XIAO Qianjun, YANG Xiaoyi

1.Department of Automation, Chongqing Industry Polytechnic College, Chongqing 401120, China;2.School of Education Science, Chongqing Normal University, Chongqing 400047, China

ApplicationofHSICBasedFusionControlStrategyinServoTrackingSystem

XIAO Qianjun1*, YANG Xiaoyi2

1.DepartmentofAutomation,ChongqingIndustryPolytechnicCollege,Chongqing401120,China;2.SchoolofEducationScience,ChongqingNormalUniversity,Chongqing400047,China

Aimedatthepuzzleofbeingdifficulttosatisfythehighprecisioncontrolandstrongrobustnessofservo-systemtrackingbyPIDcontroller,thepaperexploredasortofHSICbasedfusioncontrolstrategy.Inthepaper,itanalyzedtheinfluencefactorsforservo-systemtrackingincontrolprecisionandrobustness,discussedthecontrolstrategyofservo-systemwithuncertainty,designedthecontrolmodel,andconstructedthecontrolalgorithm.Ittookatypicaltwo-orderobjectwithdelayasanexample,gavetheanalysisofsystemsimulationexperiment,andthesimulationcurveofsystemdemonstratedthatitwouldbestronginrobustnessandbetterincontroltrackingqualitybyfusioncontrolstrategy.Theresearchresultsshowthatitisfeasibleandeffectivetotheproposedfusioncontrolstrategyforcomplexservo-systemtracking.

complexservo-system,trackingcontrol,human-simulatedintelligentcontroller,fusioncontrolstrategy

1.Introduction

Owing to the suffered from influences resulted in lots of uncertainty factors, it is difficult to build strict mathematical model for complex servo-system tracking, and it always results in being hard to actualize effectively the tracking control of high precision and strong robustness by means of conventional PID controller[1-2]. Based on the analysis of influence factor and control strategy[3-4], the following proposes a sort of HSIC based fusion control strategy in servo tracking system and discusses its application in complex servo-system.

2.Puzzles in tracking control of complex servo-system

The puzzles in complex servo-system rest with approximation degree for actual servo-system (such as Radar tracking system), and it always loses sight of cybernetics essence and only takes the theoretical model into account in the process of design. The main aspects include that it does not fully consider the features in cybernetics characteristic happened in control process.

① Feature of control parameter. In the physical world, the characteristic represents as the unknown, time varying, randomness, fuzziness and distribution and so on sometimes. Under a certain condition, it maybe has the difference as many as hundreds, and even as many as ten thousands, and it is closely related with lots of uncertainty factors such as product structure, and technology condition and so on. The approximation degree of parameter always depends on the experience and skill of designers, and it is biggish in parameter variance distribution.

② Feature of large & pure delay and large inertia. For example, generally the distribution of temperature field is always asymmetrical and non steady in work-piece processing for physical system. Its description of relation should be made by partial differential equation, and there exists obvious lag in heat transfer. Here taking the sampling periodT=100 msec，the lag time of temperature response is expressed byτ,and the influence of time lag has to be taken into account in the algorithm. The test shows that theτ/Trate of system maybe is rather large. From this test, it can be seen that the heating process is a large & pure delay and large inertia system probably.

③ Nonlinear feature. In the conventional PID control, it always supposes the system to be linear, but it is nonlinear in fact. The linear system must be satisfy the condition of both homogeneity and linear superposition additivity, but lots of system do not satisfy the equationf(k1x1+k2x2)=k1f(x1)+k2f(x2), therefore the system is obviously a nonlinear system.

④ Feature in time varying and lag unknown. Lots of uncertainty factors result in unknown and time varying of controlled process. Usually we always suppose the system to be time invariant system and the time lag of system (process) is known beforehand.

⑤ Influence of environment interference. The environment interference has the characteristic of diversity, randomness, and unknown, and it always full of uncertainty and is uncontrollable.

From the features of cybernetics mentioned above, the main puzzles are summarized as follows if the conventional PID controller is adopted. The uncertainty results in that it is difficult to construct the strict math model by analytical method in quantity, and the object model and interference is unknown or acquirable by system identification. For control problems such as “unknown”, “uncertainty” or “little-known”, it is very hard to make the mathematical description, so it also is impossible to make the control effectively by means of conventional PID controller. Conventional control theory applies differential equations, state equations and all kinds of mathematic transformation as the research tools, they are numerical computing approach in nature and belong to the constant control category and strictly subjected to the structures of controlled process, which should be described easily with mathematic models. For the problem of strong non-linearity, semi-structure and non-structure, the most important and imperative problem is how to make the math description in some extent.Problems of system complexity and reliability are difficult to make the coordination in performance, and according to the viewpoint of system engineering, the generalized control process includes operating objects and its surroundings usually. But some factors are strongly coupled and inter-constrained, besides, the surroundings is also very complex and varies continuously. Therefore the conventional control method can’t solve the control problem effectively. Although there are contradictory phenomena between robustness and sensitivity in system, but the contradiction is not obvious in reliability of simple control systems. However for complex tracking process, it is possible to cause the collapse of the whole control system because of condition change. So it is necessary to explore more effective control strategies and control algorithms.

3.Selection of strategy and model for servo-system control

3.1.SelectionofHSICbasedfusioncontrolstrategy

The key factor of influencing control effect is control parameter, but through disposable parameter tuning it is impossible to make the system be in the state of optimizing from beginning to end by means of PID control parameter. Although lots of control strategies supply the feasibility to solve the puzzles[5-6], but there still are lots of puzzles needed to be solved. For example, due to the influence of uncertainty, the neural network control is hardly able to collect the whole experiment samples from the known experience, and therefore it is very difficult to realize the effective control generally because of the method limitation. The knowledge based expert control system is also difficult to extract the dynamic characteristic information to express the feature of system and construct the perfect knowledge repository, and therefore it is also difficult to realize effective tracking control for complex servo-system. The fuzzy control can not depict the feature of servo-system fully, because the factors of uncertainty are too much, therefore the fuzzy control is unnecessarily a good choice for high precision servo-system tracking control. Here the intelligence based fusion control strategy based on HSIC is selected. The HSIC (Human Simulated Intelligent Controller) has itself basic characteristic such as multi-hierarchical organization of information processing and decision, characteristic recognition on line and characteristic memory, multi-modal pattern of combining with open and closed loop control as well as quantitative control and qualitative decision, application of reasoning logic of heuristic and intuition. Because its basic property is to simulate the control behavior of control expert, its control algorithm is the multi-modal control pattern. The algorithm can perfectly coordinate lots of demand of mutual contradiction in the control engineering indexes, for example in robustness and smooth character. Therefore it is closer to the actual engineering, and the control method is to execute the alternate use among multi-modal control algorithms. For high precision servo tracking system, it is maybe a sort of more wise choice to makes their fusion between HSIC and EC, namely it adopts the control strategy based on intelligence fusion.

3.2.Generalizedcontrolmodelforservotrackingsystem

Fig.1 Generalized control model

4.Design of basic control algorithm

The control rule set can be derived from prototype of HSIC, and it can be expressed as the following.

In general speaking, the design of human simulated intelligent controller can be summarized as the following basic steps.

1) Establish the design target trace

2) Establish the characteristic model

3) Design the control rule and control mode set

Aimed at the difference between instantaneous index and certain characteristic state in the characteristic model located by moving state of the system, as well as moving trend of ideal trace, it simulates the decision behavior of human simulated control, designs the control mode. For the design of controller in running layer, the target trace of ideal deviation is shown by broken line in Fig.2. In order to make the trace being consistent with practical deviation and ideal target trace, it can take the following control countermeasure.

If the deviation is getting great, corresponding to area ① in Fig.2, then the Bang-Bang control mode would be adopted, namely control action is large as possible.

If the deviation and its change rate are getting very small, corresponding to area ⑤ in Fig.2, then the PI algorithm would be adopted so as to eliminate the system steady deviation.

If the deviation is getting larger, corresponding to area ② in Fig.2, then proportional control mode would be adopted. And at the same time, it has to introduce a weak differential coefficient so as to endure the deviation change rate not being too large.

In the process of reducing deviation, if the deviation change rate is getting lower or equal to pre-specified rate, corresponding to area ④ in Fig.2, then the control mode of proportional plus differential mode would be adopted.

In the process of reducing deviation, if the deviation change rate is getting great than pre-specified rate, corresponding to area ③ in Fig.2, then the strong differential control based on proportional mode has to be introduced so as to make deviation change rate reducing as soon as possible.

Fig.2 Characteristic model of running control

Therefore the system running can be partitioned as five sorts of control mode according to the deviation and deviation change rate of servo-system.

Mode 1un=sgn(en)·Umax，|en|>e1

Summarized hominine control experience based on prototype algorithm, the set of control rule in complex servo-system with uncertainty is a suit of control rule corresponded to different characteristic status.

5.Experiment of system simulation

In order to simplify the analysis and have the representativeness, it takes a servo tracking object controlled as an example, and assumes the model of servo-system to be a familiar two-order with a lag node. For convenience of comparison of control effect, it respectively adopts the control algorithm of PID plus estimate, PID and HSIC based fusion control to make the experiment simulation, and makes the comparative study for the system response. Assume the model of controlled object is as the following.

In the experiment of system simulation, it respectively takes the delay to be asτ=2 s,τ=10 s andτ=20 s, and the response curve of simulation result is respectively shown as in the Fig.1, Fig.4 and Fig.5. The Fig. 6 shows the response when the external pulse interference is joined, and the pulse width is 0.2 s with amplitude being 0.5. From Fig.3 to Fig.6 it can be seen that the control performance of HSIC based fusion control strategy is to excel the PID controller and the controller of PID plus estimator (optimal PID controller) no matter what the speedability, stability and overshoot of the system response or anti-interference capability of the system. From Fig.7 to Fig.10 it respectively shows the system response curve of corresponding slope signal with different gradient for different control input setting, and it can be seen that the deviation produced by the controller of HSIC based fusion control strategy is smaller from beginning to end than the controller of PID plus estimate, and that is to say that the input tracking performance for servo-system is better than the optimal PID controller. The Fig.10 shows that even if the external interference (impulse interference width=10 s & amplitude=0.5) is joined then the controller of HSIC based control strategy still keeps better tracking performance in control precision and robustness for servo-system.

Fig.3 Unit step response under τ=2 s

Fig.4 Unit step response under τ=10 s

Fig.5 Unit step response under τ=20 s

Fig.6 Interference response under τ=15 s

Fig.7 Response under ramp slope is 0.1

Fig.8 Response under ramp slope is 1

Fig.9 Response under ramp slope is 10

Fig.10 Response under interference

6.Conclusions

From theory demonstration and analysis of simulation result mentioned above, it can obtain the following conclusions. ① the HSIC based fusion control strategy is not sensitive for change of control parameter. When the time constant or lag time of controlled object happens in change it can still track the setting value of controlled process in non-overshoot and own the strong robustness. ② It has obvious tracking superiority for large lag object, and with the increase ofτ/Tvalue the system gets more difficult to control and after theτ/Tvalue reaching a certain grade, other control strategy can not make any steady control, but the HSIC based fusion control strategy still owns better control performance. ③ When the external interference is joined the PID control is very sensitive, but the proposed control strategy is still keeps better control effect, and therefore it shows the strong anti-interference capability.

In summary, it can be seen that the HSIC based fusion control strategy represents better control tracking quality, and the tracking precision and robustness excels obviously the other control algorithm, and specially it should be a sort of preferred control tracking strategy for complex servo-system of being difficult to control.

[1] PENG Li,LIN Ying,YANG Yi.Exploring on Related Technique in the Control of Complicated System[J].Journal of Southwest China Normal University:Natural Science,2004,29(6):1066-1068.

[2] CAI Zixing,ZHOU Xiang,LI Meiyi.A Novel Intelligent Control Method Evolutionary Control[C]// Proceedings of the 3’d World Congress on Intelligent Control and Automation.[S.l.]:IEEE,2000：387-390.

[3] LI Shirong.Fuzzy control Neural control and Intelligent Cybernetics[M].Harbin:Harbin Industry University press,1999.

[4] LI Zushu,TU Yaqing.Human simulated intelligent controller[M].Beijing:National defense industry press,2003.

[5] YANG Biao,ZHANG Zengke.Dynamic Characteristic Parameter Setting Method for Human-simulated Intelligent Controller[J].Information and Control,2004,33(6):670-673.

[6] CHENG Zhong,YAN Wei,LI Zushu,et al.NPID controller based on HSIC[J].Control and Decision,2003,18(6):694-698.

基于HSIC的融合控制策略在伺服跟踪系统中的应用

肖前军1*，杨小义2

1.重庆工业职业技术学院 自动化系，重庆 401120；2.重庆师范大学 教育科学学院，重庆 400047

针对PID控制器用于伺服跟踪系统难以满足高控制精度与强鲁棒性的难题，探讨了一种基于HSIC的融合控制策略。分析了影响伺服系统跟踪控制精度与鲁棒性的因素，讨论了具有不确定性伺服系统的控制策略，设计了控制模型，构造了控制算法。以典型二阶滞后对象为例，给出了系统的仿真实验分析结果，仿真曲线显示该融合控制策略的鲁棒强、控制跟踪品质较好。研究结果表明，对于具有不确定性的复杂伺服跟踪系统，提出的融合控制策略是可行和有效的。

伺服系统；跟踪控制；仿人智能控制；融合控制策略

TP273

2012-10-15

*XIAO Qianjun. E-mail: xiaoqianjun2003@126.com

10.3969/j.issn.1001-3881.2013.06.019

2012-10-25