Evolving Simulated Modular Robots
نویسندگان
چکیده
We evolved virtual modular robots for locomotion using a direct and a generative encoding. An important benefit of modular robots in evolutionary robotics is the ability to quickly evaluate various morphologies and control systems in reality. Our use of a modular approach to construct robot morphologies is thus geared towards enabling us to construct robots in reality. However, our approach can also be used to solely construct virtual creatures. The advantage of our system is that due to accurate collision detection and our models being based on real world counterparts, the evolved behaviors can seem more realistic. Simulating virtual creatures usually takes up a lot of computational power and in this example we tried to reduce the computational power required through using an generative encoding based on an L-System, similar to what Hornby [1] implemented. Although there are other types of generative encodings L-Systems have, to our knowledge, not been frequently implemented in any modular robotic system. We compared both the direct encoding and the generative encoding for its performance [2] and the video can be found at https://www.youtube.com/watch?v=HCDftic1AdA. In the generative encoding multiple modules have the exact same control parameters which was color coded as illustrated in fig. 1. Through the reuse of genetic material we think that a generative encoding makes useful abstractions of both morphology and control which may be valuable for creating physical robots.
منابع مشابه
Crawling Out of the Simulation: Evolving Real Robot Morphologies Using Cheap, Reusable Modules
A current issue in evolutionary robotics involves the coevolution of robot controllers and body morphologies built from modular parts. As part of ongoing research, a model for the evolution of the morphologies and neural network controllers of robots is described. Several robots are evolved for locomotion in simulation built from modules representing cheap, preexisting parts and one is physical...
متن کاملEvolving Symmetric and Modular Neural Network Controllers for Multilegged Robots
Controllers for multilegged robots are characterized by modularity and symmetry. However, the controller symmetries necessary for generating appropriate gaits are often difficult to determine analytically. This paper utilizes a nature-inspired approach called Evolution of Network Symmetry and mOdularity (ENSO) to evolve such controllers automatically. It uses group theory to mutate symmetry sys...
متن کاملEvolving Motion Controllers for Articulated Robots
Articulated robots have a much more complex relationship between motor outputs and locomotion than wheel robots have. We can simplify this relationship using an abstraction layer of cyclic motor outputs corresponding to different kinds of locomotion. One way to create these cyclic motor outputs is by evolving a controller network consisting of oscillators, motor output nodes, and weighted edges...
متن کاملSelf-Organization of Communication in Evolving Robots
In this paper we present the results of an experiment in which a collection of simulated robots that are evolved for the ability to solve a collective navigation problem develop a communication system that allow them to better cooperate. The analysis of the obtained results indicates how evolving robots develop a non-trivial communication system and exploit different communication modalities.
متن کامل