VARIABLE GAIN ADAPTIVE CONTROL SYSTEMS FOR MACHINE TOOLS by A. Galip Ulsoy, Assistant Professor and Yoram Koren, Goebel Professor Technical Report No. UM-MEAM-82-7 A Progress Report for the Period 9/81 to 8/82 For Grant MEAM-8112629 Submitted to the Production Research Division of the National Science Foundation on September 30, 1982

VARIABLE GAIN ADAPTIVE CONTROL SYSTEMS FOR MACHINE TOOLS by A. Galip Ulsoy and Yoram Koren Department of Mechanical Engineering and Applied Mechanics University of Michigan Ann Arbor, Michigan PROGRAM OBJECTIVES: Computer Numerically Controlled (CNC) machine tools are gaining widespread acceptance in industry. With CNC machine tools the specification of the operating parameters (i.e., feeds and speeds) is left to the experience and judgement of the part programmer, who tends to select conservative values, and consequently reduce production rates. Adaptive control (AC) systems are aimed at providing an additional level of control to automatically adjust operating parameters on-line so as to maximize metal removal rates (MRR). AC systems, however, are not widely used in practise due to the following factors: (i) A lack of detailed understanding of the machining process. (ii) Reliability problems with sensors and other hardware associated with AC systems. (iii) Machining time may comprise as little as 5% of the total production time, thus, increases in MRR do not have a significant impact on production rates without significant improvement in tool changing, parts handling, etc. (iv) Performance and stability problems associated with AC systems due to the variable nature of the machining process. The objective of our program is to address the last problem through the development of high performance, stable AC systems. Specifically our goals are: (i) To design a variable-gain or parameter adaptive AC system which will adapt the AC controller to the changing process parameters and ensure stability and good performance over the full range of machining operations. (ii) To develop a laboratory system to evaluate the variable-gain AC controller through machining tests on an NC lathe and CNC milling machine. PROGRAM ACHIEVEMENTS (9/81 to 8/82): During the first year of our program, we have made progress in both the theoretical and experimental studies required to achieve the above stated goals.

(a) The Structure of the AC Loop (b) Hardware Configuration Figure 1. Conventional Adaptive Control System for Turning.

Theoretical Studies: We began our theoretical investigation by reviewing the literature pertinent to AC systems for machine tools, and through digital simulation studies of a candidate Variable-gain AC system. These studies have been presented in the two papers listed in the "Documents" section and are summarized below. Typically AC systems for machine tools are classified into two types [1-3]: 1) Those using adaptive control for optimization (ACO) extremizea performance index (usually an economic function) subject to process and system constraints. 2) Those using adaptive control with constraints(ACC) maximize machining parameters (e.g., cutting -peed or feedrate) subject to process and system constraints (e.g., allowable cutting force). ACC Systems do not use a performance index. Due to difficulties in formulating realistic performance indices and in measuring required variables in a process environment, ACO applications have been limited mainly to grinding. Most of the systems used in practice for milling, turning, and drilling are of the ACC type. A conventional ACC system, as shown in Fig. 1 for a CNC lathe, is a feedback control system where the feed (f) is manipulated to maintain a required value (FR) of the cutting force (F ). The process block in Fig. l(a) contains the control loops of the CNC controller, the cutting process, and the force transducer, as illustrated in Fig. l(b). The cutting process itself is part of the control loop, and variations in the parameters of the cutting process affect the performance of the AC system. Note that while this type of system is termed "adaptive" in the manufacturing literature, it is not an adaptive system in the sense defined in the control literature [4-7]. An adaptive system in this latter sense, in addition to adapting the feed to the cutting force, must also adapt the AC controller to the changing parameters of the cutting process. Here we refer to such systems as "parameter adaptive systems." Many researchers have recognized the need for parameter adaptive control systems, in order to achieve stability and good performance over a wide range of operating conditions [23-29]. Mathias [29] described a commercial AC system with " automatic gain control", where "the controller gain is automatically reduced at the onset of feedrate oscillations." Gieseke [28] reported an AC system with a PI controller where the P and I action gains are functions of the spindle speed. Weck [26] has described an AC system which uses digital logic to switch the controller gains based on the operating conditions. 2

FR + -t AC CONTROLLER / PROCESS Fc I _ 1' CONTROLLER ADAPTATION PROCESS ESTIMATION I I Figure 2 Structure of a Variable-Gain AC System with Explicit Parameter Estimation Figure 3 Structure of an Adaptive Model Following Control System

Stute [25] has described two alternative schemes for controller gain adjustment. One uses a cutting process model to estimate the process gain and adjust the controller gain accordingly (see Fig. 2).t The second scheme uses a digital PID algorithm whose gains are a function of the manipulated variable (feedrate). These parameter adaptive systems, however, all represent preliminary attempts at a practical solution and not a theoretically based design. The most advanced work to date on parameter adaptive control systems for machine tools has been reported by Masory and Koren [23,24]. They have developed a variable-gain AC system for turning based on cutting force measurement and manipulation of the feedrate. Fig. 2 shows the structure of their variable-gain system. By comparing this structure to that shown in Fig. 1 for a conventional AC system, we note that a parameter estimation block [30,31] and a controller adaptation block have been added. The parameter estimation block provides estimates of the cutting process parameters which vary with depth-of-cut and spindle speed. The controller adaptation block uses the estimated parameter values to adjust the controller gain such that a desired constant value of the open-loop gain is maintained [23,24]. Masory and Koren have theoretically and experimentallly (using a 70HP CNC/AC lathe) verified the feasibility of a veriable-gain AC system for turning. Our current project is aimed at extending the work of Masory and Koren by conducting further studies to: 1) Determine the best strategies and structure for parameter adaptive control. 2) Develop practical methods for the selection of adaptation parameters and sampling periods. 3) Evaluate selected designs (from #1 above) through actual machining tests. To date we have concentrated on the first task which requires the investigation of controller algorithms, parameter estimation algorithms, and controller adaption algorithms for a variable-gain system such as shown in Fig. 2 Other structures for parameter adaptive control, such as the Adaptive Model Following Control (AMFC) system shown in Fig. 3, will be investigated in the near future [6,7,32]. Digital simulation studies for the structure shown in Fig. 2 have shown excellent qualitative agreement with the experimental results in [23,24]. Thus, these studies can be expected to provide useful information for evaluation and design. The simulation is based on the following equations for the turning process [23,24], 3

4. I I I 3. 3 4) 4 2. 0 1) 04 a) 1. n I I I 0.0 2.5 5.0 7,5 10.0 Time, t (sec) Figure 4 Depth-of-Cut versus Time for Simulation Results in Figs. 6-12

TF + F = K f (1) c c p where Fcis the product of the actual cutting force and the sensor gain and the A/D converter gain; f is the feed; K is the process gain and depends on the depth-of-cut, the spindle speed, properties of the tool and workpiece, and the feed itself; and T is the process time constant. The feed is related to the digital command signal (u) from the computer, *- " ~ _2 f + 2wf + w f = K u (2) n n s where 5 is the damping ratio, and wn is the natural frequency of the CNC servo-loop dynamics. Ks is the servo-loop and D/A converter gain. The adaptive controller uses an integral policy, u = Kc (FR-F)dt = K E dt (3) 0 O The process estimation is also based on an integral policy, Km = C (F-uKm)dt (4) where K is the model gain corresponding to K in Eq. (1). where K is the model gain corresponding to K in Eq. (1). Finally, an integral policy is used for the Bontroller gain adaptation, Kc = (K-K K)dt (5) 0 where K is the desired value of the system open-loop gain which is selected based on stability and performance considerations. Although the simplest strategies have been employed, equations (1)-(5) lead to a sixth-order system of nonlinear equations, the analysis of which is not trivial. The effect of sampling is then accounted for to derive the corresponding difference equations on which the simulation results presented below are based. It should be noted that the cutting process model in Eq. (1) is intended for control system analysis and design, and does not attempt to describe the fundamental physical processes of a machining operation. Using simple integral policies for the controller, parameter estimation, and controller gain adaptation necessitates the selection of two parameters: cl for the parameter estimation and c2 for the controller gain adaptation (see Eqs. (4) and (5)). The effects of these parameters on system performance is illustrated in Fig. 5-12. Fig. 4 shows the stepwisejchange in depth-of-cut (a) that was used in all the simulation results presented in Figs. 5-12. The sampling 4

250. 200. 150. 100. 50. 0. 0.0 2.5 5.0 7.5 10.0 Time, t (Sso) Figure 5 Simulated AC System Force versus Time for Conventional (C1 = 0.0, C2 = 0.0) 250. 200. 150. 100. 50. O. 0.0 2.5 5.0 7.5 10.0 Figure 6 Time, t (sioa Simulated Force versus Time for a Variable-Gain AC System with cl = 0.001 and c2 = 0.5

3. 0 c -(4 0 e4 r-I r-4 0 4-) 0 2. 1. 0. 0.0 2.5 5.0 7.5 10.0 Figure 8 3. E 2. $4 0 + 1. 0/ 0 4 r-I 0, 0. 0.0 Time, t (se:) Simulated Model Gain versus Time for a VariableGain AC System with cl = 0.001 and c2 = 0.5 2.5 5.0 7.5 Time, t (bee) 10.0 Figure 8 Simulated Model Gain versus Time for a VariableGain AC System with cl = 0.001 and c2 = 0.5

250. 200. U 150. 0 "4 U 100. 0 0. 0. 0.0 2.5 5.0 7.5 Time, t (see) 10.0 Figure 9 Simulated Force versus Time for a Variable-Gain AC System With c1 =.015 and c2 = 0.1 1 2

3. I I 1- At ----- 2. 4-) (0 i'I U^~~~~~~~~~~~ I~~%. 01I, 4J,,, 0.0 2.5 5.0 7.5 10.0 Time, t (eec) Figure 10 Simulated Controller Gain versus Time for a Variable-Gain AC System with cl =.015 and C2 = 0.1

4.0 3,5 3.0 2.5 2.0 04 1.5 0 1.0 0.5 0.0 -0,5 I 0.0 2.5 5.0 7.5 10.0 Time, t (sec) Figure 11 Simulated Model Gain versus Time for a Variable-Gain AC System with C1 =.015 and c2 = 0.1

800. 700. -- o r-I O H r-4.r 0 a) U o 0 a) 4-) 3 r-4 0 U) 600. 500. 400. 300. 0.001 0.01 0.1 Controller Adaptation Parameter, c2 1.0 Figure 12 Absolute Error versus Controller Gain Adaptation Parameter c for Several Values of the Parameter Estimation Parameter cl

period used was At = 0.1 second in all cases. Fig. 5 shows the cutting force response for the conventional AC System in Fig. 1 (i.e., cl = c2 = 0). The system is seen to be unstable at large depths-of-cut. This instability is remedied by using the variable-gain approach. Fig. 6 shows the cutting force response with cl = 0.001 and c2 = 0.5. The system is now stable, and Figs. 7 and 8 show how the controller gain (Kc) and estimated process gain (Km) are varied to achieve this improved performance. Another simulation with a larger c1 and smaller c~ is also presented in Figs. 9-11. Fig. 9 shows the cutting orce response with cl = 0.015 and C2 = 0.1. The cutting force response is again stable and, as shown in Fig. 10, the controller gain (Kc) is adapted to the changing depth-of-cut in the process. Fig. 11, however, shows that the estimation of the model gain (Km) is beginning to exhibit instability at low depths-of-cut. The effects of the parameters cl and c2 on system performance is illustrated in Fig. 12, where the absolute force error criterion is plotted versus c2 for several values of cl. Small values of cl and c2 lead to poor performance and instability. This is to be expected since the variablegain nature of the system is lost and the system behavior approaches that shown in Fig. 5. Poor performance and instability also results with large cl and c2 values. There is, however, a region of cl and c2 values for which good performance is obtained. The designer of the AC system must select the c1 and c2 values within this region in order to ensure stability and satisfactory performance of the entire AC/CNC system. It is clear that practical methods for the selection of adaptation parameters and sampling period must be established if parameter adaptive AC/CNC systems are to find widespread/industrial acceptance. Experimental Studies: During the past year, we have assembled the equipment necessary to undertake the proposed experimental studies. A PDP-11/23 laboratory computer system, complete with analog-to-digital converters, digital-to-analog converters, programmable clock, and digital (parallel) input-output ports, has been purchased. The computer system is equipped with the usual peripherals (CRT, disk drives, and printer), and has the following software systems available: (i) RT-11 operating system with MACRO assembler and FORTRAN, and (ii) UCSD P-system operating system with UCSD Pascal and FORTRAN'77. This computer system will be used to implement the variablegain AC controller on both a LeBlond NC Lathe and a Bridgeport CNC milling machine, which are available in our departmental laboratories. The interfacing of the computer system to the machine tools requires cutting force and/or power sensors which provide the feedback to the computer from the cutting process. The 5

a) A PDP D C 11/23 D Computer D ~~- A C Digital I/O PDP D 11/23 C Computer D A C Force Dvnamometer CNC Milling Machine Feed Override. =. i1 Force Dynamometer 1 v NC F Lathe E W E i~~eeQ ~ b) Digital I/O Oeea Override I 1 J [7 1 — ~~~~~~~~~~~~ I - - -- _ 1 I Figure 13 Schematic of Laboratory System for a) CNC Milling Machine, and b) NC Lathe

computer manipulates the feedrate on the machine tools through the feedrate override circuits. The experimental systems are illustrated schematically in Fig. 13. Note that both cutting force and power signals are available as feedback signals on the NC Lathe system, and their relative usefulness for AC will also be investigated. Due to the feedrate override electronics on the NC lathe, the feedrate cannot be continously adjusted. It can be adjusted in discrete steps of 15% of the full programmed feedrate from 0% to 150%. Also note that a power monitor is not used on the CNC mill, since the mill feed drives are stepping motors. The feedrate on the CNC mill can be continously adjusted in the range 0% to 150% of the full programmed feedrate. While the required computer-machine tool interfaces, as described above, are essentially complete, they have not been tested. In the coming months we will be testing the system, developing the required software modules, and designing the required machining tests. The machining tests will be designed to evaluate, (i)'The comparative performance of conventional versus variable-gain AC systems. (ii) The performance of variable-gain AC systems for widely varying tool-workpiece properties. (iii) The relative advantages of cutting force sensors versus power monitors as the feedback element in a variable-gain AC system for turning. We expect to begin our machining tests on the CNC mill during the summer of 1983, and to conduct machining tests on the NC lathe during the Fall of 1983. REFERENCES 1. Koren, Y., Computer Control of Manufacturing Systems, McGraw-Hill, New York, 1983. 2. Pressman, R.S., and Williams, J.E., Numerical Control and Computer-Aided Manufacturing, John Wiley and Sons, New York, 1977. 3. Wick, C., "Automatic Adaptive Control of Machine Tools," Manufacturing Engineering, Sept. 1977, pp. 38-45. 4. Groover, M.P., "A Definition and Survey of Adaptive Control Machining," SME Paper No. MS70-561, 1970. 5. Mishkin, E., et al., Adaptive Control Systems, McGraw-Hill, New York, 1961. 6. Landau, I.D., "A Survey of Model Reference Adaptive Techniques - Theory and Applications," Automatica, Vol. 10, 1974, pp. 353-379. 7. Landau, Y.D., Adaptive Control, Marcel Dekker, New York, 1979. 6

8. Anonymous, "In Process Tool Wear Sensors," Survey on Sensors and Transducers, CIRP Technical Committee Cutting Working Group Equipment, Circulated During 1977. 9. Kegg, R.L., "Production Experience with Adaptive Controls," 3rd NC Robot Automation Conference, 1978. 10. Porter, B., and Summers, R.D.M.J., "Adaptive Machine-Tool Control-The State of the Art," Machinery, Feb. 5, 1969. 11. Peklenik, J., "Analysis of the Adaptive Control of Manufacturing Systems - A Critical Assessment," CIRP Fourth International Seminar on Optimization of Manufacturing Systems, 1972. Uncorrected Preprint. 12. Carpenter, D., "Adaptive Control," Proceedings of the Machine Tool Task Force Conference, Vol. 4, Oct. 1980, pp. 31-40. 13. Mathias, R.A., Boock, W., and Welch, A., "Adaptive Control: Monitoring and Control of Metal-Cutting Processes," Proceedings of the Machine Tool Task Force Conference, Vol. 4, Sect. 7.13, Oct. 1980. 14. Amitay, G., Malkin S., and Koren,Y., "Adaptive Control Optimization of Grinding," Trans. of ASME, JOurnal of Engineering for Industry, Vol. 103, No. 1, Feb. 1981, pp. 102-111. 15. Mathias, R.A., "Adaptive Control for the Eighties," Personal Communication. 16. Anonymous, "Adaptive Control Goes Into Production," Steel, Apr. 29, 1968. 17. Mathias, R.A., "Adaptive Control of the Milling Process," Proceedings of the IEEE Machine Tools Industry Conference, Paper No. 34CP67-716, Oct. 1967. 18. Cole, L.M., Numerical Control Programming Languages, NorthHolland Publishing Company, Chapter 4.5, 1970. 19. Donahue, E.J., "Applications of Adaptive Control in the Aerospace Industry," SME Paper No. MS76-274, Apr. 1976. 20. Mathias, R.A., "Adaptive Control for Machining Centers," SME Paper No. MS76-196, 1976. 21. Mayer, J.E., "Estimated Requirements for Machine Tools During the 1980-1990 Period, " Proceedings of the Machine Tool Task Force Conference, Vol. 2, Oct. 1980, pp. 31-41. 22. Birla, S.K., "Sensors for Adaptive Control and Machine Diagnostics," Proceedings fo the Machine Tool Task Force Conference, Vol. 4, Sect. 7.12, Oct. 1980. 23. Masory, O., and Koren, Y., "Adaptive Control System for Turning," Annals of the CIRP, Vol. 29, No. 1, 1980. 24. Koren, Y., and Masory, 0., "Adaptive Control with Process Estimation," Annals of the CIRP, Vol. 30, No. 1, 1981, pp. 373-376. 25. Stute, G., "Adaptive Control," Proceedings of the Machine Tool Task Force Conference, Vol. 4, Sect. 7.14, Oct. 1980. 26. Weck, M., "Adaptive Control in Turning," Proceedings of the Machine Tool Task Force Conference, Vol. 4, Sect. 7.15, Oct. 1980. 7

27. Brecker, J.N., and Shum, L.Y., "Tool Collision and Machine Considerations in Adaptive Control Systems," Annals of the CIRP, Vol. 25, No. 1, 1976, pp. 319-322. 28. Gieseke, E., "Adaptive Control Constraint and Automatic Cut Distribution System for Turning Operations," Personal Communication, 1977, (Based on a Doctoral Dissertation, TH Aachen, 1973). 29. Mathias, R.A., "Software Adaptive Control Optimum Productivity for CAM," SME Paper No. MS77-252, 1977. 30. Eykhoff, P., System Identification, Parameter and State Estimation, John Wiley and Sons, London, 1974. 31. Astrom, K.J., and Eykhoff, P., "System Identification - A Survey," Automatica, Vol. 7, No. 2, March 1971, pp. 123-162. 32. Unbehauen, H., (Ed.), Methods and Applications in Adaptive Control, Springer-Verlag, New York, 1980. 33. Centner, R., "Final Report on Development of Adaptive Control Techniques for a Numerically Controlled Milling Machine," Technical Documentary Report ML-TDR-64-279, Aug. 1964. 34. Colwell, L.V., Frederick, J.R., and Quackenbush, L.J., Research in Support of Numerical and Adaptive Control in Manufacturing, The University of Michigan, Ann Arbor, 1969. 35. Porter, B., and Summers, R.D.M.J., "The Performance of Self-Optimalizing Strategies in the Adaptive Control of the Metal Cutting Process," International Journal of Machine Tool Design and Research, Vol. 8, 1968, pp. 217237. 36. Huber, J., and Centner, R., "Test Results with an Adaptively Controlled Milling Machine," ASTME Paper No. MS68-638, 1968. 37. Weck, M., and Mueller, W., "Chatter Vibration Sensors for Adaptive Control Systems for Turning and Milling Operations," NAMRC, No. 4, 1976. 38. Weck, M., et al., "Adaptive Control for Face-Milling Operations with Strategies for Avoiding Chatter-Vibrations and for Automatic Cut Distribution," Annals of the CIRP, Vol. 24, No. 1, 1975, pp. 405-409. 39. Stute, G., and Kapajiotidis, N., "Integration of Adaptive Control Constraint (ACC) into a CNC," Annals of the CIRP, Vol. 24, No. 1, 1975, pp. 441-415. 40. DeFilippi, A., and Ippolito, R., "Adaptive Control of Milling Machines," SME Paper No. MS73-173, 1973. 41. Bedini, R., and Lisini, C., "Computer Control of Milling Machines," SME Paper No. MS73-173, 1973. 4`42. Nakazawa, K., "Improvement of Adaptive Control of Milling Machine by Non-Contact Cutting Force Detector," Proceedings of the 16th International Machine Tool Design and Research Conference, pp. 109-116. 44. Bedini, R., and Pinotti, P.C., "A Hardwired Logic for the Adaptive Control of a Milling Machine," International Journal of Machine Tool Design and Research, Vol. 16, 1976, pp. 193-207. 8

45. Novak, A., and Colding, B., "Performance of AC Systems in Relation to Measurable Parameters," Annals of the CIRP, Vol. 23, No. 1, 1974, pp. 33,34. 46. DeFilippi, A., and Ippolito, R., "The Influence of Constraints on Cutting Conditions Optimization," Annals of the CIRP, Vol. 24, No. 1. 1975, pp. 417-421. 47. Beadle, B.R., and Bollinger, J.G., "Computer Adaptive Control of a Machine Tool," Annals of the CIRP, Vol. 16, 1971, pp. 61-65. 48. Beadle, B.R., and Bollinger, J.G., "A Constrained Search Adaptive Controller for Metal Cutting," Second NAMRC, 1974, pp. 267-284. 49. Brecker, J.N., and Shum, L.Y., "Reducing Tool Wear with Air Gap Sensing," SME Paper No. TE77-333, 1977. 50. Sata, T., et al., "Newly Developed Adaptive Control Systems of the Turning Process," CIRP Seminars on Manufacturing Systems, 1973. 51. Welch, A., "Advance Machining Using Computer Aided Technology," SME Paper No. MS77-253, 1977. 52. Abraham, R.G., "Westinghouse Progress in Adaptive Control," Westing house Machine Tool Forum, Paper No. TP6, 1973. 53. Mathias, R.A., "An Effective System for Adaptive Control of the Milling Process," ASTME Paper No. MS68-202, 1968. 54. Milner, D.A., "Adaptive Control & Automation," Information Technology, Paper No. JCIT3, North-Holland Publishing Company, 1978, pp. 721-725. 55. Watanabe, T., Iwai, S., and Nawata, Y., "An Adaptive CNC System of a Milling Machine Tool," Proceedings of the IFAC International Symposium, Oct. 1977, pp. 195-205. 56. Watanabe, T., and Iwai, S., "An Application of a MiniComputer to a Computer Numerical Control System of a Machine Tool," Presented at the IEEE Conference, Nov. 1981. 57. Brecker, J.N., and Shum, L.Y., "Tool Collision and Machine Considerations in Adaptive Control Systems," Annals of the CIRP, Vol. 25, No. 1, 1976, pp. 319-322. 58. Meuller, P.A., "Trainable Adaptive Control for Automated Machining." SME Paper No. MS72-132, 1972. 59. Hudson, C.A., "Computers in Manufacturing," Science, Vol. 215, Feb. 12, 1982, pp. 818-825. 60. Tlusty, J., and Elbestawi, M., "Constraints in Adaptive Control with Flexible End Mills," Annals of the CIRP, Vol. 28, No. 1, 1979, pp. 253-255. 61. Kannatey-Asibu, E. Jr., and Dornfeld, D.A., "Ouantitative Relationships for Acoustic Emission from Orthogonal Metal Cutting," ASME Paper No. 80-WA/Prod-26, 1980. 62. Watanabe, T., and Iwai, S., "Real Time Programming of Computer Numerical Control of a Machine Tool in CAM," Presented at the IFAC international Conference, 1981. 9

OBJECTIVES FOR THE NEXT YEAR (9/82 to 8/83) The program objectives for the next year are: 1) To continue analytical and simulation studies of candidate controller designs. Specifically we will investigate an Adaptive Model Following Controller (AMFC) design, stability requirements, methods for selecting adaptation parameters, and methods for selecting the sampling frequency. 2) To develop the required software on the PDP-11/23 system. Specifically we will develop programs to test the interfaces that have been completed, to implement the AC on the milling machine, and to compare different AC strategies. Required modifications to implement the AC on the lathe will be started in August 1983. 3) To conduct a series of machining tests on the CNC milling machine. These tests will involve conventional AC controllers as well as candidate oarameter adaptive AC controllers. Several different tool and material combinations will be tested. These tests will provide the results required for comparison between conventional and parameter adaptive AC strategies. DOCUMENTATION 1. Ulsoy, A.G., Koren, Y., and F. Rasmussen, "Principal Developments in the Adaptive Control of Machine Tools." in Measurement and Control for Batch Manufacturing, ASME Booklet, New York, 1982. 2. Ulsoy, A.G., and Y. Koren, "Variable-Gain Adaptive Control Systems for Machine Tools," NSF Workshop on Manufacturing Systems and Productivity, Dearborn, Michigan, March 1982. COLLABORATORS Materials Technology Laboratories Equipment Group TRW Inc. Cleveland, OH 44117 ACKNOWLEDGEMENTS We would like to acknowledge the assistance of Fred Rasmussen and Dr. Oren Masory with the literature review, and of Jerry Raski with the simulation studies. Mike Cambell assisted with the computer-machine tool interfacing, and Carol K. Bovan helped in preparing the manuscript. Jack Lawrance at TRW provided many helpful suggestions regarding the machining tests. 10