P. Nordin, and W. Banzhaf, #An on-line method to evolvebehavior and to control a miniature robot in real time with genetic programming", in Adaptive Behavior,5#2#: 107-140, 1997.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....will likely take too much time because of the speed limit of a physical robot. The work in [3] shows that for a visualreaching task, it will take 2,000 hours with their equipments to learn the task. It is possible to reduce the time by running GP in simulation that samples data from the real world [9, 11, 13]. Another approach to cope with changes is to subject the evolved system to perturbation expecting that the resulting solution will be more tolerant. The work such as [7, 12] introduce perturbation at every step of evolutionary process. They report limited success. From the experience of our ....

Nordin, P. and Banzhaf W., "An on-line method to evolve behavior and to control a miniature robot in real time with genetic programming", in Adaptive Behavior, 5(2) : 107-140, 1997.

....as books. Genetic Algorithms and Robotics: A heuristic strategy for optimization, 338] 4.2 Journal articles The following list contains the references to every journal article included in this bibliography. The list is arranged in alphabetical order by the name of the journal. Adaptive Behavior, [86, 237] Advanced Technology for Developers, 414] Artif. Intell. Eng. UK) 218, 233] Artificial Intelligence, 119] Artificial Life, 83, 147, 197] BioSystems, 351] Comput. Ind. Eng. UK) 229] Control Engineering Practice, 90] IEE Colloq. Dig. 262] IEE Conf. Publ. ETSI konferenssi, ....

....Ashlock, Dan, 207] Aspragathos, N. A. 235] Aspragathos, Nikos A. 208, 307] Atmar, J. Wirt, 350] Aydin, K. K. 126] Baba, N. 101] Baba, Norio, 65] Baek, Seung Min, 314] Baffes, Paul T. 327, 328] Balakrishnan, Karthik, 210, 257, 286] Baluja, Shumeet, 258] Banzhaf, Wolfgang, [198, 202, 204, 237, 238, 287] Barnes, D. P. 88] Barrett, David, 437] Bartscht, E. 33] Beer, Randall D. 216] Bennett III, Forest H. 288] Bersano Begey, Tommaso F. 242] Bessi ere, Pierre, 160, 329, 416, 417, 418, 419, 420, 421, 422, 423, 424] Bikdash, M. 296] Biondi, Joelle, 146] Blume, Christian, 102, 370] ....

[Article contains additional citation context not shown here]

Peter Nordin and Wolfgang Banzhaf. An on-line method to evolve behavior and to control a miniature robot in real time with genetic programming. Adaptive Behavior, 5(2):107--140, Autumn 1996. y(SBS V. 29 No. 29) ga96aNordin.

....of the journal. ACM Computer Surveys, 651] Acta Crystallographica Section D: Biological Crystallography, 430] Acta Electronica Sinica (China) 102, 261, 371, 702] Acta Physica Sinica, 188] Acta Polytechnica Scandinavica, Electrical Engineering Series (Finland) 758] Adaptive Behavior, [333] Adv. Eng. Softw. UK) 517, 657, 689, 832] Advanced Composites Letters, 324] AES J. Audio Eng. Soc. 872] AI Communications, 141] AIAA Journal, 612, 870] AIAA Journal on Disc, 230, 337, 512, 660, 815] Analytical Chemistry, 24, 374, 633] Ann. Oper. Res. Netherlands) 182, 192, ....

....[479] Bakirtzis, A. G. 89, 233] Bala, Jerzy W. 593] Balakrishnan, P. V. 344] Balas, G. J. 23] Balasekar, S. 584] Balogh, S. 629] Baluja, Shumeet, 594] Bandyopadhyay, Sanghamitra, 747] Bangalore, Arjun S. 24] Bangalore, Shanthamallikarjuna Shivappa, 425] Banzhaf, Wolfgang, [333, 445, 846, 880] Barnes, J. W. 75] Barron, M. 373] Bartels, Christian, 45] Basile, Luciano, 595] Baskaran, Subbiah, 100] Bastian, A. 46] Baumgarten, G. 605] Beasley, J. E. 196] Beauchamp, James, 680, 891] Becker, Bernd, 889] Beckers, Mischa L. M. 47] Beder, Jay H. 626] Beer, Randall D. ....

[Article contains additional citation context not shown here]

Peter Nordin and Wolfgang Banzhaf. An on-line method to evolve behavior and to control a miniature robot in real time with genetic programming. Adaptive Behavior, 5(2):107--140, Autumn 1996. ySBS V. 29 No. 29 ga96aNordin.

....which involve real robots, on the one hand, and control architectures implemented as neural networks, on the other hand. More general reviews are to be found in [18] 4] 22] 52] 14] 37] and [20] Examples of evolutionary robotics applications involving non neural controllers are [5] [56], 15] or [26] II. The review Since 1994, about 30 papers have been published that describe results obtained when the neural controllers of real robots have been automatically designed through an evolutionary process. These papers are classified in Table I below, according to a general ....

.... dynamical processes like developmental programs than when they just change static structures like the chromosomes of traditional genetic algorithms [29] Another pending issue is that of assessing whether it is easier to evolve neural controllers than, for example, explicit control programs (e.g. [56]) or production rules systems (e.g. 5] although it has been argued that the former approach offers over the latter the advantages of generating smoother fitness landscapes [4] and of facilitating realistic injections of noise in specific parts of the controller [18] Likewise, it is presently ....

Nordin, P. and Banzhaf, W. An On-Line Method to Evolve Behavior and to Control a Miniature Robot in Real Time with Genetic Programming. Adaptive Behavior. 5, 2, 107-140. 1996.

....On Programming Computers by Means of Natural Selection and Genetics, 61] total 4 books 4.2 Journal articles The following list contains the references to every journal article included in this bibliography. The list is arranged in alphabetical order by the name of the journal. Adaptive Behavior, [320, 67] AI Expert, 186] Automatisierungstechnik (Germany) 386] BYTE, 128] Chin. J. Electron. China) 445] Comput. Chem. Eng. 350] Comput. Chem. Eng. UK) 454] Comput. Econ. Netherlands) 333] Computer Graphics, 525] Computers Mathematics with Applications, 370] Electr. Power ....

....523] Araki, Miyuhiko, 76] Arus, C. 287] Ashlock, Dan, 383, 459] Atkin, Marc C. 83, 476, 477] Atlan, Laurent, 478, 479, 508] Aytekin, T. 177] Backer, Gerriet, 265] Bal ate, Mojm ir, 384] Ballard, Dana H. 123, 155, 167, 178] Balogh, S. 350] Bancroft, C. 338] Banzhaf, Wolfgang, [143, 215, 240, 255, 253, 257, 259, 260, 266, 283, 309, 320, 323, 363, 372, 376, 385, 432, 480, 481] Barclay, Peter J. 293] Barklund, Jonas, 306] Barnes, D. P. 124] Barton, Geoffrey W. 210, 239, 316, 371] Bartton, G. W. 454] Bassanini, A. 515] Bastian, A. 386] Bauer, Eric T. 179] Beard, Nick, 482, 521] Bell, Larry, 381] Bengio, Samy, 85] Bengio, Yoshua, 85] Benini, Luca, ....

[Article contains additional citation context not shown here]

Peter Nordin and Wolfgang Banzhaf. An on-line method to evolve behavior and to control a miniature robot in real time with genetic programming. Adaptive Behavior, 5(2):107--140, Autumn 1996. y(SBS V. 29 No. 29) ga96aNordin.

.... and the environment [4] Asada et al. decomposes the input state space [20] 28] For simple robot tasks in a unknown and dynamic environment, researchers have applied evolution based learning algorithms to low level control architecture such as LISP like programming languages [17] 24][22], production rules (classifier systems) 30] 5] 10] and neural networks [3] 7] 23] 1] For our robot task and control architecture, we choose the evolutionary approach because we are most interested in navigation in a unknown environment. The evolutionary approach uses genetic algorithm (GA) ....

Peter Nordin and Wolfgang Banzhaf. An on-line method to evolve behavior and to control a miniature robot in real time with genetic programming. Adaptive Behavior, 5(2):107--140, 1997.

....of intelligent systems with life like properties by means of simulated evolution. An essential distinction among the approaches in evolutionary robotics is in the genetic structure that undergoes adaptation. Neural network controllers [3] genetic fuzzy systems [2] 7] 16] and genetic programming [13] were proposed as learning techniques for the control of real robots. We utilized the evolutionary learning method proposed in the previous section to adapt a wall following behavior of a mobile robot that is implemented by means of fuzzy control rules. The evolutionary design process takes place ....

P. Nordin, W. Banzhaf, "An On-Line Method to Evolve Behavior and to Control a Miniature Robot in Real Time with Genetic Programming" Adaptive Behaviour, 5 (2), pp. 107-140, (1997).

.... pop(t) recombine and mutate parents to create pop(t 1) determine fitness of pop(t 1) t = t 1 until best individual is good enough Figure 1: General Evolutionary Computation Algorithm of robot control architecture should be evolved There are a number of options: high level code [40] machine code [37], parameter settings for a hand designed system [39] situationaction rules [7] and entire rule based strategies [15] Perhaps the most innovative direction, however, is the combination of evolutionary computation with artificial neural networks. Neural networks allow the evolutionary process to ....

P. Nordin and W. Banzhaf. An on-line method to evolve behavior and to control a miniature robot in real time with genetic programming. Adaptive Behavior, 5(2):107-- 140, 1997.

....for this task. It is fast enough even without a real time kernel. A piece of interface software was written to control the servos via the serial RS232 line. 3 EVOLVING ROBOT CONTROLERS To evolve control programs we applied GP [2, 8] as a learning method. A variant of stochastic sampling as in [14] has been used where the fitness evaluation for an individual is performed by allowing the individual to control the robot only for a short time (here, about 10 seconds) 3.1 THE GENETIC PROGRAMMING SYSTEM Here, a simple tree based steady state GP algorithm is applied. An individual controls ....

Peter Nordin and Wolfgang Banzhaf. An on-line method to evolve behavior and to control a miniature robot in real time with genetic programming. Adaptive Behaviour, 5(2):107--140, 1996.

....to unseen environments and tasks. In order to overcome the latter problem artificially generated noise is added sometimes to the simulated environment. We have earlier reported on a first series of experiments using GP to control a real robot trained in real time with actual sensor values [21] [22]. In such an environment, the system has to evolve robust controllers because noise is present everywhere and the number of real life training situations is infinite. In addition, it is highly impractical to reset the robot to a predefined state before evaluating a fitness case. Consequently, we ....

....returns an action in the form of a vector of two motor speeds: f(s 1 ; s 2 ; s 3 ; s 4 ; s 5 ; s 6 ; s 7 ; s 8 ) fm 1 ; m 2 g (1) Function f models the simple stimulus response behavior of the robot. Our original approach was to evolve this function through interaction with the environment [21] [22]. The second more efficient approach reported here generates a simulation or world model instead of deriving motor speeds directly from sensor input. This involves another function, g, which codes the relation between motor speed values, sensory inputs and fitness. Figure 6: The Training ....

[Article contains additional citation context not shown here]

Nordin P. and Banzhaf W. (1997). An On-Line Method to Evolve Behavior and To Control a Miniature Robot in Real Time with Genetic Programming. Adaptive Behavior, 5(2), 107 --- 140.

.... Several experiments have also been performed where a controller program has been evolved directly through genetic programming [11, 17, 28] We, too, have reported earlier on our first experiments using GP to control a real robot that has been trained in real time with actual sensor values [23, 25]. In a real environment our system had to evolve robust controllers because noise was present everywhere and the number of real life training situations was virtually infinite. Consequently we were forced to devise an online learning method which ensured learning of behavior while each ....

.... of the robot partly can be attributed to the built in generalization capabilities of the genetic programming system [21, 22] We have further demonstrated that the memory based GP control system can evolve much smoother and less chaotic behavior than the non memory GP control system described in [23]. We would like to evaluate the usefulness of our approach with agent systems that have a wider set of possible actions. In such systems it would be infeasible to use exhaustive search for finding the best action according to a world model. Handley has previously demonstrated the feasibility of GP ....

P. Nordin and W. Banzhaf (1997a) An On-Line Method to Evolve Behavior and to Control a Miniature Robot in Real Time with Genetic Programming. Adaptive Behavior 5 107 - 140.

....the fitness has, on the other hand, a serious disadvantage: It will require a very long time to train the system. There is a method, however, to maintain good generalization abilities by using a large number of fitness cases while not requiring that much computing power, stochastic sampling 1 [NB95c,NB97]: From a large set of fitness cases a small sample is chosen randomly before a fitness evaluation takes place. Fitness evaluation is performed only on the small sample, and hence is less expensive than evaluation on the entire fitness case set. Thus, no program from the population can really ....

Peter Nordin and Wolfgang Banzhaf. An on-line method to evolve behavior and to control a miniature robot in real time with genetic programming. Adaptive Behavior, 26:107 -- 140, 1997.

No context found.

P. Nordin, and W. Banzhaf, #An on-line method to evolvebehavior and to control a miniature robot in real time with genetic programming", in Adaptive Behavior,5#2#: 107-140, 1997.

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