Back to reality: Crossing the reality gap in evolutionary robotics

@article{Zagal2004BackTR,
  title={Back to reality: Crossing the reality gap in evolutionary robotics},
  author={Juan Crist{\'o}bal Zagal and Javier Ruiz-del-Solar and Paul A. Vallejos},
  journal={IFAC Proceedings Volumes},
  year={2004},
  volume={37},
  pages={834-839}
}
Abstract In this work a new method to evolutionary robotics is proposed, it combines into asingle framework, learning from reality and simulations. An illusory sub-system is incorporated as an integral part of an autonomous system. The adaptation of the illusory system results from minimizing differences of robot behavior evaluations in reality and in simulations. Behavior guides the illusory adaptation by sampling task-relevant instances of the world. Thus explicit calibration is not required… Expand

Figures from this paper

Combining Simulation and Reality in Evolutionary Robotics
TLDR
An algorithm is presented in which only the differences between the behavior fitness obtained in reality versus that obtained in simulations are used as feedback for adapting a simulation, which can be achieved through the co-evolution of robot behavior and simulator. Expand
From simulation to reality: migration of humanoid robot control
TLDR
This thesis investigated how to create a mechanism that provides a smooth gradient to transfer humanoid robot control from simulated robot to real robot, and developed motions in the simulations, with the help of Machine Learning, and transferred them to real robots successfully. Expand
Physical Scaffolding Accelerates the Evolution of Robot Behavior
TLDR
It is suggested that a simulation-to-reality pipeline may strike a good balance between accelerating evolution in simulation while anchoring the results in reality, free the investigator from having to prespecify the robot's morphology, and pave the way to scalable, automated, robot-generating systems. Expand
Dynamic Balancing of Mobile Robots in Simulation and Real Environments
TLDR
This work describes a method using a combination of repetitive splines and genetic algorithms to evolve a simulated control system for a humanoid robot, which is subsequently transferred to a real robot hardware. Expand
A Comparison of Sampling Strategies for Parameter Estimation of a Robot Simulator
TLDR
This work investigates sample selection, which is the method of choosing the robot controllers, evaluated in reality, that guide simulator tuning, and finds that strategies that select samples that show high fitness in simulation greatly outperform those that do not. Expand
Sim-to-Real Transfer with Neural-Augmented Robot Simulation
TLDR
This work introduces a method for training a recurrent neural network on the differences between simulated and real robot trajectories and then using this model to augment the simulator, which can be used to learn control policies that transfer significantly better to real environments than policies learned on existing simulators. Expand
Self-modeling in humanoid soccer robots
TLDR
The applicability, potential benefits, open problems and expected contributions that an emerging set of self-modeling techniques might bring on the development of humanoid soccer robots are discussed and directions of research are proposed. Expand
Quantifying the Reality Gap in Robotic Manipulation Tasks
TLDR
It is quantified the accuracy of various simulators compared to a real world robotic reaching and interaction task, and shows the relative strengths and weaknesses of numerous contemporary simulators, assisting researchers in the field in their selection of appropriate simulators for their use cases. Expand
UCHILSIM: A Dynamically and Visually Realistic Simulator for the RoboCup Four Legged League
TLDR
It is believed that the use of a highly realistic simulator might speed up the progress in the four legged league by allowing more people to participate in the challenge. Expand
Learning to Kick the Ball Using Back to Reality
TLDR
This work presents an approach for the automatic generation (from scratch) of ball-kick behaviors for legged robots, relying on the use of UCHILSIM, a dynamically accurate simulator, and the Back to Reality paradigm to evolutionary robotics, a recently proposed method for narrowing the difference between simulation and reality during robot behavior execution. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 24 REFERENCES
UCHILSIM: A Dynamically and Visually Realistic Simulator for the RoboCup Four Legged League
TLDR
It is believed that the use of a highly realistic simulator might speed up the progress in the four legged league by allowing more people to participate in the challenge. Expand
Learning to Kick the Ball Using Back to Reality
TLDR
This work presents an approach for the automatic generation (from scratch) of ball-kick behaviors for legged robots, relying on the use of UCHILSIM, a dynamically accurate simulator, and the Back to Reality paradigm to evolutionary robotics, a recently proposed method for narrowing the difference between simulation and reality during robot behavior execution. Expand
Evolution Versus Design: Controlling Autonomous Robots
  • P. Husbands, I. Harvey
  • Computer Science
  • Proceedings of the Third Annual Conference of AI, Simulation, and Planning in High Autonomy Systems 'Integrating Perception, Planning and Action'.
  • 1992
This paper sets out and justifies a methodology for the development of the control systems, or ‘cognitive architectures)) of autonomous mobile robots. It will be argued that the design b y hand ofExpand
A Description of the rUNSWift 2003 Legged Robot Soccer Team
This paper describes the 2003 world champion legged robot soccer team, rUNSWift. The 2003 rUNSWift team is enhanced in a num- ber of ways over previous teams; both long and short range collabo-Expand
Policy gradient reinforcement learning for fast quadrupedal locomotion
  • Nate Kohl, P. Stone
  • Computer Science, Engineering
  • IEEE International Conference on Robotics and Automation, 2004. Proceedings. ICRA '04. 2004
  • 2004
TLDR
A machine learning approach to optimizing a quadrupedal trot gait for forward speed using a form of policy gradient reinforcement learning to automatically search the set of possible parameters with the goal of finding the fastest possible walk. Expand
Evolutionary Robotics: The Biology, Intelligence, and Technology of Self-Organizing Machines
Evolutionary robotics is a new technique for the automatic creation of autonomous robots. Inspired by the Darwinian principle of selective reproduction of the fittest, it views robots as autonomousExpand
UChile 1 Team Description Paper
The U-Chile-1 four legged team is an effort of the Department of Electrical Engineering of the University of Chile in order to foster research in robotics at high level. U-Chile-1 was composed thisExpand
Minimal simulations for evolutionary robotics
TLDR
This thesis puts forward new techniques for evolving control systems for real robots using easy-to-build, fast-running simulations and discusses the best types of neural network, encoding scheme, genetic algorithm and genetic operators to use-and how to use them. Expand
Organismically-inspired robotics : homeostatic adaptation and teleology beyond the closed sensorimotor loop
In 1972, and later in 1979, at the peak of the golden era of Good Old Fashioned Artificial Intelligence (GOFAI), the voice of philosopher Hubert Dreyfus made itself heard as one of the few callsExpand
German Team 2003
TLDR
An architecture that provides mechanisms for parallel development of the robot is introduced, divided into modules that have well-defined tasks and interfaces and which can be switched at runtime. Expand
...
1
2
3
...