Pub Date : 1995-08-27DOI: 10.1109/ISIC.1995.525041
B. Stilman
Considers an example of application of linguistic geometry to the search multiagent problem with an extremely high branching factor, this problem is computationally hard for conventional approaches. The high computational complexity is explained by the allowance of concurrent motions of all the agents. This results in a significant growth of the branching factor (up to 300) because all the combinations of simultaneous motions are legal. Another difficulty is that when making a move, each side is uncertain about the concurrent motions of the other side. The linguistic geometry tools generated a solution of the problem demonstrating a dramatic search reduction for this totally concurrent system.
{"title":"Total concurrency and uncertainty in linguistic geometry","authors":"B. Stilman","doi":"10.1109/ISIC.1995.525041","DOIUrl":"https://doi.org/10.1109/ISIC.1995.525041","url":null,"abstract":"Considers an example of application of linguistic geometry to the search multiagent problem with an extremely high branching factor, this problem is computationally hard for conventional approaches. The high computational complexity is explained by the allowance of concurrent motions of all the agents. This results in a significant growth of the branching factor (up to 300) because all the combinations of simultaneous motions are legal. Another difficulty is that when making a move, each side is uncertain about the concurrent motions of the other side. The linguistic geometry tools generated a solution of the problem demonstrating a dramatic search reduction for this totally concurrent system.","PeriodicalId":219623,"journal":{"name":"Proceedings of Tenth International Symposium on Intelligent Control","volume":"79 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134082882","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1995-08-27DOI: 10.1109/ISIC.1995.525059
J. Albus, A. Lacaze, A. Meystel
Autonomous learning in the architectures of intelligent control requires special procedures performed upon acquired knowledge. This affects the structure of world representation and it is intimately linked with mechanisms of behavior generation. This paper illuminates algorithms of autonomous learning performed via nested clustering which is goal driven and exercises simulation of decision making process.
{"title":"Algorithm of nested clustering for unsupervised learning","authors":"J. Albus, A. Lacaze, A. Meystel","doi":"10.1109/ISIC.1995.525059","DOIUrl":"https://doi.org/10.1109/ISIC.1995.525059","url":null,"abstract":"Autonomous learning in the architectures of intelligent control requires special procedures performed upon acquired knowledge. This affects the structure of world representation and it is intimately linked with mechanisms of behavior generation. This paper illuminates algorithms of autonomous learning performed via nested clustering which is goal driven and exercises simulation of decision making process.","PeriodicalId":219623,"journal":{"name":"Proceedings of Tenth International Symposium on Intelligent Control","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128529659","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1995-08-27DOI: 10.1109/ISIC.1995.525039
A. Wildberger
Exploratory research in intelligent and adaptive control systems at the Electric Power Research Institute (EPRI) attempts to combine the goal of finding innovative and valuable applications specific to the electric power industry with the equally important goal of establishing fundamental theory or reliable procedures for the design and development of adaptive and intelligent systems. The largest block of this research consists of twenty-one projects that make up the "Intelligent Control Systems Initiative", jointly funded by EPRI and the National Science Foundation. Most of these projects began in 1993 and will continue for three years. The rest of EPRI's exploratory research projects, that address intelligent and adaptive control in general, are part of the program in applied mathematics and information science (AM&IS). This area emphasizes fundamental theory and design procedures. However, the AM&IS program also includes projects that seek to invent and test new techniques and new applications whose success appears relatively risky at the beginning of the project.
{"title":"Survey of projects relating to intelligent control at the Electric Power Research Institute","authors":"A. Wildberger","doi":"10.1109/ISIC.1995.525039","DOIUrl":"https://doi.org/10.1109/ISIC.1995.525039","url":null,"abstract":"Exploratory research in intelligent and adaptive control systems at the Electric Power Research Institute (EPRI) attempts to combine the goal of finding innovative and valuable applications specific to the electric power industry with the equally important goal of establishing fundamental theory or reliable procedures for the design and development of adaptive and intelligent systems. The largest block of this research consists of twenty-one projects that make up the \"Intelligent Control Systems Initiative\", jointly funded by EPRI and the National Science Foundation. Most of these projects began in 1993 and will continue for three years. The rest of EPRI's exploratory research projects, that address intelligent and adaptive control in general, are part of the program in applied mathematics and information science (AM&IS). This area emphasizes fundamental theory and design procedures. However, the AM&IS program also includes projects that seek to invent and test new techniques and new applications whose success appears relatively risky at the beginning of the project.","PeriodicalId":219623,"journal":{"name":"Proceedings of Tenth International Symposium on Intelligent Control","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122527032","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1995-08-27DOI: 10.1109/ISIC.1995.525076
D. Krantz, Maria L. Gini
We have developed and demonstrated a technique to allow a follower vehicle to closely repeat the path established by a leader vehicle. The leader records the range-to-obstruction to either or both sides as well as its own estimated position and orientation. A non-model-based technique is used to correlate the follower's estimate of position to the leader's. The technique is tolerant of changes in the environment between passes of the leader and follower, and eliminates the correspondence problem in a model-based system where both the leader and follower need to correctly identify a particular object as being the same landmark. In this paper, we describe the active cross-path correlation algorithm and initial experimental results.
{"title":"Active cross-path correlation for autonomous vehicle path following","authors":"D. Krantz, Maria L. Gini","doi":"10.1109/ISIC.1995.525076","DOIUrl":"https://doi.org/10.1109/ISIC.1995.525076","url":null,"abstract":"We have developed and demonstrated a technique to allow a follower vehicle to closely repeat the path established by a leader vehicle. The leader records the range-to-obstruction to either or both sides as well as its own estimated position and orientation. A non-model-based technique is used to correlate the follower's estimate of position to the leader's. The technique is tolerant of changes in the environment between passes of the leader and follower, and eliminates the correspondence problem in a model-based system where both the leader and follower need to correctly identify a particular object as being the same landmark. In this paper, we describe the active cross-path correlation algorithm and initial experimental results.","PeriodicalId":219623,"journal":{"name":"Proceedings of Tenth International Symposium on Intelligent Control","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126757311","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1995-08-27DOI: 10.1109/ISIC.1995.525043
L. Carmichael, G. Saridis
This paper introduces a novel methodology, reliability-based complexity (RBC), that uses system models and a priori statistics in order to regulate the sensor measurements and algorithms (static and dynamic information) utilized by intelligent machines during the execution of some task. The objective is to produce tasks with the greatest level of performance and the least amount of cost. Task performance is evaluated through the development of reliability estimates that measure the individual types of uncertainties present in the task. These reliability estimates, in conjunction with cost measures, are incorporated into the mathematical framework developed in RBC. Within this framework, the optimal information selected for each task ensures that the task will be executed with maximum reliability and minimum cost. A case study involving robotic assembly is presented in order to illustrate these results.
{"title":"Reliability-based complexity in intelligent machines","authors":"L. Carmichael, G. Saridis","doi":"10.1109/ISIC.1995.525043","DOIUrl":"https://doi.org/10.1109/ISIC.1995.525043","url":null,"abstract":"This paper introduces a novel methodology, reliability-based complexity (RBC), that uses system models and a priori statistics in order to regulate the sensor measurements and algorithms (static and dynamic information) utilized by intelligent machines during the execution of some task. The objective is to produce tasks with the greatest level of performance and the least amount of cost. Task performance is evaluated through the development of reliability estimates that measure the individual types of uncertainties present in the task. These reliability estimates, in conjunction with cost measures, are incorporated into the mathematical framework developed in RBC. Within this framework, the optimal information selected for each task ensures that the task will be executed with maximum reliability and minimum cost. A case study involving robotic assembly is presented in order to illustrate these results.","PeriodicalId":219623,"journal":{"name":"Proceedings of Tenth International Symposium on Intelligent Control","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126404297","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1995-08-27DOI: 10.1109/ISIC.1995.525086
P. Gonsalves, A. Caglayan
Demonstrates fuzzy logic terminal guidance for a surface-to-surface missile. A conventional proportional-integral-derivative (PID) guidance scheme is first employed to form the basis for the fuzzy terminal guidance. The PID guidance scheme employs heading and flight path angle errors derived from sensor imagery of target location to compute angular rate commands. The fuzzy logic terminal guidance design procedure involves translating the conventional PID guidance scheme into a fuzzy control structure with the associated fuzzification of input/output variables, encoding of decision-making knowledge into rulebases, and defuzzification. A novel scheme is implemented in which fuzzy rulebases are used to generate and combine PD and PI portions of the guidance commands. Additionally, the fuzzy terminal guidance includes rulebases to control speed and to perform gain scheduling. Performance evaluations demonstrate the feasibility of the approach.
{"title":"Fuzzy logic PID controller for missile terminal guidance","authors":"P. Gonsalves, A. Caglayan","doi":"10.1109/ISIC.1995.525086","DOIUrl":"https://doi.org/10.1109/ISIC.1995.525086","url":null,"abstract":"Demonstrates fuzzy logic terminal guidance for a surface-to-surface missile. A conventional proportional-integral-derivative (PID) guidance scheme is first employed to form the basis for the fuzzy terminal guidance. The PID guidance scheme employs heading and flight path angle errors derived from sensor imagery of target location to compute angular rate commands. The fuzzy logic terminal guidance design procedure involves translating the conventional PID guidance scheme into a fuzzy control structure with the associated fuzzification of input/output variables, encoding of decision-making knowledge into rulebases, and defuzzification. A novel scheme is implemented in which fuzzy rulebases are used to generate and combine PD and PI portions of the guidance commands. Additionally, the fuzzy terminal guidance includes rulebases to control speed and to perform gain scheduling. Performance evaluations demonstrate the feasibility of the approach.","PeriodicalId":219623,"journal":{"name":"Proceedings of Tenth International Symposium on Intelligent Control","volume":"100 4","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133720932","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1995-08-27DOI: 10.1109/ISIC.1995.525082
R. Rajagopalan, R. Cheng, V. R. Ayyadevara, G. Huard
This paper presents the system architecture and performance of a robotic workcell capable of deburring used components. Such a workcell should possess the ability to automatically probe, reconstruct the surface geometry, determine the tool path for further machining, and finally carry out edge finishing. The robot chosen is YAMAHA Zeta-1 robot designed specifically for deburring. Probing is accomplished by means of a displacement sensor. This obviates the need for an additional probing stage in the process. The workcell controller has an information processing part consisting of a PC-parallel processor network, an interface to communicate with the controller of the deburring robot and a hardware interface for the probe. The surface of the workpiece is probed by the robot, directed along a path prescribed by the PC-parallel processor network, based on the original CAD database. The probed points, thus collected are used to reconstruct the surface probed and extrapolate to obtain the edge profile. The tool path constructed from the edge profile is then used to direct the robot during machining. Results of edge tracking of a saddle surface in three dimensions is presented. The geometric tacking error is found to be within the resolution of the command signals (0.063 mm).
{"title":"System architecture and edge tracking performance of an automated robotic deburring workcell","authors":"R. Rajagopalan, R. Cheng, V. R. Ayyadevara, G. Huard","doi":"10.1109/ISIC.1995.525082","DOIUrl":"https://doi.org/10.1109/ISIC.1995.525082","url":null,"abstract":"This paper presents the system architecture and performance of a robotic workcell capable of deburring used components. Such a workcell should possess the ability to automatically probe, reconstruct the surface geometry, determine the tool path for further machining, and finally carry out edge finishing. The robot chosen is YAMAHA Zeta-1 robot designed specifically for deburring. Probing is accomplished by means of a displacement sensor. This obviates the need for an additional probing stage in the process. The workcell controller has an information processing part consisting of a PC-parallel processor network, an interface to communicate with the controller of the deburring robot and a hardware interface for the probe. The surface of the workpiece is probed by the robot, directed along a path prescribed by the PC-parallel processor network, based on the original CAD database. The probed points, thus collected are used to reconstruct the surface probed and extrapolate to obtain the edge profile. The tool path constructed from the edge profile is then used to direct the robot during machining. Results of edge tracking of a saddle surface in three dimensions is presented. The geometric tacking error is found to be within the resolution of the command signals (0.063 mm).","PeriodicalId":219623,"journal":{"name":"Proceedings of Tenth International Symposium on Intelligent Control","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132359196","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1995-08-27DOI: 10.1109/ISIC.1995.525051
M. Hanson, R. Tolson
Kinematically redundant robots are robots that have more degrees of freedom than necessary to complete a desired task. Traditionally, the extra degrees of freedom have been used to optimize a single criterion such as joint torques minimization, obstacle avoidance, or minimization of flexible base vibrations. Because these approaches do not consider hardware limitations such as joint and rate limits, optimizing a single criterion often leads to high joint velocities and instabilities. To overcome these problems, it has been suggested that multiple criteria be optimized. Although optimizing multiple criteria offers the possibility of stabilizing joint solutions, it is difficult to choose the required weights associated with each criterion. This paper presents results using a fuzzy logic supervisor to decide the relative importance of each criterion and compute time-varying weights.
{"title":"Using a fuzzy supervisor to optimize multiple criteria in redundant robots","authors":"M. Hanson, R. Tolson","doi":"10.1109/ISIC.1995.525051","DOIUrl":"https://doi.org/10.1109/ISIC.1995.525051","url":null,"abstract":"Kinematically redundant robots are robots that have more degrees of freedom than necessary to complete a desired task. Traditionally, the extra degrees of freedom have been used to optimize a single criterion such as joint torques minimization, obstacle avoidance, or minimization of flexible base vibrations. Because these approaches do not consider hardware limitations such as joint and rate limits, optimizing a single criterion often leads to high joint velocities and instabilities. To overcome these problems, it has been suggested that multiple criteria be optimized. Although optimizing multiple criteria offers the possibility of stabilizing joint solutions, it is difficult to choose the required weights associated with each criterion. This paper presents results using a fuzzy logic supervisor to decide the relative importance of each criterion and compute time-varying weights.","PeriodicalId":219623,"journal":{"name":"Proceedings of Tenth International Symposium on Intelligent Control","volume":"74 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132191947","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1995-08-27DOI: 10.1109/ISIC.1995.525046
Seul Jung, T. Hsia
In this paper a neural network force/position control scheme is proposed to compensate uncertainties in both robot dynamics and unknown environments. The proposed impedance control allows us to regulate force directly by specifying a desired force. Training signals are proposed for a feedforward neural network controller. The robustness analysis of the uncertainties in environment position is presented. Simulation results are presented to show that both the position and force tracking are excellent in the presence of uncertainties in robot dynamics and unknown environments.
{"title":"Neural network techniques for robust force control of robot manipulators","authors":"Seul Jung, T. Hsia","doi":"10.1109/ISIC.1995.525046","DOIUrl":"https://doi.org/10.1109/ISIC.1995.525046","url":null,"abstract":"In this paper a neural network force/position control scheme is proposed to compensate uncertainties in both robot dynamics and unknown environments. The proposed impedance control allows us to regulate force directly by specifying a desired force. Training signals are proposed for a feedforward neural network controller. The robustness analysis of the uncertainties in environment position is presented. Simulation results are presented to show that both the position and force tracking are excellent in the presence of uncertainties in robot dynamics and unknown environments.","PeriodicalId":219623,"journal":{"name":"Proceedings of Tenth International Symposium on Intelligent Control","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133057202","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1995-08-27DOI: 10.1109/ISIC.1995.525050
J. Ware, J. Undery
This paper describes a system that locates and identifies components in an automated manufacturing process. The system uses a network of processors (an array of transputers) to construct and hold the workspace model, and to extract the feature measurements used to facilitate component identification. A MLP artificial neural network is then used to identify the components using the feature measurements obtained from the model. In an earlier version of this system goodness-of-fit was used to classify components, however, that method has drawbacks that neural networks overcome. The original design of the system was modular enabling a straightforward substitution of the component classification methods.
{"title":"Locating and identifying components in a robot's workspace using a hybrid computer architecture","authors":"J. Ware, J. Undery","doi":"10.1109/ISIC.1995.525050","DOIUrl":"https://doi.org/10.1109/ISIC.1995.525050","url":null,"abstract":"This paper describes a system that locates and identifies components in an automated manufacturing process. The system uses a network of processors (an array of transputers) to construct and hold the workspace model, and to extract the feature measurements used to facilitate component identification. A MLP artificial neural network is then used to identify the components using the feature measurements obtained from the model. In an earlier version of this system goodness-of-fit was used to classify components, however, that method has drawbacks that neural networks overcome. The original design of the system was modular enabling a straightforward substitution of the component classification methods.","PeriodicalId":219623,"journal":{"name":"Proceedings of Tenth International Symposium on Intelligent Control","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1995-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127943482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：