Motion generation of multi-legged robot in complex terrains by using estimation of distribution algorithm

Motion generation is one of the most important and challenging problems in multi-legged robot research. Most of the existing methods show a good fulfillment of the requirements of robots in structured environments. However, it still faces many challenges to generate motions effectively and quickly for multi-legged robot works in complex environments. In this paper, we put forward a method which converts the motion generation problem into a Multi-objective Optimization Problem (MOP), which will make the robot not only run as fast as possible, but also save energy, and then use a distribution estimation algorithm, the trend prediction model method, to obtain motions for a six-legged robot. Experiments show that this method is effective.

[1]  Min Jiang,et al.  Embodied concept formation and reasoning via neural-symbolic integration , 2010, Neurocomputing.

[2]  Manuela M. Veloso,et al.  An evolutionary approach to gait learning for four-legged robots , 2004, 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (IEEE Cat. No.04CH37566).

[3]  Prahlad Vadakkepat,et al.  Genetic algorithm-based optimal bipedal walking gait synthesis considering tradeoff between stability margin and speed , 2009, Robotica.

[4]  Fan Zhang,et al.  Fuzzy neural network based dynamic path planning , 2012, 2012 International Conference on Machine Learning and Cybernetics.

[5]  Min Jiang,et al.  Integration of Global and Local Metrics for Domain Adaptation Learning Via Dimensionality Reduction , 2017, IEEE Transactions on Cybernetics.

[6]  Naoyuki Kubota,et al.  Evolving spiking neural network for robot locomotion generation , 2015, 2015 IEEE Congress on Evolutionary Computation (CEC).

[7]  Yukinori Kakazu,et al.  Acquiring Adaptive Gaits For Many-Legged Robots by Reinforcement Learning. , 1994 .

[8]  Ben Goertzel,et al.  Improving machine vision via incorporating expectation-maximization into Deep Spatio-Temporal learning , 2014, 2014 International Joint Conference on Neural Networks (IJCNN).

[9]  Min Jiang,et al.  Trend Prediction Model Based Multi-Objective Estimation of Distribution Algorithm , 2016 .

[10]  Isao Shimoyama,et al.  Dynamics in the dynamic walk of a quadruped robot , 1989, Adv. Robotics.

[11]  Hod Lipson,et al.  Evolving robot gaits in hardware: the HyperNEAT generative encoding vs. parameter optimization , 2011, ECAL.

[12]  黄 忠强,et al.  Trend Prediction Model Based Multi-Objective Estimation of Distribution Algorithm , 2016 .

[13]  Jie Yan,et al.  A Review of Gait Optimization Based on Evolutionary Computation , 2010, Appl. Comput. Intell. Soft Comput..

[14]  Naoyuki Kubota,et al.  Cyclic motion generation for intelligent robot by evolutionary computation , 2013, 2013 IEEE Workshop on Robotic Intelligence in Informationally Structured Space (RiiSS).

[15]  Takenori Obo,et al.  Motion generation of multi-legged robot by using knowledge transfer in rough terrain , 2016, 2016 IEEE Symposium Series on Computational Intelligence (SSCI).

[16]  Kalyanmoy Deb,et al.  Multiobjective optimization , 1997 .

[17]  Hiroaki Satoh,et al.  Minimal generation gap model for GAs considering both exploration and exploitation , 1996 .

[18]  Fernando Santos Osório,et al.  Applying Genetic Algorithms to Control Gait of Physically Based Simulated Robots , 2006, 2006 IEEE International Conference on Evolutionary Computation.

[19]  Dominic Anthony Messuri.  Optimization of the locomotion of a legged vehicle with respect to maneuverability , 1985 .

[20]  Pedro Larrañaga,et al.  A Review on Estimation of Distribution Algorithms , 2002, Estimation of Distribution Algorithms.

[21]  Dilip Kumar Pratihar,et al.  Soft computing-based gait planners for a dynamically balanced biped robot negotiating sloping surfaces , 2009, Appl. Soft Comput..

[22]  Kalyanmoy Deb,et al.  Optimal path and gait generations simultaneously of a six-legged robot using a GA-fuzzy approach , 2002, Robotics Auton. Syst..