【正文】 k a y kk b y k du k b u k duk c v k dv k c v k dvkr??????????????? ? ? ?? ? ? ? ?? ? ? ? ????????? （ 3） Where bj( k) is the value of the membership function of the jut local model at the current value of the scheduling variable s(k). Normalized triangular membership functions are used, as illustrated in Fig. 2. The sc hed uling variable s(k) is calcul ated using coe fficients kr, ky, ku, and kv, using the w eighted sum ( ) ( ) ( ) ( 1 ) ( )r y u vs k k r k k y k k u k k v k? ? ? ? ? (4) The coefficients are co nfigured b y the e ngi neer according to the na ture of the process nonlinearity. . Online Learning Agent (OLA) The OLA scans the buffe r of recent real time signals, prepared by the SPA, and estimates the parameters of the local models that are excited by the signals. The most recently derived parameters are submitted to the MIA onl y w he n the y pass the verificatio n test a nd are proved to be be tter tha n the existing set. The OLA is invoked upon demand from the OS or autonomousl y, w hen an interval of the process signals w ith sufficient excitation is available for processing. It processes the signals batch w ise and using the local learning ap proach. An advantage of the batchw ise concept is that the decision on w hether to adapt the model is not performed in real time but follow ing a delay that allow s for inspection of the identi fication result before it is applied. Thus, be tter mea ns for co ntr ol over data selection is provided. A problem of distribution of the putation time required for identification appears w ith batchw ise processing of data (opposed to the online recursive processing that is typically used in adaptive controllers).This pr oblem is resolved using a multi tasking operation system. Since the OLA typically requires considerably more putation than the real time control algorithm, it runs in the background as a low priority task. The follow ing procedure, illustrated in Fig. 3, is executed w hen the OLA is invoked. . Copy signal buffer The b uffer of the real time signals is maintained b y the SP A. Whe n the OLA is invoked, the relevant section of the buffer is copied for further processing. . Excitation check A q uick e xci ta tio n c hec k is perfor med a t the s tart, so that processing o f the signals is performed only w hen they contain excitation. If the standard deviations of the signals r(k), y(k), u(k), and v(k) in the active buffer are below their thresholds, the execution is cancelled. . Copy active MFM from MIA The online learning procedure alw ays pares the new ly identified local models w ith the previous set of parameters. Therefore, the active MFM is copied fro m the MIA w here it is stored. A de fa ult set of model par a meters is used for the local models that have not ye t bee n ide ntified (see Section ). . Select local models A local model is selected if the sum of its membership functions bj(k) over the active buffer normalized by the active buffer length exceed s a given threshold. Only the selected local models are included in further processing. . Identification The local model parame ters are identified using the fuzzy i nstr ume ntal variables (FIV) identification method developed by Blazˇ icˇ et al. (2021). It is an extension of the linear instrumental variables identification procedure (Lj ung, 1987) for the specified MF M, based o n the local learning approach (MurraySmith amp。 Ha168。 Seb, 1999) indicate that w hile there is a considerable and grow ing market for advanced controllers, relatively few vendors offer turnkey products. Excellent results of advanced co ntrol concep ts, based on fuzzy para me ter sc hed uling (Ta n, Hang, amp。 Nonline ar control。 Mode lbase d control。 Araki,1998。 Sebor g, 1994。 Section 3 gives a brief description of the CT。 Juricˇ ic180。 Kocijan, Zˇ unicˇ ,Strmcˇ nik, amp。 Jennings, 1995). The main purpose is to simplify controller configuration by partial automation of the missioning procedure, w hich is typically performed by the control engineer. ABS solve difficult problems by assigning tasks to w orked softw are age nts. T he so ftw are age nts are characterized b y properties suc h as autonomy (operation w ithout direct intervention of humans), social ability (interaction w ith other agents), reactivity (perception and response to the environme nt), proactiveness (goal directed behaviour,ta ki ng the initiative), etc. This w ork does not address issues of ABS theory, but rather the application of the basic concepts of ABS to the field of process systems engineering. In this context, a number of limits have to be considered. For example: initiative is restricted, a high degree of reliability and predictability is dema nded, i nsi ght i nto the proble m do mai n is limited to the sensor readings, specific hardw are platforms are used, etc. The ASPECT controller is an efficient and user friendly engineering tool for implementation of parameter scheduling control in the process industry. The missioning of the controller is simplified by automatic experimentation and tuning. A distinguishing feature of the controll er is that the algorithms are adapted for implementatio n on PLC or open controller Industrial hardw are platforms. The controller parameters are automatically tuned from a nonlinear process model. The model is obtained from operating process signals by experimental modelling,using a novel online learning procedure. This procedure is based on model identification using the local learning approach (MurraySmith amp。 Bruijn,2021), multiplemodel control (Dougherty amp。 He nso n amp。外文文獻(xiàn)翻譯 Advanced control algorithms embedded in a programmable Abstract This paper presents an innovative selftuning nonlinear controller ASPEC