Resilient Machines Through Continuous Selfmodeling Odf
- Slides: 17
Download presentation
Resilient Machines Through Continuous Self-Modeling Josh Bongard, Victor Zykov, and Hod Lipson, Science, Vol. 314, pp. 1118 -1121, 2006. Pattern Recognition 2010. 04. 06 Seung-Hyun Lee Soft Computing Lab.
Contents • Introduction • Motivation • Self Modeling • Experiments • Conclusion 1 / 15 16 S FT COMPUTING @ YONSEI UNIV. KOREA
Introduction • Animals – After injured, create qualitatively different compensatory behaviors • Robots – How robots can deal with this sort of unexpected damage? self modeling 2 / 15 16 S FT COMPUTING @ YONSEI UNIV. KOREA
Motivation • How can robot learn its own morphology? – Direct observation? – Database of past experience? • How can robot synthesize complex behaviors or recover from damage? – Trial and error? slow, costly, risky! • In this paper, – Inferring morphology: self-directed exploration – Complex behavior or recovering from damage: synthesize new behaviors using the resulting self models 3 / 15 16 S FT COMPUTING @ YONSEI UNIV. KOREA
Overall Process Self Modeling Prediction Modeling Testing S FT COMPUTING @ YONSEI UNIV. KOREA 4 / 15 16
Testing Self Modeling • In this process – Performs an arbitrary motor action – Records the resulting sensory data 5 / 15 16 S FT COMPUTING @ YONSEI UNIV. KOREA
Modeiling Self Modeling • Model synthesize component – Synthesizes a set of candidate self-models • Method – Before damage(topological modeling) • Greedy random-mutation hill climber algorithm • 16 parameters Robot initially knows how many body pars it is composed of, the size, weight and mass of each part, and angle-movement relations • 15 random models • 200 iterations • Evaluation: Euclidean distance between the centroid and where the centroid should be – After damage(parametric modeling) • Self-model is frozen • 8 parameters (volumes and masses are scaled by 10%~200%) 6 / 15 16 S FT COMPUTING @ YONSEI UNIV. KOREA
Prediction Self Modeling • Action synthesize component – Find a new action most likely to elicit the most information from the robot based on the current self model inferred 7 / 15 16 S FT COMPUTING @ YONSEI UNIV. KOREA
Self Modeling • After self modeling procedures(16 times repetition) – Create desired behaviors (D) – Execute by the physical robot 8 / 15 16 S FT COMPUTING @ YONSEI UNIV. KOREA
Self Modeling 9 / 15 16 S FT COMPUTING @ YONSEI UNIV. KOREA
Robot Experiments • Speculation – 4 upper and lower leg parts and a main body – 8 motorized joints(-90 ~ 90 degree range) • 0 degree: flat • Positive degree: upwards • Negative degree: downwards – 2 tilt sensors • Self model representation – Planar topological arrangement • Damage – Disabled one leg 10 /16 15 S FT COMPUTING @ YONSEI UNIV. KOREA
Design Experiments • Control variables – Computational efforts(250, 000 internal model simulations) – Physical actions(16) • Three algorithms – Algorithm 1: 16 random physical actions batch training(modeling) – Algorithm 2: Physical actions self modeling random action selection – Algorithm 3(proposed): Physical actions self modeling actions selection 11 /16 15 S FT COMPUTING @ YONSEI UNIV. KOREA
Result Experiments 12 /16 15 S FT COMPUTING @ YONSEI UNIV. KOREA
Result Experiments à Model-driven algorithm is more accurate than random baseline algorithms à A robot that actively chooses action on the basis of its current set of hypothesized self-models has a better chance of successfully inferring its own morphology 13 /16 15 S FT COMPUTING @ YONSEI UNIV. KOREA
Result Experiments à Automatically generated self-model was sufficiently predictive to allow the robot to consistently develop forward motion patterns without further physical trials 14 /16 15 S FT COMPUTING @ YONSEI UNIV. KOREA
Result Conclusion • Contribution – First physical system • Autonomously recover its own morphology with little prior knowledge • Optimize the parameters of its morphology after unexpected change – Show the possibility of unknown cognitive process • • Which organisms actively create and update self models in the brain? How and which sensor-motor signals are used to do this? What form these model take? Does human utilize multiple competing models? 15 /16 15 S FT COMPUTING @ YONSEI UNIV. KOREA
Thank you
Source: https://slidetodoc.com/resilient-machines-through-continuous-selfmodeling-josh-bongard-victor/
0 Response to "Resilient Machines Through Continuous Selfmodeling Odf"
Post a Comment