of Θ, the mass function m(A) represents how strongly the evidence supports A. The mass m(Θ) represents the uncertainty of the evidence.

Dempster’s rule of combination.

Dempster’s rule of combination combines two BPAs in a way that the new BPA represents a consensus of the contributing pieces of evidence, it sets intersection putting the emphasis on the common elements of evidence. Dempster’s rule of combination is the orthogonal sum of m1 and m2, denoted by (m1m2), shown as follows [31, 32]: (5) where A, B, and C are subsets of 2|Θ|, k is a normalization constant representing the conflict coefficient of two BPAs, k is defined as follows [31, 32]:(6)

Pignistic probability transformation.

Pignistic probability transformation BetP(⋅) in the transferable belief model (TBM) [59] is commonly chosen to transform a BPA to probability distribution. BetP(A) is defined as follows: (7) where A ∈ 2θ, |A| is the cardinality of subset A.

For detailed information related to evidence theory, one can refer [31, 32, 46, 59].

2.3 The planar inverted pendulum system

The planar inverted pendulum system [1, 2] consists of a rod, a cart and two rails in the orthogonal axes, as is shown in Fig 1. The cart can move along the rail in Y axis, and the rail in Y axis can move along the rail in X axis, so the cart can move within the planar plane XOY. The rod is set on the cart by a Hooke’s joint, it can rotate around X axis and Y axis. Normally, the pendulum deviates from OZ direction with θx in XOZ plane, and θy in YOZ plane. Fig 1 also shows the positive direction for each axis and each angle with the plus symbol (+), so the unmarked direction is the negative direction. The stabilization of the planar inverted pendulum system includes the equilibrium of the rod accomplished by the pendulum controller and controlling the cart back to the origin of coordinates by the cart controller in both X-axis and Y-axis.

According to the control logic, if the pendulum deviates from the upright position clockwise (Clockwise is chosen as the positive direction), the pendulum controller should control the cart move toward the right direction (Defining on the left is the positive direction), in this way, the pendulum will go back toward its equilibrium position by rotating anticlockwise. On the contrary, if the pendulum deviates from the upright position anticlockwise, the pendulum controller should control the cart move toward the left direction, and the pendulum will go back to its equilibrium position by rotating clockwise.

The cart controller will control the cart indirectly, this is because the equilibrium of the rod is the foundation of the stabilization of the whole system, so controlling the cart back to the origin of coordinates has a lower priority level than the rod’s equilibrium. If the cart leaves the original point toward the positive direction, the cart controller should keep its movement to the positive direction, this behavior will lead the pendulum deviate from the upright position anticlockwise, so the pendulum controller will control the cart move toward the negative direction, thus the pendulum will rotate clockwise and go back to its equilibrium position, also, the cart will go back to its original position automatically. If the cart leaves the original point toward the negative direction, the control logic will be on the contrary.

More detailed information about the mathematical model and the physical structure of the planar inverted pendulum is presented in [1, 2, 5, 8], and the stability of the system in the sense of Lyapunov is proved in [6].

3 Fuzzy-evidential controller

For each axis of the planar inverted pendulum system, the control mode is assumed to be the same, which means the controller for each axis is designed with the same way. The control force for each axis comes from a single motor, however, according to the control logic analyzed above, there should be two controllers in each axis, a controller for the rod and a controller for the cart. Thus, in each axis, after designing a fuzzy controller for the rod and another one for the cart, respectively, a coordinator for these two controllers is necessary to output the final control force of the motor, denoted as F. The function of the control force F, which is similar to the adaptive sliding-mode control in [5], is shown as follows: (8) where m1 and m2 are the output of the designed coordinator based on fuzzy inference and evidential reasoning, which will be shown in detail in section 3.2. The F1 and F2 are the output of the fuzzy nine-point controller of the rod and the cart, which will be shown in detail in section 3.1.

The proposed fuzzy-evidential controller for the planar inverted pendulum system is shown in Fig 2. In Fig 2, θ is the angular deviation of the rod, and its derivative means the angular speed. x is the displacement deviation of the cart, and its derivative means the speed of the cart.

thumbnail
Fig 2. Fuzzy-evidential controller for the planar inverted pendulum system.

https://doi.org/10.1371/journal.pone.0160416.g002

3.1 Fuzzy controller of the rod and cart

Multi-point controller is a kind of direct and simple control strategy based on control experience [60], and a nine-point controller [61] can be regarded as a special case of the multi-point controller. In this paper, the fuzzy nine-point controller is chosen to be the fuzzy controller for the rod and the cart in Fig 2. Take the control variable θ for example, the principle of the nine-point controller is shown as Fig 3.

In Fig 3, a phase plane consists of the control variable θ and its gradient is divided into nine parts by the zero zone value θ0, −θ0, and . There are nine principles in fuzzy nine-point controller [61]. If the system state is θ > θ0 and , which means the angular deviation of the rod is very big in positive direction, so the system needs the strongest control effect f4+ in positive direction to enlarge the output of the rod’s fuzzy controller and reduce the deviation of the rod. The other eight control zones can be under control in a similar way, the rest eight principles are shown as follows.

  • If the system state is θ > θ0 and , applying a strong control effect f3+ in positive direction to the system.
  • If the system state is θ > θ0 and , applying a weak control effect f2+ in positive direction to the system.
  • If the system state is |θ| ≤ θ0 and , applying the weakest control effect f1+ in positive direction to the system.
  • If the system state is θ < −θ0 and , applying the strongest control effect f4− in negative direction to the system.
  • If the system state is θ < −θ0 and , applying a strong control effect f3− in negative direction to the system.
  • If the system state is θ < −θ0 and , applying the weak control effect f2− in negative direction to the system.
  • If the system state is |θ| ≤ θ0 and , applying the weakest control effect f1− in negative direction to the system.
  • If the system state is |θ| ≤ θ0 and , applying the maintaining control effect f0.

Based on the principle of fuzzy nine-point controller, the fuzzy controller of the rod is shown in Table 1, as well as the fuzzy controller of the cart in Table 2. Note that, in Table 2, the positive direction of the parameter is contrary to Table 1, since the fuzzy controller of the cart is an indirect controller, as is explained by the control logic of the system in Section 2.3.

According to Figs 1 and 2 and the control strategy mentioned above, for each axis there is only one motor to fulfill the fuzzy control strategy between the rod and the cart. So, a coordinator is needed. In this paper, the coordinator is based on fuzzy inference and evidential reasoning, named the fuzzy-evidential controller, as is shown in detail in the next subsection.

3.2 Fuzzy-evidential coordinator

In a close loop feedback control system, the values of control variables are often changing from time to time to keep the dynamic stability of the controlled system, such as the angle of the rod and the displacement of the cart in the planar inverted pendulum system. In this section, the dynamic mass function or dynamic BPA is designed to express the dynamic characteristic of the control variable in the frame of fuzzy inference.

In Fig 4, Θ = {A1,A2,A3,A4,A5) is defined as the frame of discernment with five events, where Ai (i = 1,2,3,4,5) is the fuzzy partition of |θ| and the value of |θ| ranges from a small value to a big one. In this paper, |θ| is defined as 0 ≤ |θ| ≤ 1. The membership function corresponding to |θ| in Fig 4 is defined as u(Ai) (i = 1,2,3,4,5), which satisfies: (9) For any given value θ in Fig 4, there are at least three zero values among the membership function u(Ai) (i = 1,2,3,4,5), which will lead to a jump value of a controller in real application, and the discontinuous value in a controller usually has a bad control effect on the controlled object. So the nonzero mass value for each single subset of 2|Θ| is constructed as follows: (10) subject to: (11) where BetP (Ai) is the Pignistic probability for the ith fuzzy partition, the mass function m(A1,A2,A3,A4,A5) is redistributed among all the single subset of 2|Θ| according to BetP (Ai). The k is a scale factor, and a big value of k is considered. Since, if the value of k is too small, a large value of the originally nonzero membership function in Fig 4 will be redistributed to the fuzzy partition with a zero value, which is contrary to the control logic. According to the control experience, it is recommended to define k as a constant parameter and k = 0.9. The Eq (11) satisfies Eq (4), which can be derived as follows: (12)

Recall the weighted average defuzzifier as Eq (2) and the singleton fuzzifier as Eq (1), remember the priority level of the rod is higher than the cart. Mathematically, the coordinator for the fuzzy controller of the rod and the cart is defined as follows: (13) where m1 is the weight coefficient for the fuzzy controller of the rod and m2 is the weight coefficient for the fuzzy controller of the cart. Since m(Ai) is time-varying because of the dynamic characteristics of the control system, m(Ai) is a dynamic BPA. The bi and b6−i are the equilibrium coefficients to balance the priority level between the fuzzy controller of the rod and the fuzzy controller of the cart. If the deviation of the angle |θ| is bigger than the zero zone value θ0, the fuzzy controller of the rod will work as the master controller which means bi is a big value and the value of b6−i will be a small one. Table 3 shows a recommended value of bi and the corresponding b6−i (i = 1,2,3,4,5) for the planar inverted pendulum system.

A numerical example.

This is an example to show how to calculate the fuzzy-evidential coordinator m1 and m2 in Eq (13).

Take the point α in Fig 4 as an example. The values of the membership function are u (A1) (α) and u (A2) (α). It can be proved that u (A1) (α) + u (A2) (α) = 1, so it satisfies Eq (9). According to Eqs (10) and (11), the dynamic mass function is calculated as follows: (14) Finally, with Eqs (13) and (14), the value of the fuzzy-evidential coordinator for the fuzzy controller of the rod and the cart is calculated as follows: (15)

4 Experiment

The stability of the planar inverted pendulum system in the sense of Lyapunov is proved in [6]. The purpose of this experiment is to verify the effectiveness of the proposed fuzzy-evidential controller. The experiment is realized in MATLAB, the nonlinear model of the planar inverted pendulum in reference [8] is chosen to be the model of the controlled system in simulation platform. Each fuzzy evidential controller, as well as the planar inverted pendulum is expressed as a S-function in MATLAB.

The results of the control experiment are shown in Figs 5, 6 and 7, where the solid line represents the control variable in X-axis and the star line represents the control variable in Y-axis, the control cycle of the close loop is 0.005 second.

In Fig 5, both in X-axis and Y-axis, the initial deviation of the rod from its upright position is 7 degree, which is about 0.122 radian. The zero zone value of the fuzzy controller for the rod in Table 1 is θ0 = 0.007 rad and rad/s in both X-axis and Y-axis. According to Fig 6, the zero zone value of the fuzzy controller for the rod in Table 1 is x0 = 0.018 m and m/s in both X-axis and Y-axis. After about 0.5 second, the displacement of the cart becomes stable, and keep adjustment near the zero point. The control variables in Figs 5 and 6 will go to a stable state in about 0.5 second, which means the stabilization of the controlled system. Compared with the control effect in reference, the method in this paper has a better control effect than the LQR method [8] and the fuzzy controller [12], the comparative result among these three methods is shown in Table 4. The overshoot of the angle and displacement in the proposed method is smaller than the LQR method in [8], so the adjusting time of the proposed method is shorter than LQR. Also, there is less zero value in the proposed method, and the control variable is smoother than the classical fuzzy controller in [12], which means the fuzzy-evidential coordinator is more effective, so the time to the stabilization needed by the controlled system decreases obviously.

Fig 7 shows the control effect in each axis, which is an acceleration control variable comes from the motor of each axis. In Fig 7, the control effect in each axis is a big value at first, then it decreases to a small value and maintain fine tuning, this is in harmony with Figs 5 and 6. The final control effect for X-axis ranges from [-45,35] and [-35,45] for Y-axis.

The data of the experiment results can be found in Supporting Information.

5 Conclusion

This paper proposes a new fuzzy-evidential controller for the stabilization of the planar inverted pendulum system. After designing a fuzzy controller for the rod and the cart, a fuzzy-evidential coordinator for these two controllers is proposed based on fuzzy inference and evidential reasoning. The experimental result shows the effectiveness of the proposed method, as well as the merit that the new controller can force the system into its stable state faster than the method without evidential reasoning or based on LQR in the references. The following work includes extending this method to other application areas, as well as the more complicated inverted pendulum, i.e. the double and triple planar inverted pendulum.

Supporting Information

S1 Table. The data of the experiment results.

https://doi.org/10.1371/journal.pone.0160416.s001

(XLSX)

Acknowledgments

The authors would like to thank the reviewers and the editor for their valuable comments and suggestions to improve this work. The work is partially supported by National Natural Science Foundation of China (Grant No. 60904099) and Natural Science Basic Research Plan in Shaanxi Province of China (Grant No. 2016JM6018).

Author Contributions

Conceptualization: Yongchuan Tang, Deyun Zhou, Wen Jiang. Data curation: Yongchuan Tang. Formal analysis: Yongchuan Tang, Deyun Zhou, Wen Jiang. Funding acquisition: Wen Jiang. Investigation: Yongchuan Tang. Methodology: Yongchuan Tang, Deyun Zhou, Wen Jiang. Project administration: Deyun Zhou. Resources: Deyun Zhou. Software: Yongchuan Tang. Supervision: Wen Jiang. Validation: Yongchuan Tang, Deyun Zhou, Wen Jiang. Writing - original draft: Yongchuan Tang. Writing - review & editing: Yongchuan Tang.

References

  1. 1. Wai RJ, Chang LJ. Stabilizing and tracking control of nonlinear dual-axis inverted-pendulum system using fuzzy neural network. Fuzzy Systems, IEEE Transactions on. 2006;14(1):145–168.
  2. 2. Wai RJ, Kuo MA, Lee JD. Design of cascade adaptive fuzzy sliding-mode control for nonlinear two-axis inverted-pendulum servomechanism. Fuzzy Systems, IEEE Transactions on. 2008;16(5):1232–1244.
  3. 3. Wang LX. Stable adaptive fuzzy controllers with application to inverted pendulum tracking. Systems, Man, and Cybernetics, Part B: Cybernetics, IEEE Transactions on. 1996;26(5):677–691.
  4. 4. Muškinja N, Tovornik B. Swinging up and stabilization of a real inverted pendulum. Industrial Electronics, IEEE Transactions on. 2006;53(2):631–639.
  5. 5. Wai RJ, Chang LJ. Adaptive stabilizing and tracking control for a nonlinear inverted-pendulum system via sliding-mode technique. Industrial Electronics, IEEE Transactions on. 2006;53(2):674–692.
  6. 6. Shen J, Sanyal AK, Chaturvedi NA, Bernstein D, McClamroch H. Dynamics and control of a 3D pendulum. In: Decision and Control, 43rd IEEE Conference on. vol. 1. IEEE; 2004. p. 323–328.
  7. 7. Shin HK, Kim BK. Energy-efficient gait planning and control for biped robots utilizing the allowable ZMP region. Robotics, IEEE Transactions on. 2014;30(4):986–993.
  8. 8. Liu F, Tang Y, Qi Q. Stabilize the planar single inverted pendulum based on LQR. In: Automation and Logistics, 2011 IEEE International Conference on. IEEE; 2011. p. 238–242.
  9. 9. Zhang JL, Zhang W. LQR self-adjusting based control for the planar double inverted pendulum. Physics Procedia. 2012;24:1669–1676.
  10. 10. Howimanporn S, Parnichkun M. Control of an X-Y planar inverted pendulum using PSO-based SMC. International Journal of Robotics & Automation. 2015;30(4):310–321.
  11. 11. Wai RJ, Kuo MA, Lee JD. Cascade direct adaptive fuzzy control design for a nonlinear two-axis inverted-pendulum servomechanism. Systems, Man, and Cybernetics, Part B: Cybernetics, IEEE Transactions on. 2008;38(2):439–454.
  12. 12. Tang Y, Liu F, Qi Q, Yang Y. Stabilizing planar inverted pendulum system based on fuzzy nine-point controller. TELKOMNIKA Indonesian Journal of Electrical Engineering. 2014;12(1):422–432.
  13. 13. Chang SS, Zadeh LA. On fuzzy mapping and control. Systems, Man and Cybernetics, IEEE Transactions on. 1972;SMC-2(1):30–34.
  14. 14. Wang LX, Mendel JM. Fuzzy basis functions, universal approximation, and orthogonal least-squares learning. Neural Networks, IEEE Transactions on. 1992;3(5):807–814.
  15. 15. Zadeh LA. Fuzzy sets. Information & Control. 1965;8(65):338–353.
  16. 16. Zadeh LA. Toward a theory of fuzzy information granulation and its centrality in human reasoning and fuzzy logic. Fuzzy sets and systems. 1997;90(2):111–127.
  17. 17. Sabahi F, Akbarzadeh-T MR. A qualified description of extended fuzzy logic. Information Sciences. 2013;244:60–74.
  18. 18. Wen C, Wang Y, Xu X. Fuzzy Information Fusion Algorithm of Fault Diagnosis Based on Similarity Measure of Evidence. Lecture Notes in Computer Science. 2008;5264(11):506–515.
  19. 19. Liu HC, Liu L, Lin QL. Fuzzy failure mode and effects analysis using fuzzy evidential reasoning and belief rule-based methodology. Reliability, IEEE Transactions on. 2013;62(1):23–36.
  20. 20. Liu HC, You JX, You XY, Shan MM. A novel approach for failure mode and effects analysis using combination weighting and fuzzy VIKOR method. Applied Soft Computing. 2015;28:579–588.
  21. 21. Guo J. A risk assessment approach for failure mode and effects analysis based on intuitionistic fuzzy sets and evidence theory. Journal of Intelligent & Fuzzy Systems. 2016;30(2):869–881.
  22. 22. Akyar H. Fuzzy Risk Analysis for a Production System Based on the Nagel Point of a Triangle. Mathematical Problems in Engineering. 2016;2016(2016):3080679.
  23. 23. Yager RR, Filev DP. Including probabilistic uncertainty in fuzzy logic controller modeling using Dempster-Shafer theory. IEEE transactions on systems, man, and cybernetics. 1995;25(8):1221–1230.
  24. 24. Geng T, Lv Y, Wang M, Liu Y. Expert Self-Tuning Using Fuzzy Reasoning for Proportional-Integral-Derivative Controller. Journal of Computational and Theoretical Nanoscience. 2015;12(7):1287–1291.
  25. 25. Li NJ, Wang WJ, Hsu CCJ. Hybrid particle swarm optimization incorporating fuzzy reasoning and weighted particle. Neurocomputing. 2015;167:488–501.
  26. 26. Azadi M, Jafarian M, Saen RF, Mirhedayatian SM. A new fuzzy DEA model for evaluation of efficiency and effectiveness of suppliers in sustainable supply chain management context. Computers & Operations Research. 2014;54:274–285.
  27. 27. Sabahi F, Akbarzadeh-T MR. Introducing validity in fuzzy probability for judicial decision-making. International Journal of Approximate Reasoning. 2014;55(6):1383–1403.
  28. 28. Jiang W, Luo Y, Qin X, Zhan J. An improved method to rank generalized fuzzy numbers with different left heights and right heights. Journal of Intelligent & Fuzzy Systems. 2015;28(5):2343–2355.
  29. 29. Chou CC. A generalized similarity measure for fuzzy numbers. Journal of Intelligent & Fuzzy Systems. 2016;30(2):1147–1155.
  30. 30. Deng Y. Fuzzy Analytical Hierarchy Process Based On Canonical Representation on Fuzzy Numbers. Journal of computational analysis and applications. 2017;22(2):201–228.
  31. 31. Dempster AP. Upper and Lower Probabilities Induced by a Multi-valued Mapping. Annals of Mathematical Statistics. 1967;38(2):325–339.
  32. 32. Shafer G. A Mathematical Theory of Evidence. Princeton: Princeton University Press; 1976.
  33. 33. Jiang W, Wei B, Qin X, Zhan J, Tang Y. Sensor Data Fusion Based on a New Conflict Measure. Mathematical Problems in Engineering. 2016;(Accepted):5769061.
  34. 34. Jiang W, Wei B, Xie C, Zhou D. An evidential sensor fusion method in fault diagnosis. Advances in Mechanical Engineering. 2016;8(3):1–7.
  35. 35. Gordon J, Shortliffe EH. A method for managing evidential reasoning in a hierarchical hypothesis space. Artificial Intelligence. 1985;26(3):323–357.
  36. 36. Yang JB, Singh MG. An evidential reasoning approach for multiple-attribute decision making with uncertainty. Systems, Man and Cybernetics, IEEE Transactions on. 1994;24(1):1–18.
  37. 37. Yang JB, Xu DL. Evidential reasoning rule for evidence combination. Artificial Intelligence. 2013;205:1–29.
  38. 38. Fu C, Yang JB, Yang SL. A group evidential reasoning approach based on expert reliability. European Journal of Operational Research. 2015;246(3):886–893.
  39. 39. Chin KS, Fu C, Wang Y. A method of determining attribute weights in evidential reasoning approach based on incompatibility among attributes. Computers & Industrial Engineering. 2015;87:150–162.
  40. 40. Liu ZG, Pan Q, Dezert J, Martin A. Adaptive imputation of missing values for incomplete pattern classification. Pattern Recognition. 2016;52:85–95.
  41. 41. Liu ZG, Pan Q, Mercier G, Dezert J. A new incomplete pattern classification method based on evidential reasoning. Cybernetics, IEEE Transactions on. 2015;45(4):635–646.
  42. 42. Liu HC, Lin QL, Ren ML. Fault diagnosis and cause analysis using fuzzy evidential reasoning approach and dynamic adaptive fuzzy Petri nets. Computers & Industrial Engineering. 2013;66(4):899–908.
  43. 43. Jiang W, Xie C, Wei B, Zhou D. A modified method for risk evaluation in failure modes and effects analysis of aircraft turbine rotor blades. Advances in Mechanical Engineering. 2016;8(4):1–16.
  44. 44. Su X, Mahadevan S, Xu P, Deng Y. Dependence assessment in Human Reliability Analysis using evidence theory and AHP. Risk Analysis. 2015;35:1296–1316. pmid:25847228
  45. 45. Wang YM, Yang JB, Xu DL. Environmental impact assessment using the evidential reasoning approach. European Journal of Operational Research. 2006;174(3):1885–1913.
  46. 46. Deng Y. Generalized evidence theory. Applied Intelligence. 2015;43(3):530–543.
  47. 47. Su X, Mahadevan S, Han W, Deng Y. Combining dependent bodies of evidence. Applied Intelligence. 2015;44(3):634–644.
  48. 48. Jiang W, Zhan J, Zhou D, Li X. A Method to Determine Generalized Basic Probability Assignment in the Open World. Mathematical Problems in Engineering. 2016;2016(2016):3878634.
  49. 49. Graham JH. A fuzzy logic approach for safety and collision avoidance in robotic systems. International Journal of Human Factors in Manufacturing. 1995;5(4):447–457.
  50. 50. Gutiérrez-Estrada J, Sanz EDP, López-Luque R, Pulido-Calvo I. SEDPA, an expert system for disease diagnosis in eel rearing systems. Aquacultural engineering. 2005;33(2):110–125.
  51. 51. Binaghi E, Madella P. Fuzzy Dempster-Shafer reasoning for rule-based classifiers. International Journal of Intelligent Systems. 1999;14(6):559–583.
  52. 52. Suh DY, Eisner RL, Mersereau RM, Pettigrew RI. Knowledge-based system for boundary detection of four-dimensional cardiac magnetic resonance image sequences. Medical Imaging, IEEE Transactions on. 1993;12(1):65–72.
  53. 53. Sun X, Tan J, Wen Y, Feng C. Rolling bearing fault diagnosis method based on data-driven random fuzzy evidence acquisition and Dempster-Shafer evidence theory. Advances in Mechanical Engineering. 2016;8(1):1687814015624834.
  54. 54. Velagic J, Lacevic B, Perunicic B. A 3-level autonomous mobile robot navigation system designed by using reasoning/search approaches. Robotics and Autonomous Systems. 2006;54(12):989–1004.
  55. 55. Deng Y. A Threat Assessment Model under Uncertain Environment. Mathematical Problems in Engineering. 2015;2015(2015):878024.
  56. 56. Jiang W, Xie C, Luo Y, Tang Y. Ranking Z-numbers with an improved ranking method for generalized fuzzy numbers. Journal of Intelligent and Fuzzy Systems. 2016;p. Accepted.
  57. 57. Filev DP, Yager RR. A generalized defuzzification method via BAD distributions. International Journal of Intelligent Systems. 1991;6(7):687–697.
  58. 58. Yager RR, Filev D. On the issue of defuzzification and selection based on a fuzzy set. Fuzzy sets and Systems. 1993;55(3):255–271.
  59. 59. Smets P, Kennes R. The transferable belief model. Artificial Intelligence. 1994;66(2):191–234.
  60. 60. Johnson C, Kress R, Roemer R, Hynynen K. Multi-point feedback control system for scanned, focused ultrasound hyperthermia. Physics in medicine and biology. 1990;35(6):781–786. pmid:2367547
  61. 61. Qi Q, Huang W, He Q, Zhao Y, Huang Q. The nine-point controller. In: 2008 3rd IEEE Conference on Industrial Electronics and Applications. IEEE; 2008. p. 645–648.