Low-power VLSI Design of Fuzzy Logic Based Automatic Controller for Total Artificial Heart Bashir I. Morshed Department of Electronics
References 2 [1] H. Cheung-Hwa, "Fuzzy logic automatic control of the Phoenix-7 total artificial heart", Japanese Society for Artificial Organs, accepted for publication on Feb. 2004. [2] H. C. Kim, et al., "Development of a microcontroller based automatic control system for the electrohydraulic total artificial heart", IEEE Trans. Biomedical Engineering, vol. 44, no. 1, pp. 77-89, 1997. [3] M. Sasaki, et al., "Fuzzy multiple-input maximum and minimum circuits in current mode and their analyses using bounded-difference equations", IEEE Trans. Computers, vol. 39, no. 6, pp. 768-774, 1990. [4] M. Sasaki, F. Ueno, "A Fuzzy logic function generator (FLUG) implemented with current mode CMOS circuits", Intl. Symp. Multiple- Valued Logic, pp. 356-362, 1991. [5] M. Sasaki, F. Ueno, "A VLSI implementation of Fuzzy logic controller using current mode CMOS circuits", Intl. Conf. Industrial Fuzzy Control and Intelligent Systems, pp. 215-220, 1993. [6] M. Sasaki, F. Ueno, "A Novel Implementation of Fuzzy logic controller using new meet operation", IEEE Conf. Fuzzy Systems, vol. 3, pp. 1676-1681, 1994.
Contents 3 Introduction Artificial heart Controller algorithms Fuzzy logic Fuzzification and defuzzification Design challenges Preliminary results Time table Conclusion
Objective 4 To design an automatic controller to be used in totallyimplantable artificial heart. Desired properties: Low-power and high speed operation, Real-time monitoring and control, Proper operation within some supply voltage variation, Self-regulating and adaptive. Suitable techniques: Fuzzy logic based control algorithm, CML based CMOS (full-custom ASIC) technology. Design challenges: CML based special circuitry to implement Fuzzy functions, Combining CML blocks to implement Fuzzy logic blocks.
Introduction 5 Artificial hearts are being implanted to the patients with critical heart problems, so that they can survive until a heart transplant is possible. Milestones: 1952: First successful open heart surgery by F. John Lewis. 1967: Christiaan Barnard performs the first whole heart transplant. 1982: Willem DeVries first implanted a permanent artificial heart designed by Robert Jarvik. Ongoing researchers are focusing on durable and adaptive artificial hearts so that they can operate independently and reliably without any human control or monitoring.
Total Artificial Heart (TAH) 6 A TAH is an implantable device entirely replacing the human heart for a certain period of time. Must be capable of doing all functions of heart according to specific needs of human body. Two mechanical pumps (diaphragm type) replace the ventricles and are controlled by internal electronic device. Controller Simplified blocks of TAH [2]
7 Proposed Configuration [2]
Controller for TAH Automatic controller monitors and regulates: heart rate, percent systole, drive pressure, separately for both of the left and the right ventricles. Factors to be considered: Patient's physical activity, requirement of oxygen circulation, blood pressure at preload and afterload, Starling's Response, etc. 8
Some Interesting Phenomena Starling's Response: deals with the variable heart volume and rate needed for specific individual. The well trained athlete's heart rate does not increase as much nor as quickly as that of an average individual during strenuous activity. Pulmonary edema: if the right ventricle is over-driven, then fluid can actually be forced into the lungs. The amount of blood flowing from the right ventricle has to be sufficient enough to supply the required amount of blood to the left ventricle. The pressure at afterload on the right ventricle must be less compared to that of the left side. 9
Controller Algorithm 10 Various controlling algorithms and techniques: PID based, RAM/ROM lookup table based, Fuzzy logic based, etc. The advantages of Fuzzy logic based controller: Simple rule based operation makes FL very fast, Potential for real-time automatic control eliminating continuous manual monitoring under varying hemodynamic conditions, Inherent adaptability property of the logic, Highly stable nature and nonlinear control surface, No ADC/DAC.
Controller Design Design options: Micro-controller based, PLA/PAL based, Full custom ASIC. Advantage of full custom ASIC: Efficient and compact design blocks of Fuzzy rules using current mode logic (CML) [3-6], Very low power and high speed of operation, Efficient implementation of control algorithm. 11
Fuzzy Logic (FL) 12 Lotfi A. Zadeh (1973) is the founder of Fuzzy logic. It is basically a convenient way to map an input space to an output space. The basic idea is soft computing alike human logic. Rather than attempting to model a system mathematically, FL incorporates a simple rule-based approach to a solving control problem IF (X) AND/OR (Y) THEN (Z). There are mainly two different models: Mamdani Sugeno
Discrete vs Continuous Logic Continuous Logic (Fuzzy) Discrete Logic (Digital) 13
Precision vs Significance Digital Robot Fuzzy Robot 14
15 Structure of Fuzzy Logic
16 Membership Functions
17 Fuzzification and Defuzzification
18 Mamdani Model
19 Sugeno Model
Example 1 (Mamdani): Rules If then If then If then 20
21 Example 1 (Mamdani): Surface
Example 2 (Sugeno): Rules If or then If and then If or then 22
23 Example 2 (Sugeno): Surface
24 Conventional vs Fuzzy Logic
Design Challenges 25 Design/verification of the following functional blocks: Minimum/Maximum operation (current mode), FL Function generator (FLUG) with rules block, Membership function circuit, Meet operation (replacing weighted average), Bias circuit, variable gain current mirror, etc. Modifications to adopt membership functions other than Gaussian and Mamdani model. Combining the basic blocks and acceptable operation with supply voltage variation. Compatible input/output to meet requirements of TAH.
26 Preliminary Results
Preliminary Results (cont.) Delay effect 27 Minimum device size
Time Table Research Steps Begin End Literature survey 01 Feb. 28 Feb. Design of Fuzzy logic functions 28 Feb. 13 Mar. Design of Fuzzy blocks for TAH controller 14 Mar. 27 Mar. Simulation and verification of all blocks 28 Mar. 05 April Project presentation - 06 April Preparing final report 06 April 17 April Submission of final report - 18 April 28
Conclusion Project goals: The designs proposed by Sasaki will be tested on CMOS 0.18u technology. A few design blocks of the Fuzzy controller for TAH will be designed, tested and verified. Proper operation for operating voltage variation within certain degree of errors. Limitations for simulation: Modeling TAH and various parameters associated with that. The adaptability property. 29