Robot device and learning method of robot device
First Claim
1. A robot apparatus comprising:
- a plurality of legs respectively connected to a body to support said body;
driving means for independently moving each of said legs;
generating means for generating a plurality of walking patterns of the robot, each walking pattern being prescribed by a plurality of parameters;
control means for controlling said driving means based on said plurality of parameters so that the robot walks according to said plurality of walking patterns;
evaluating means for evaluating an adaptability of each of said plurality of walking patterns to each of a plurality of walking surfaces, whereby the robot evaluates said walking patterns by searching for, detecting and walking toward a target; and
updating means for updating said plurality of parameters based on the robot'"'"'s own evaluation of said walking patterns, whereby said updating means updates said parameters prescribing said walking patterns according to a genetic algorithm.
1 Assignment
0 Petitions
Accused Products
Abstract
A walking-type robot device and its learning method are disclosed, wherein the robot device is caused to perform walking that accords with the parameters for controlling the walking, the walking is evaluated, and the parameters are updated so that the very evaluation is enhanced. Besides, a walking-type robot device is provided with a controlling means for controlling the robot so as to cause it to perform walking that accords with parameters which prescribe the driving phase of each leg of the time of walking, an evaluating means for evaluating the velocity of the walking, and a parameter updating means for updating the parameters so that the evaluation of the walking by the evaluating means is enhanced.
33 Citations
7 Claims
-
1. A robot apparatus comprising:
-
a plurality of legs respectively connected to a body to support said body;
driving means for independently moving each of said legs;
generating means for generating a plurality of walking patterns of the robot, each walking pattern being prescribed by a plurality of parameters;
control means for controlling said driving means based on said plurality of parameters so that the robot walks according to said plurality of walking patterns;
evaluating means for evaluating an adaptability of each of said plurality of walking patterns to each of a plurality of walking surfaces, whereby the robot evaluates said walking patterns by searching for, detecting and walking toward a target; and
updating means for updating said plurality of parameters based on the robot'"'"'s own evaluation of said walking patterns, whereby said updating means updates said parameters prescribing said walking patterns according to a genetic algorithm. - View Dependent Claims (2)
-
-
3. A learning method for a walking robot, comprising the steps of:
-
generating walking patterns of the robot, each walking pattern being prescribed by a plurality of parameters;
controlling walking patterns of the robot based on said plurality of parameters;
evaluating an adaptability of each of said plurality of walking patterns to each of a plurality of walking surfaces, whereby the robot evaluates said walking patterns by searching for, detecting and walking toward a target; and
updating said plurality of parameters based on the robot'"'"'s own evaluation of said walking patterns, whereby said parameters prescribing said walking patterns are updated according to a genetic algorithm. - View Dependent Claims (4, 5, 6)
-
-
7. A robot apparatus comprising:
-
a plurality of legs respectively connected to a body to support said body;
driving means for independently moving each of said legs;
generating means for generating a plurality of walking patterns of the robot, each walking pattern being prescribed by a plurality of parameters;
control means for controlling said driving means based on said plurality of parameters so that the robot walks according to said plurality of walking patterns;
evaluating means for evaluating an adaptability of each of said plurality of walking patterns to each of a plurality of walking surfaces, whereby the robot evaluates said walking patterns by searching for, detecting and walking toward a target, wherein the evaluating means is operable to;
search and detect a first target located in a first position;
cause the robot to walk, according to a first walking pattern, toward the first target;
calculate the velocity and straightness of the robot while the robot walks toward the first target;
search and detect a second target located in a second position;
cause the robot to walk, according to a second walking pattern, toward the second target; and
calculate the velocity and straightness of the robot while the robot walks toward the second target; and
updating means for updating said plurality of parameters based on the robot'"'"'s own evaluation of said walking patterns, whereby said updating means updates said parameters prescribing said walking patterns according to a genetic algorithm.
-
Specification