Lund, H. H. et al. "Evolving Robot Morphology." Proceedings of IEEE Fourth International Conference on Evolutionary Computation. (1997).  Home/Search   Document Details and Download   Summary   Related Articles   Check

....of existing agents in detail. The only way is to observe them carefully and use human intelligence to draw conclusions. Usually, the behavior of such agents is non deterministic, and their control systems are sophisticated, often coupled with morphology and very strongly connected functionally [12]. Thus for the purposes of studying behaviors and populations of individuals, one needs high level, intelligent support tools [10] It is not likely that automatic tools will soon be able to produce understandable, non trivial explanations of sophisticated artificial agents. However, it is ....

Lund H.H., Hallam J. and Lee W.-P. (1997) Evolving Robot Morphology. In: Proceedings of IEEE 4th International Conference on Evolutionary Computation. NJ. IEEE Press. Invited paper.

.... of the recent convergence in the maturation of physicsbased simulation packages and increase of raw computing power of personal computers [18] Research in this area generally falls into two categories: 1) the evolution of controllers for creatures with fixed [9] or parameterized morphologies [13, 15], and (2) the evolution of both the creatures morphologies and controllers simultaneously [4, 8, 12] Some work has also been carried out in evolving morphology alone [6] and evolving morphology with a fixed controller [11] Related work using mobile robots have also shown promising results in ....

Henrik H. Lund, John Hallam, and Wei-Po Lee. Evolving robot morphology. In Proceedings of CEC'97, pages 197--202, Piscataway, NJ, 1997. IEEE Press.

....agents in detail [3,5,11] The only way is to observe them carefully and use human intelligence to draw conclusions. Usually, the behavior of such agents is non deterministic, and their control systems are sophisticated, often coupled with morphology and very strongly connected functionally [12]. Thus for the purposes of studying behaviors and populations of individuals, one needs high level, intelligent support tools [9] In the real world, there are human biologists who investigate living organisms. In artificial worlds, such work is very tedious and time consuming, and it is not ....

Lund H.H., Hallam J. and Lee W.-P. (1997) Evolving Robot Morphology. In: Proceedings of IEEE 4th International Conference on Evolutionary Computation. NJ. IEEE Press. Invited paper.

.... the recent convergence in the maturation of physics based simulation packages and increase of raw computing power of personal computers [21] Research in this area generally falls into two categories: 1) the evolution of controllers for creatures with fixed [10, 18] or parameterized morphologies [15, 17], and (2) the evolution of both the creatures morphologies and controllers simultaneously [4, 9, 12, 14] Some work has also been carried out in evolving morphology alone [6] and evolving morphology with a fixed controller [13] Related work using mobile robots have also shown promising results ....

Henrik H. Lund, John Hallam, and Wei-Po Lee. Evolving robot morphology. In Proceedings of the 4th IEEE International Conference on Evolutionary Computation, pages 197--202, Piscataway, NJ, 1997. IEEE Press.

....badly. This is certainly not the way to get a fitness value of the controller in a real physical environment. A. EHW Controllers EHW controllers refer to those EHW that are used primarily as controllers for robots or any other devices (such as ATM switches or multiplexors) 37] 38] 56] [60]. Mizoguchi et al. 56] used EHW to control an artificial ant to follow the John Muir Trail. The trail was placed on a grid. The controller of the artificial ant, which was implemented by EHW through simulation, had one input and two outputs. The input contained information about whether the trail ....

H. H. Lund, J. Hallam, and W.-P. Lee, "Evolving robot morphology, " in Proc. 1997 IEEE Conf. Evolutionary Computat. (ICEC'97). Piscataway, NJ: IEEE, pp. 197--202.

....located at strategic positions on the retinal surface. Lichtensteiger and Eggenberger [33] evolved the morphology of a compound eye of a robot asked to move on a straight trajectory and observed that evolved individuals displayed higher density of photoreceptors in the front. Lund et al. [34] co evolved in simulation the control system and some characteristics of the body of Khepera like robots that were selected for their ability to navigate while avoiding obstacles. The morphological features that were allowed to change were the diameter of the body, the distance between the two ....

Lund, H.H. et al. (1997) Evolving robot morphology. In: Proceedings of the Fourth International Conference on Evolutionary Computation. New York: IEEE Press

....body and environment. Therefore, we have been working to co evolve both the brain and the body, simultaneously and continuously, from a simple controllable mechanism to one of sufficient complexity for a particular specialized task. Although we were not first to propose brain body coevolution [6, 24, 27], we have reduced it to practice several times. Over the next decade, we see three technologies that are maturing past threshold to make possible a new industry of inexpensive automatically designed machines. One is the increasing fidelity of advanced mechanical simulation, stimulated by profits ....

H. Lund, J Hallam, and W. Lee. Evolving robot morphology. In Proceedings of IEEE Fourth International Conference on Evolutionary Computation, pages 197--202. IEEE Press, 1997.

.... passive dynamics has made clear that a careful choice of morphology can lead to locomotion without any actuation or controller at all [16] Examples now abound that demonstrate the evolution of both the morphology and control of simulated agents [19, 20, 10, 4, 15] as well as real world robots [14, 9, 13] is possible. However, we argue in [2] that the coupled evolution of both morphology and control of adaptive agents is not as interesting in and of itself, but rather the implications of such studies open up a host of research questions regarding the evolution of adaptive behaviour that are not ....

Lund, H. H., Hallam, J., Lee. W.-P.: Evolving Robot Morphology. In: Proceedings of of the IEEE Fourth International Conference on Evolutionary Computation. IEEE Press (1997).

....the robots have a weak status of embodiment. e.g. the body of the robot is static, the position and characteristics of the sensors and actuators are modi ed and adapted to the environment by hand, not by genuine development (compare with recent studies on the evolution of robot morphology, e.g. [54]) The body (the robot s mechanical and electronical parts) is not living , and its state does not depend on the internal dynamics of the control program. If the robot s energy supply is interrupted (the robot dies ) the robot s body still remains in the same state. This is a fundamental ....

Henrik Hautop Lund, John Hallam, and Wei-Po Lee. Evolving robot morphology. In Proceedings of IEEE 4th International Conference on Evolutionary Computation. IEEE Press, 1997.

....in the previous subsection, could all be characterized as various degrees of software situatedness. Recently, however, researchers have begun to apply evolutionary methods also to the construction of physical structures and robot morphologies (in simulation) e.g. Funes and Pollack 1997; Lund et al. 1997), in some cases in co evolution with robot controllers (Cliff and Miller 1996, Lund and Miglino 1998. Cliff and Miller (1996) for example, simulated the co evolution of eyes (number and position of visual sensors) and brains (ANN controllers mapping visual input to motor output) of simple ....

Lund, Henrik H., Hallam, John, and Lee, W. (1997). Evolving robot morphology. In Proceedings of the IEEE Fourth International Conference on Evolutionary Computation. IEEE Press.

....and mutual determination of body, nervous system and environment, has been largely neglected so far in embodied AI research. A small number of researchers have, however, begun to study the evolution of physical structures and robot morphologies (e.g. Funes Pollack 1997; Lund et al. 1997), in some cases in co evolution with controllers, as, for example, the work of Cliff Miller (1996) in which co evolution of eyes (optical sensors) and brains (control networks) has been applied (in simulation) to pursuing and evading agents. The approach of (co ) evolutionary robotics is ....

Lund, H. H.; Hallam, J. & Lee, W. (1997) Evolving robot morphology. Proceedings of the IEEE Fourth International Conference on Evolutionary Computation.

....out of parts. Husbands 96] uses a distributed genetic algorithm and a distributed genetic algorithm hybridized with gradient decent techniques to evolve the cross section of optimal aircraft wing boxes. Lichtensteiger 99] presents a study of evolution of the morphology of the compound eye. In [Lund 97] evolution of a morphology of an auditory hardware is discussed. Mark 98] presents a framework for the study of sensor evolution in a continuous 2 dimensional virtual world (XRaptor) Finally, in recent years work has began on co evolving form and function for autonomous agents. Sims 94] ....

Lund, H. H., J. Hallam, and W-P. Lee. "Evolving Robot Morphology." Proceedings of IEEE Fourth International Conference on Evolutionary computation; IEEE Press, NJ, 1997.

....stimulation of one sensory modality causing perceptions in different senses (Cytowic 1995) But what about artificial or simulated new sensor channels (Cariani 1992) considers the construction of new sensors in real devices and (Pask 1959) provides an example by constructing an artificial ear. (Lund, Hallam, Lee 1997) evolve ears for the Khepera robot and (Lee, Hallam, Lund 1996) simulate an agent acquiring distance information from any number of sensors. There also exist several approaches coding some sensor parameters genetically (Cliff, Harvey, Husbands 1993) Menczer Belew 1994) Todd Wilson ....

Lund, H. H.; Hallam, J.; and Lee, W.-P. 1997. Evolving robot morphology. In Proc. IEEE 4th Int. Conference on Evolutionary Computation. NJ: IEEE.

....Gomi, T. 72] Gopalan, Vijayarangan, 296] Gorges Schleuter, Martina, 370] Gorrini, V. 105, 266] Gough, N. E. 62] Grefenstette, John J. 205, 371] Grocholewska Czurylo, A. 29] Grosenbaugh, Mark, 437] Gruau, Fr ed eric C. 106] Hajek, M. 17] Hall, Ernest L. 54] Hallam, J. [302, 320] Halme, Aarne, 137, 161, 192] Hamada, Kazuro, 298] Hamam, Y. M. 153] Han, Woong Gie, 314] Handa, H. 101] Handley, Simon G. 121, 372, 373, 374, 375] Handroos, H. 310] Haneda, H. 319] Hao, Hong, 189] Harashima, Fumio, 131, 166, 291] Hart, John, 138] Harvey, Inman, 75, 107, ....

....Kishan K. 273] Kumbla, K. 190] Kuniyoshi, Y. 228] Kuruma, Toshiji, 89] Kwok, D. P. 78, 400] Lee, C. S. George, 11, 20] Lee, C. S. G. 34] Lee, Chi Ho, 268] Lee, Jiann Der, 172, 226, 269] Lee, M. A. 79] Lee, Seung Ik, 272] Lee, T. H. 244] Lee, Wei Ming, 252] Lee, Wei Po, [302, 320] Lehtinen, Hannu, 56] Leitch, Donald Dewar, 80, 145, 176] Leung, C. H. 66] Leung, T. P. 400] Lewis, M. Anthony, 401] Li, G. 317] Liegeois, A. 177, 185] Lin, C. S. 303] Lin, Fang Chang, 214] Lin, Hoi Shan, 73, 108, 406] Logan, Brian, 230] Lopez, Luis R. 413] Lott, ....

[Article contains additional citation context not shown here]

Henrik Hautop Lund, J. Hallam, and Wei-Po Lee. Evolving robot morphology. In Proceedings of IEEE International Conference on Evolutionary Computation, pages 197--202, Indianapolis, IN, 13.-16. April 1997. IEEE, New York, NY. y(CCA47417/97) ga97aLund.

....is thus how to automate the integrated design of bodies and brains using a coevolutionary learning approach. Brain body coevo lution is a popular idea, but evolution of robot bodies is usually restricted to adjusting a few morphological parameters in an otherwise xed, human engineered automaton [5 7]. We propose that the key is to evolve both the brain and the body, simultaneously and continuously, from a simple controllable mechanism to one of sucient complexity for a task. We then require a replication process, that brings an exact copy of the evolved machine into reality. Finally, the ....

Lund, H., Hallam, J., Lee, W.: Evolving robot morphology. In: Proceedings of IEEE Fourth International Conference on Evolutionary Computation, IEEE Press (1997) 197-202

....discussed so far have only been concerned with the evolution adaptation of artificial nervous systems. Recently, however, researchers have begun to apply evolutionary methods also to the construction of physical structures and robot morphologies (in simulation) e.g. Funes and Pollack 1997; Lund et al. 1997), in some cases in co evolution with controllers (Cliff and Miller 1996, Lund and Miglino 1998) Cliff and Miller (1996) for example, simulated the co evolution of eyes (optical sensors) and brains (ANN controllers) of simple robotic agents which pursued and evaded each other in a ....

Lund, Henrik H., Hallam, John, and Lee, W. (1997). Evolving robot morphology. In Proceedings of the IEEE Fourth International Conference on Evolutionary Computation. IEEE Press.

....were successfully transferred to the physical robot afterwards. A counter example of the evolution of homing behavior realized entirely on a physical robot is given by Floreano Mondada (1996) Another interesting application of evolutionary methods is the evolution of agent morphologies (e.g. Lund et al. 1997; Sims, 1994) sometimes in co evolution with the agent s control mechanism (Cliff Miller, 1996; Lund Miglino, 1998) This approach aims to overcome what Funes Pollack (1997) call the chicken and egg problem of evolutionary robotics: Learning to control a complex body is dominated by ....

Lund, H. H., Hallam, J. & Lee, W. (1997). Evolving Robot Morphology. Proceedings of IEEE Fourth International Conference on Evolutionary Computation. IEEE Press.

....bodily units with some of the solidarity of organisms, cf. Section 2. While integrated robot body units are some way off in the future, a small but growing number of researchers have begun to study the evolution of physical structures and robot morphologies, e.g. Funes and Pollack, 1997, Lund et al. 1997a] in some cases in co evolution with controllers [Cliff and Miller, 1996, Lund and Miglino, 1998] For example, Cliff and Miller (1996) simulated the co evolution of eyes (optical sensors) and brains (ANN controllers) of simple robotic agents which pursued and evaded each other in a ....

Lund, H. H., Hallam, J., and Lee, W. (1997a). Evolving robot morphology. In Proceedings of the IEEE Fourth International Conference on Evolutionary Computation. IEEE Press.

.... to justify from an evolutionary perspective and potentially misleading [7] But evolution of robot bodies remains limited mostly to adjusting parameters of the sensory configuration (sensor positioning, and direction, receptive fields, and so on) 8, 27] Recent work by Lund, Hallam and Lee [29] addresses some morphological issues, such as body size and wheel radius together with evolution of a robot control program. The field of Evolutionary Hardware, which does adaptation of reconfigurable electronics, makes a strong claim, that evolution can take advantage of emergent physical ....

Lund, H., Hallam, J and Lee, W. (1997). Evolving Robot Morphology. Invited paper in Proceedings of IEEE Fourth International Conference on Evolutionary Computation. IEEE Press, NJ.

....to an environment needs an adequate body to inhabit. In nature, the brain for a body, and the body for a brain are exquisitely intertwined and co adapted after millions of years of coevolution. The idea of co evolving bodies and brains is becoming popular. Recent work by Lund, Hallam and Lee (Lund et al. 1997; Lee et al. 1996) for example, evolves in simulation a robot control program simultaneously with some parameters of its morphology such as sensor number and positioning and body size. Our work attempts to build from the opposite shore: we are using evolutionary techniques to create structures, ....

Lund, H., Hallam, J and Lee, W. (1997). Evolving Robot Morphology. Invited paper in Proceedings of IEEE Fourth International Conference on Evolutionary Computation. IEEE Press, NJ.

....on the evolution of control software for real robots. Evolutionary Robotics has become a field on its own [15] Some rely on carefully designed simulations [4] while others apply evolution directly in the real robot [6] Hybrid techniques [13] are a mixture of the two. Lund, Hallam and Lee [11] [14] have evolved in simulation both a robot control program and some parameters of its physical body (sensor number and positioning, body size, etc. Their last paper [14] addresses the possibility of coevolving a robot controller and auditory morphology for the task of (cricket) phonotaxis. They ....

....apply evolution directly in the real robot [6] Hybrid techniques [13] are a mixture of the two. Lund, Hallam and Lee [11] 14] have evolved in simulation both a robot control program and some parameters of its physical body (sensor number and positioning, body size, etc. Their last paper [14] addresses the possibility of coevolving a robot controller and auditory morphology for the task of (cricket) phonotaxis. They contemplate the possibility of designing a Lego robot simulator. 3 The Physical Model The resistance of the plastic material (ABS acrylonitrile butadiene styrene) of Lego ....

Lund, H., Hallam, J and Lee, W. (1997). Evolving Robot Morphology. Invited paper in Proceedings of IEEE Fourth International Conference on Evolutionary Computation. IEEE Press, NJ.

....a vivid demonstration of what is possible if the full physical manifestation of an agent can be put under evolutionary control. Evolving the physical properties of physical robots in the real world presents much more difficulty. Although there have been a few attempts to evolve a robot s actuators (Lund, Hallam, and Lee 1997; Fukuda 1989) the only real successes I am aware of concern the evolution of a robot s sensor morphology. This was first done by Harvey, Husbands, and Cliff (1994) where controllers were evolved for a visually guided robot that could perform a simple shape discrimination task. Each input to the ....

Lund, H., J. Hallam, and W.-P. Lee (1997). Evolving robot morphology. In Proceedings of the IEEE 4th International Conference on Evolutionary Computation. IEEE Press.

....However, successful results on real robots presented in recent years already show the practical feasibility of the method. One of the nice features of the evolutionary approach is its generality, that is the possibility to apply it to several different aspects robotics (e.g. mechanical design [19] or circuit design [29] and to several different control specifications. Furthermore, it can be fruitfully combined with other forms of adaptation (such as learning) and or with traditionally pre designed solutions. ....

H. H. Lund, J. Hallam, and W-P. Lee. Evolving robot morphology. In Proceedings of the IEEE 4th International Conference on Evolutionary Computation. IEEE Press, 1997.

....would hit an obstacle badly. This is certainly not the way to get a fitness value of the controller in a real physical environment. 3. 1 EHW Controllers EHW controllers refer to those EHW which are used primarily as controllers for robots or any other devices (such as ATM switches or multiplexors) [37, 38, 56, 55, 57, 58, 59, 60]. Mizoguchi et al. 56] used EHW to control an artificial ant to follow the John Muir Trail. The trail was placed on a grid. The controller of the artificial ant, which was implemented by EHW through simulation, had one input and two outputs. The input contained information about whether or not ....

H. H. Lund, J. Hallam, and W.-P. Lee, "Evolving robot morphology," in Proc. of the 1997 IEEE Conf. on Evolutionary Computation (ICEC'97), pp. 197--202, IEEE Press, Piscataway, NJ, USA, 1997.

....Evolutionary algorithms are versatile tools for searching large design spaces with little a priori domain specific knowledge. They have thus found use in evolutionary robotics in the design of behavior and control programs for robots, and in some cases, even in the design of robot bodyplans [Lund et al. 1997]. In our past work we have used evolution to design neural network controllers for a simulated robot under a variety of environmental constraints. The robot had the task of clearing a square room (much like the environment used in this paper) by push12 ing the boxes to the walls. Evolution found ....

Lund, H., Hallam, J., & Lee, W-P. (1997). Evolving Robot Morphology. In: IEEE 4th International Conference on Evolutionary Computation. IEEE Press.

No context found.

H. H. Lund, J. Hallam, and W.-P. Lee. Evolving Robot Morphology. In Proceedings of IEEE Fourth International Conference on Evolutionary Computation, pages 197-202, NJ, 1997. IEEE Press.

....LEGO Lab, 1999. 3. 3 Integrating New Hardware with LEGO Mindstorms We choose to use the LEGO Mindstorms robotic kit as the hardware platform for the adaptive pet robot, because this robotic kit allows fast prototyping, flexibility in the robot morphology (which we view as essential to robotics [4]) and easy integration of new hardware. For instance, for the adaptive pet robot, we made a number of new sensors available for the LEGO Mindstorms robotic kit, including digital compass sensor, bend sensor, directional hearing sensor, and speech recognition (through communication with host ....

Henrik Hautop Lund, John Hallam, and Wei-Po Lee. "Evolving Robot Morphology". Invited Paper. In Proceedings of IEEE 4th International Conference on Evolutionary Computation. IEEE Press, NJ, 1997.

....etc. This study is important and the students should learn about this relationship, since the performance of a robot controller is critical dependent on the robot body plan. This can, for instance, be shown with an evolutionary approach when co evolving robot controllers and robot body plans [5, 6]. For this educational purpose, we use the LEGO Mindstorms robots. We find this robot platform a good choice in this educational context, because it allows students with no engineering background to assemble robots with different body plans. The robots are easily assembled with the well known LEGO ....

Lund, H. H., Hallam, J., and Lee, W.-P. "Evolving Robot Morphology" In Proceedings of IEEE 4th International Conference on Evolutionary Computation. IEEE Press, NJ, 1997.

....we argued that for the simulation reality transfer it is advantageous to use control systems that allow generalisation over noise and other discrepancies between the simulator and the real world. Therefore, we used neural networks as control systems for the robot in that study. Secondly, we [1] [4] showed how it is possible to co evolve robot controllers and robot body plans. Some of this work was done in simulation only, because of the difficulties in designing self modifying hardware. However, we showed in [4] how to modify pre amplifiers, synthesizers, and mixers in an auditory system ....

....networks as control systems for the robot in that study. Secondly, we [1] 4] showed how it is possible to co evolve robot controllers and robot body plans. Some of this work was done in simulation only, because of the difficulties in designing self modifying hardware. However, we showed in [4] how to modify pre amplifiers, synthesizers, and mixers in an auditory system for a mobile robot. Our argument was that one can co evolve robot controllers and robot body plans in simulation, build the physical robot according to the evolved robot body plans, download the evolved control system, ....

H. H. Lund, J. Hallam, and W.-P. Lee. Evolving Robot Morphology. In Proceedings of IEEE Fourth International Conference on Evolutionary Computation, NJ, 1997. IEEE Press. Invited paper.

....robots. However, the circuit architecture is only a part of the hardware system, and ideally we would like to evolve the whole system. The hardware of a robot consists of both the circuit, on which the control system is implemented, and the sensors, motors, and physical structure of the robot. We [15] have previously argued that True EHW should evolve the whole hardware system, since the evolution and performance of the electronic hardware is largely dependent on the other parts of the hardware that constitute the system. The latter part is what we called the robot body plan. A robot body plan ....

....to change their architectures and behaviours dynamically and autonomously with their environment. Indeed, the design of a complete autonomous system must include self organisation, and self organisation means adaptation of both morphology and control architecture. Together with Lee and Hallam, we [13, 15] therefore investigated the possibility of co evolution between robot controllers and robot body plans in simulation, and we are currently building a LEGO robot system that allows evolution of robot body plans for real robots. The experiments showed that controllers and robot body plans are ....

H. H. Lund, J. Hallam, and W.-P. Lee. Evolving Robot Morphology. In Proceedings of IEEE Fourth International Conference on Evolutionary Computation, NJ, 1997. IEEE Press. Invited paper.

No context found.

Lund, H. H. et al. "Evolving Robot Morphology." Proceedings of IEEE Fourth International Conference on Evolutionary Computation. (1997).

No context found.

H. Lund, J Hallam, and W. Lee. Evolving robot morphology. In Proceedings of IEEE Fourth International Conference on Evolutionary Computation, pages 197--202. IEEE Press, 1997.

No context found.

H. Lund, J. Hallam, and W. Lee. Evolving robot morphology. In Proceedings of IEEE Fourth International Conference on Evolutionary Computation. IEEE Press, 1997.

No context found.

Lund, H.H., Hallam, J. and Lee, W.-P. Evolving Robot Morphology. In Proceedings of IEEE 4th International Conference on Evolutionary Computation. IEEE Press, 1997

No context found.

H. Lund, J Hallam, and W. Lee. Evolving robot morphology. In Proceedings of IEEE Fourth International Conference on Evolutionary Computation, pages 197--202. IEEE Press, 1997.

No context found.

H. Lund, J Hallam, and W. Lee. Evolving robot morphology. In Proceedings of IEEE Fourth International Conference on Evolutionary Computation, pages 197--202. IEEE Press, 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