Innately adaptive robotics through embodied evolution

Mahdavi, Siavash Haroun; Bentley, Peter J.

doi:10.1007/s10514-006-5941-6

Cited by 23 publications

(16 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To our knowledge, TBR-Evolution is the first algorithm designed to generate a large number of efficient gaits without requiring to learn each of them separately, or to test complex controllers for each direction. In addition, our experiments only rely on internal, embedded measurements, which is critical for autonomy, but not considered in most previous studies (e.g., Kohl and Stone (2004); Zykov et al (2004); Chernova and Veloso (2004); Yosinski et al (2011); Mahdavi and Bentley (2006)).…”

Section: Conclusion and Discussionmentioning

confidence: 99%

“…When dealing with physical legged robots, the majority of studies only considers the evolution of un-driven walking controllers and, most of the time, the task consists in finding a controller that maximizes the forward walking speed (Zykov et al, 2004;Chernova and Veloso, 2004;Hornby et al, 2005;Berenson et al, 2005;Yosinski et al, 2011;Mahdavi and Bentley, 2006). Papers on alternatives to evolutionary algorithms, like policy gradients (Kohl and Stone, 2004;Tedrake et al, 2005) or Bayesian optimization (Calandra et al, 2014;Lizotte et al, 2007), are also focused on robot locomotion along a straight line.…”

Section: Evolving Walking Controllersmentioning

confidence: 99%

See 1 more Smart Citation

Evolving a Behavioral Repertoire for a Walking Robot

Cully

Mouret

2016

Evolutionary Computation

View full text Add to dashboard Cite

Numerous algorithms have been proposed to allow legged robots to learn to walk. However, most of these algorithms are devised to learn walking in a straight line, which is not sufficient to accomplish any real-world mission. Here we introduce the Transferability-based Behavioral Repertoire Evolution algorithm (TBR-Evolution), a novel evolutionary algorithm that simultaneously discovers several hundreds of simple walking controllers, one for each possible direction. By taking advantage of solutions that are usually discarded by evolutionary processes, TBR-Evolution is substantially faster than independently evolving each controller. Our technique relies on two methods: (1) novelty search with local competition, which searches for both high-performing and diverse solutions, and (2) the transferability approach, which combines simulations and real tests to evolve controllers for a physical robot. We evaluate this new technique on a hexapod robot. Results show that with only a few dozen short experiments performed on the robot, the algorithm learns a repertoire of controllers that allows the robot to reach every point in its reachable space. Overall, TBR-Evolution introduced a new kind of learning algorithm that simultaneously optimizes all the achievable behaviors of a robot.

show abstract

Section: Conclusion and Discussionmentioning

confidence: 99%

Section: Evolving Walking Controllersmentioning

confidence: 99%

Evolving a Behavioral Repertoire for a Walking Robot

Cully

Mouret

2016

Evolutionary Computation

View full text Add to dashboard Cite

show abstract

“…To increase the speed of operation, this algorithm reduces the number of candidate solutions at the expense of accuracy. Similarly, EA has been applied for damage recovery on a snake robot with a damaged body in [23] and a four-legged robot with a broken leg [24].…”

Section: Resilient Robotsmentioning

confidence: 99%

Resilient Robots: Concept, Review, and Future Directions

Zhang

Gupta

2017

Robotics

View full text Add to dashboard Cite

This paper reviews recent developments in the emerging field of resilient robots and the related robots that share common concerns with them, such as self-reconfigurable robots. This paper addresses the identity of the resilient robots by distinguishing the concept of resilience from other similar concepts and summarizes the strategies used by robots to recover their original function. By illustrating some issues of current resilient robots in the design of control systems, physical architecture modules, and physical connection systems, this paper shows several of the possible solutions to facilitate the development of the new and improved robots with higher resilience. The conclusion outlines several directions for the future of this field.

show abstract

“…The methods usually employed are policy gradient methods [11,14,24] or evolutionary algorithms [9,13,20,4,15], which optimize the controller's performances according to a desired action. They were successfully applied to several domains, from snake crawling [18] to bipedal walking [24].…”

Section: Learning Low Level Controllersmentioning

confidence: 99%

Behavioral repertoire learning in robotics

Cully

Mouret

2013

Proceedings of the 15th Annual Conference on Genetic and Evolutionary Computation

View full text Add to dashboard Cite

Learning in robotics typically involves choosing a simple goal (e.g. walking) and assessing the performance of each controller with regard to this task (e.g. walking speed). However, learning advanced, input-driven controllers (e.g. walking in each direction) requires testing each controller on a large sample of the possible input signals. This costly process makes difficult to learn useful low-level controllers in robotics.Here we introduce BR-Evolution, a new evolutionary learning technique that generates a behavioral repertoire by taking advantage of the candidate solutions that are usually discarded. Instead of evolving a single, general controller, BR-evolution thus evolves a collection of simple controllers, one for each variant of the target behavior; to distinguish similar controllers, it uses a performance objective that allows it to produce a collection of diverse but high-performing behaviors. We evaluated this new technique by evolving gait controllers for a simulated hexapod robot. Results show that a single run of the EA quickly finds a collection of controllers that allows the robot to reach each point of the reachable space. Overall, BR-Evolution opens a new kind of learning algorithm that simultaneously optimizes all the achievable behaviors of a robot.

show abstract

Innately adaptive robotics through embodied evolution

Cited by 23 publications

References 18 publications

Evolving a Behavioral Repertoire for a Walking Robot

Evolving a Behavioral Repertoire for a Walking Robot

Resilient Robots: Concept, Review, and Future Directions

Behavioral repertoire learning in robotics

Contact Info

Product

Resources

About