頁籤選單縮合
題 名 | Multi-Input Multi-Output Self-Learning Fuzzy Logic Control for Haemodynamic Support of Septic Shock |
---|---|
作 者 | Mason,D. G.; Ross,J. J.; McGuinness,S. M.; Linkens,D. A.; Edwards,N. D.; | 書刊名 | 醫學工程 |
卷 期 | 10:4 1998.08[民87.08] |
頁 次 | 頁71-80 |
專 輯 | Special Issue on Control Methods in Anaesthesia |
分類號 | 410.1644 |
關鍵詞 | Fuzzy logic; Intelligent control; Drug infusions; Septic shock; |
語 文 | 英文(English) |
英文摘要 | Haemodynamic support of intensive care patient in septic shock via multiple intravenous infusions represents a challenging multi-input multi-output (MIMO) control problem. Patient characteristics can vary markedly in terms of their multiple circulatory responses to drug infusions. This uncertainty combined with the non-linear and time-varying nature of patient responses to infusions led us to investigate the clinical use of selflearning fuzzy logic control(SLFLC). We have previously demonstrated this direct adaptive control strategy to perform well clinically as a single-input single-output (SISO) control process. Its clinical application to a MIMO control problem such as in haemodynamic support of septic shock gives the added dimension of uncertain coupling between input and output variables. In addition, sometines ineffective drug infusiona need to be replaced by alternative drugs. Different drugs can potentially make significant changes to the coupling between input and putput variables. The MIMO implementation of SLFLC also represents a challenge in keeping the computational and menory costs at a reasonable level as the number of control system inputs and outputs increase. In addition, the architecture should maintain a good level of interpretability by the domain expert so that their insight can be utlised in the development of the control strategy, With these issues in mind a three-input three-ouput control strategy based upon self-learning fuzzy logic was devised to work in computer simulation. utilising a compartmental cardiovascular model to represent the variable patient state in septic shock Cardiovascular variables sampled by the control system include pulmonary artery diastolic pressure (PADP). mean arterial pressure (MAP). cardiac index (CI), heart rate (HR) and urine flow rate ("UFR). Three simulation scenarios were devised to represednt different types of patients in septic shock, including variable drug sesitivities, leading to the need to change to alternative drugs and hence challenging the adaptive capabilities of the control system. Preliminary computer simulations show good performance, commencing with blank rulebases. A simulation system is currently under development which will allow the proposed controller to be assessed off-line by domain experts with the aim of refining the system prior to preliminary chinical triads. |
本系統中英文摘要資訊取自各篇刊載內容。