Miller, W. T. (1987). Sensor-based control of robotic manipulators using a general learning algorithm. IEEE J. of Robotics and Automation, 3:157--165.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....an appropriate control action for any desired process output. Widrow et al. 80] used this method to control linear systems. Other researchers have applied this method to the control of nonlinear thermodynamic systems [70] and for learning the inverse kinematics [5, 45] and inverse dynamics [43, 54] of robots. Direct inverse modeling appears to be an attractive method for learning control because it is easily implemented and, in most cases, results in good controllers. However, its applicability and usefulness depend on the characteristics of the controlled process. With dynamic processes, ....

W. T. Miller. Sensor based control of robotic manipulators using a general learning algorithm. IEEE Journal of Robotics and Automation, 3:157--165, 1987.

....actual result, inverse models prove more problematic. If the correct motor command was known, which could provide an appropriate supervised error signal, then there would be no need for the inverse model. Three main approaches have been used to adapt such inverse models direct inverse modeling (Miller, 1987; Kuperstein, 1988) distal supervised learning (Jordan and Rumelhart, 1992) and feedback error learning (Kawato, 1990) The latter two models both rely on the ability to convert errors in the actual trajectory into errors in the motor command. They, unlike the direct approach, are able to acquire ....

Miller W.T. (1987). Sensor-based control of robotic manipulators using a general learning algorithm. IEEE J. Robotics and Automation, 3,157-- 165.

....an appropriate control action for any desired process output. Widrow et al. 149] used this method 23 to control linear systems. Other researchers have applied this method to the control of nonlinear thermodynamic systems [131] and for learning the inverse kinematics [10, 78] and inverse dynamics [71, 93] of robots. a) Training the inverse model (b) Using the inverse model as a controller Controller Inverse model Process Desired output Action Output Process Random action Output Target action Inverse model Action Figure 2.4. Control acquisition through direct inverse modeling. ....

Miller, W. T. Sensor based control of robotic manipulators using a general learning algorithm. IEEE Journal of Robotics and Automation, 3:157--165, 1987.

....or muscle commands) needed to achieve the desired state. As the dynamics of the arm change due to growth, damage, fatigue and changes in external loading, the inverse model must be adaptable. Three main approaches for adaptation have been proposed in such inverse models direct inverse modeling (Miller 1987; Kuperstein 1988) distal supervised learning (Jordan and Rumelhart 1992) and feedback error learning (Kawato 1990) The latter two models both rely on the ability to convert errors in the actual trajectory into changes in the motor command and are able to acquire an accurate inverse model even ....

Miller, W.T. (1987). Sensor-based control of robotic manipulators using a general learning algorithm. IEEE Journal of Robotics and Automation, 3:157--165.

....and Casta no et al. 2] These methods track feature points and effect servoing movements using an Image Jacobian which relates Cartesian movements with positional errors derived from the tracked features. Other methods include the work of Papanikopolous et al. 9] Koivo et al. 6] and Miller [12]. The basic idea behind the Image Jacobian is to model the differential relationship between the camera system and the robotic control system in order to accurately predict the effects of small changes in one system on the other. It is a linear, position dependent (i.e. non constant) transform. ....

I. W. Thomas Miller. Sensor-based control of robotic manipulators using a general learning algorithm. IEEE Journal of Robotics and Automation, RA3 (2):157--165, April 1987.

....system are correct over the entire input space. On the other hand, it is unlikely that all possible inputs will be encountered in solving a particular control or classification problem. The standard CMAC model has been applied to the real time control of robots with encouraging success (Miller [27]; Miller, Glanz, and Kraft [28] A comprehensive coverage of the CMAC models and learning algorithms is given by Dean [7] Research on the CMAC model and its variants is still in its early stage. In particular, there are very few rigorous theoretical results available. Many problems remained ....

W. T. Miller, "Sensor-Based Control of Robotic Manipulators Using a General Learning Algorithms," IEEE Journal of Robotics and Automation 3 (April 1987), 157--165.

....that are tangent to constraint surfaces in the configuration space [29] One possible solution to this limitation is to use vision based techniques to control motion in the remaining directions. Thus, much research attention has recently been focused on vision based control (see, for example, [2, 3, 9, 10, 15, 16, 22, 27, 30, 31, 32, 33, 34, 35, 37]) Although vision based control has been used successfully for a number of tasks (for example, in welding applications [1, 5, 21] none of the systems referenced above lend themselves to task level specification of goals, and therefore, there are currently no automatic planning systems that can ....

W. T. Miller. Sensor-based control of robotic manipulators using a general learning algorithm. IEEE Journal of Robotics and Automation, RA-3(2):157--165, April 1987.

....the command series while meeting the demand of changes of the plant s dynamics. Biological evidence strongly suggests that such a task can be solved with the help of learning. Effort along this route include various inverse system identification methods, such as the direct identification method (Miller, 1987; Kawato et al. 1987; Widrow et al. 1978) the indirect method, that is based on the identification of the forward model (Jordan, 1990; Werbos, 1988; Widrow, 1986) and the feedback error learning method, when the errors generated by a previously fixed stabilizing feedback controller are used to ....

W. T. Miller 1987, "Sensor based control of robotic manipulators using a general learning algorithm," IEEE J. Robotics and Automation 3, 157--165.

....These methods differ from each other primarily in the type of learning algorithm used. The learning strategies vary from neural like networks [Kuperstien 87] Mel 87] Miller 89] Walter et al. 90] to table lookup mechanisms such as the cerebellar model articulation controller (CMAC) Albus 75] Miller 87] Here, we propose a new framework for learning based visual positioning tracking. Our approach differs from previous ones in two significant ways; a) the method uses raw brightness images directly without the computation of image features, and (b) the learning algorithm introduced is based on ....

W. T. Miller, Sensor-based control of robotic manipulators using a general learning algorithm, IEEE Journal of Robotics and Automation, Vol. RA-3, No. 2, pp. 157-165, April, 1987.

....there have undoubtedly been some oversights (our apologies ) These are organized by application with a listing of the numbers from the papers in the Bibliography at the end of the chapter. Headings for theory papers and hardware papers are also included. Control (including robotics) Kinematics [10, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40] Dynamics [30, 33, 34, 36, 41, 42, 43, 44, 45] Unstable plant [34] Time delay in plant [46] Adaptive critic [47, 48, 49, 50] Walking (biped and quadruped) 35, 36, 38, 43, 44, 51] Dynamic programming [52] Chemical systems [9, 53, 54, 55, 56, 57] Optimal [52, 58, 59] Mobile [60] Manipulator robot ....

.... Dynamics [30, 33, 34, 36, 41, 42, 43, 44, 45] Unstable plant [34] Time delay in plant [46] Adaptive critic [47, 48, 49, 50] Walking (biped and quadruped) 35, 36, 38, 43, 44, 51] Dynamic programming [52] Chemical systems [9, 53, 54, 55, 56, 57] Optimal [52, 58, 59] Mobile [60] Manipulator robot [5, 10, 30, 31, 32, 33, 45, 61, 62, 63, 64, 65, 66] Fuzzy [27, 28, 67, 68, 69] Manufacturing CIM tool fault [39, 70, 71, 72] Pattern recognition Nearest neighbor methods [73] Character recognition [23, 74] Handwriting recognition [74] Signal processing [23, 75, 76, 77] Biomedical [2, 68, 77, 78, 79, 80, 81] Others Physics detectors [82] ....

W. T. Miller, "Sensor Based Control of Robotic Manipulators Using A General Learning Algorithm," IEEE Trans. Robotics Automat., Vol. RA-3, pp. 157-165, 1987.

....These methods differ from each other primarily in the type of learning algorithm used. The learning strategies vary from neural like networks [Kuperstien 87] Mel 87] Miller 89] Walter et al. 90] to table lookup mechanisms such as the cerebellar model articulation controller (CMAC) Albus 75] Miller 87] Our appearance based approach to robot vision offers a solution to servoing that differs from previous work in two significant ways; a) the method uses raw brightness images directly without the computation of image features, and (b) the learning algorithm is based on principal component ....

W. T. Miller, "Sensor-based control of robotic manipulators using a general learning algorithm," IEEE Journal of Robotics and Automation, Vol. RA-3, No. 2, pp. 157-165, April, 1987.

....on line learning, Liapunov s second Preprint submitted to Elsevier Science 24 March method. 1 Introduction A vast amount of work has dealt with neural networks for controlling a plant with known, partially known, or unknown dynamics. Techniques, such as inverse system identification (see e.g. (Miller, 1987; Kawato et al. 1987; Widrow et al. 1978) or identification based ( indirect ) methods (see e.g. Jordan, 1990; Werbos, 1988; Widrow, 1986) have been proposed for learning the inverse dynamics. For an overview see (Dean and Wellman, 1991; Miller et al. 1990; Narendra and Parthasarathy, ....

....discussion direct inverse modeling provided a better fit with SDS Control. In the next section we describe some simulation results with our neurocontroller that make use of direct inverse modeling and non variational, Hebbian type learning in a CMAC like architecture (Marr, 1969; Albus, 1971; Miller, 1987; Fomin et al. 1994; Szepesv ari and Lorincz, 1995) 5 Computer simulations In this section results of computer experiments are presented. The aim of this section is to illustrate the theory and the working of the compensation mechanism by simulations. The plant s equation, the neurocontroller ....

Miller, W. (1987). Sensor based control of robotic manipulators using a general learning algorithm. IEEE Journal of Robotics and Automation, 3:157--165.

....property, so that input vectors which are similar to previously learned input vectors, but are novel to the system, will generate output vectors that are similar to previously learned output vectors. One solution to the interpolation problem in robot learning control was presented in (Miller, 1987) with the use of the so called cerebellar model arithmetic computer (CMAC) In this algorithm an input vector is mapped to several locations in an intermediate memory, and the output vector is computed by summing over the values stored in all the locations to which the input vector was mapped. ....

....(c.f. Slotine and Li, 1987; Slotine and Li, 1991; Sadegh and Horowitz, 1990; Ortega and Spong, 1989; Wen and Bayard, 1988) and their bibliographies) and function estimation algorithms which use basis expansion functions, content addressable memories, neural networks and self organizing maps (c.f. (Miller, 1987; Messner et al. 1991a; Atkeson, 1990; Barto et al. 1983; Ritter et al. 1992; Shoureshi, 1993) and their bibliographies) In this section we will describe a learning control system for robot manipulators based on integral transforms (Messner et al. 1991a) We assume that there exist L ....

Miller, T. W. (1987). Sensor-based control of robotic manipulators using a general learning algorithm. IEEE Journal of Robotics and Automation, RA-3(2):157--165.

No context found.

Miller, W. T. (1987). Sensor-based control of robotic manipulators using a general learning algorithm. IEEE J. of Robotics and Automation, 3:157--165.

No context found.

W.T. Miller. Sensor based control of robotic manipulators using a general learning algorithm. IEEE Journal of Robotics and Automation, 3:157--165, 1987.

No context found.

Miller III, W.T. (1987), Sensor-Based Control of Robotic Manipulators Using a General Learning Algorithm, IEEE Journal of Robotics and Automation, Vol. RA-4, No. 2, pp. 157-165.

No context found.

Miller III, W.T. (1987), Sensor-Based Control of Robotic Manipulators Using a General Learning Algorithm, IEEE Journal of Robotics and Automation, Vol. RA-4, No. 2, pp. 157-165.

No context found.

W.T. Miller. Sensor-based control of robotic manipulators using a general learning algorithm. IEEE Trans. Robotics and Automation, 3(2), pp. 157--165, April 1987.

No context found.

Miller, W. T. 1987. Sensor Based Control of Robotic Manipulators Using a Generalized Learning Algorithm, IEEE J. Rob. and Automat., 3, 157-165.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC