Citation Context

... evolutionary algorithms (EAs) are employed to automatically design robots or controllers of robots; this field is now called Evolutionary Robotics (ER) (Meyer et al., 1998; Nolfi and Floreano, 2004; =-=Lipson, 2005-=-; Doncieux et al., 2009). According to ER c©200X by the Massachusetts Institute of Technology Evolutionary Computation x(x): xxx-xxx J.-B. Mouret and S. Doncieux researchers, EAs are a promising candi...

Citation Context

...dependently. This approach typically starts from a very primitive template which evolves into a complex final form capable of locomotion and other behaviors [27, 30]. Studies using co-evolution have investigated movement patterns using a predesigned initial robot, which was then subjected to evolution and resulted in solutions capable of effective locomotion with unique body structures and controllers [2, 7]. These co-evolution experiments have shown promise in producing robots capable of executing both simple and complex tasks [11, 16]. The ultimate goal of evolutionary robotics is synthesis [25, 26], in which a machine is capable of creating other machines in a process that mimics natural evolution. Realizing this goal depends in part on being able to model physical material properties in a realistic manner in simulation. New issues related to bridging simulation and reality have emerged with the introduction of 3D printing technologies. The reality gap appeared when controllers were evolved in simulation and transferred to real systems [10, 22]. Koos et al. [24] recently introduced a simulation-reality disparity measure to account for the differences in performance between the virtual a...

Citation Context

...motions that those actuators will create. These characteristics of soft robots undermine the conventional strategy of designing structure first and adding controls last. Rather, to obtain efficient movement, it is essential to design the form and the controls of a soft robot concurrently. Conceptually, design optimization techniques [13] can be applied to enable simultaneous tuning of the structure and controls of a soft robot. Design optimization methods have previously been applied to evolve rigid robots from libraries of fundamental building blocks, such as simple bars and linear actuators [14, 15]. Design optimization for soft robots is more challenging, however. Firstly, physics simulations for continuum materials, such as finite-element analysis, are computationally expensive, particularly when applied to compute the response to dynamic loading. Secondly, soft structures are particularly sensitive to geometry, so fundamental building blocks, such as rods, spheres and prisms, cannot be used to physically construct prototypes. Soft robots Proceedings of the ASME 2009 International Mechanical Engineering Congress & Exposition IMECE2009 November 13-19, Lake Buena Vista, Florida, USA I - ...

Citation Context

...fer to engineering, these characteristics may not prove efficient for the engineered product because the environment and manner in which a BID is used determines how efficient it is, and this environment may be very different from the one in which the biological analogy operated efficiently in nature. Numerous BID authors echo this point when they note that designs in biology are highly specialized for their needs and the constraints in their environments [25], that the specific set of needs biological organisms evolved to meet may not reflect human needs and existing engineering capabilities [26], and that the constraints facing optimized biological systems may be very different from the constraints facing engineered products [27]. Third, even if bioinspired efficiency characteristics successfully transfer to engineering and result in an engineering design possessing an efficiency advantage, this advantage may be countered by unintended effects in other stages of the product life cycle that make the final design less efficient as a whole. For instance, a life cycle inventory study of a lotus-inspired self-cleaning surface from Raibeck et al. [22] found that the energy and resource eff...

Citation Context

...lers [2, 7]. These co-evolution experiments have shown promise in producing robots capable of executing both simple and complex tasks [11, 16]. The ultimate goal of evolutionary robotics is synthesis =-=[25, 26]-=-, inwhich amachine is capable of creatingother machines in a process that mimics natural evolution. Realizing this goal depends in part on being able to model physical material properties in a realist...

Citation Context

...particularly attractive for controllers because they have low computational cost, and their inputs and outputs can correspond directly to sensors and actuators, respectively. In evolutionary robotics =-=[19]-=-, an artificial genome encodes a robot’s control system and possibly its morphology. The control program is downloaded into a real or simulated robot, which is then let “loose” in an environment. The ...

Citation Context

...y achieving increasing attention also because it opens the possibility of designing innovative systems or finding solutions for complex problems that are quite hard to solve with classical approaches =-=[3,6]-=-. One of the main advantages of automatic design is that the designer can specify just the minimal set of constraints and objectives on the final system; for example, when designing neural network sys...

Citation Context

...gies between neural controllers and a digital muscle based joint-level control. Background and Related Work Evolutionary robotics (Floreano et al., 2008; Cliff et al., 1993; Nolfi and Floreano, 2000; =-=Lipson, 2005-=-) harnesses the process that has produced robust natural organisms and applies it to either physical or simulated robots. Evolved behaviors such as walking (Brooks, 1989), grasping (Bongard, 2010), an...