Humanoid robot with collision avoidance and trajectory recovery capabilities

Patent metadata

Field	Value
Publication number	US-10179406-B2
Application number	US-201515311800-A
Country	US
Kind code	B2
Filing date	Jun 5, 2015
Priority date	Jun 5, 2014
Publication date	Jan 15, 2019
Grant date	Jan 15, 2019

Abstract

A humanoid robot which can move on its lower limb to execute a trajectory and is capable of detecting intrusion of obstacles in a safety zone defined around its body as a function of its speed is provided. Preferably when the robot executes a predefined trajectory, for instance a part of a choreography, the robot which avoids collision with an obstacle will rejoin its original trajectory after avoidance of the obstacle. Rejoining trajectory and speed of the robot are adapted so that it is resynchronized with the initial trajectory. Advantageously, the speed of the joints of the upper members of the robot is adapted in case the distance with an obstacle decreases below a preset minimum. Also, the joints are stopped in case a collision of the upper members with the obstacle is predicted.

First claim

The invention claimed is:

1. A method for controlling a trajectory of at least one of upper and lower members of a humanoid robot, said method comprising:
storing an initial trajectory with a target point in a memory of the robot;
acquiring, from at least a sensing procedure controlled from on-board the robot, data representative of a position of one or more obstacles; and
calculating, by a processor on-board the robot:
an envelope of said robot, said envelope depending on a speed and a direction of the robot;
a relative position of the envelope and the one or more obstacles;
a probability of collision of the envelope with one of said one or more obstacles; and,
a series of commands in order to: i) avoid collision of the envelope with said one of said one or more obstacles; and ii) at least one of rejoin the target point of the initial trajectory stored in memory within an initial timing, and preserve the general direction and content of an initial gesture, to change in at least one of space and time at least one of the trajectory or a speed of the at least one of upper and lower members of the robot.

2. The method of claim 1 , wherein the series of commands to rejoin the target point comprises:
calculating a point of the initial trajectory;
calculating a straight line between the point of the initial trajectory, and the target point;
calculating a speed of the robot in the straight line to rejoin the target point at the initial timing; and
defining the trajectory of the robot as the initial trajectory until the point of the initial trajectory, then the straight line between the point of the initial trajectory, and the target point.

3. The method of claim 1 , wherein the series of commands to preserve the general direction and content of the initial gesture comprises calculating a target position for a point of the robot, said point on a trajectory, computing a maximum joint velocity for a chain containing the point, said maximum joint velocity depending of the distance between the point and said one of said one or more obstacles.

4. The method of claim 1 , wherein the initial trajectory is referenced in space and time.

5. The method of claim 1 , wherein the sensing procedure is performed by at least one of a plurality of laser lines generators, a plurality of imaging sensors, a plurality of acoustic sensors and a plurality of contact detectors.

6. The method of claim 5 , wherein the sensing procedure is performed by at least two of a plurality of laser lines generators, a plurality of imaging sensors, a plurality of acoustic sensors and a plurality of contact detectors.

7. The method of claim 1 , wherein the data representative of a location of one or more obstacles is an extraction of a number of characteristic points from a map of pixels representing a probability of absence of said one of said one or more obstacles in said pixels.

8. The method of claim 1 , wherein an envelope of a footprint of the robot is calculated as a function of the speed of the robot and predetermined guards around the robot.

9. The method of claim 8 , wherein the series of commands is calculated to determine a changed trajectory to avoid collision of the envelope with any obstacle.

10. The method of claim 9 , wherein the series of commands is further calculated to determine a changed trajectory and a changed speed to rejoin a target point of the initial trajectory at a time when the robot should have reached the target point on the initial trajectory.

11. The method of claim 1 , wherein an upper member of the robot comprises a chain of segments articulated together by motored joints.

12. The method of claim 11 , wherein an envelope of a member of the robot is calculated as a function of a predetermined security distance around the articulated segments.

13. The method of claim 11 , wherein the series of commands is calculated to determine a reduction in angular speeds of the motors of the joints of the articulated segments when the envelope approaches said one of said one or more obstacles.

14. The method of claim 13 , wherein the reduction in angular speeds of the motors of the joints is calculated to saturate a velocity of the chain to a maximum safety value, said maximum safety value being calculated based on a target velocity of a point in the chain, and positions of said one or more obstacles relative to said point in the chain.

15. A humanoid robot comprising:
at least an upper member and a lower member;
a memory storing an initial trajectory with a target point and computer code instructions;
a plurality of sensing modules configured to acquire data representative of a position of one or more obstacles; and
a processor configured to execute said computer code instructions to calculate:
an envelope of said robot, said envelope depending on a speed and a direction of the robot;
a relative position of the envelope and the one or more obstacles;
a probability of collision of the envelope with one of said one or more obstacles; and
a series of commands in order to: i) avoid collision of the envelope with said one of said one or more obstacles; and ii) at least one of rejoin the target point of the initial trajectory stored in memory within an initial timing, and preserve the general direction and content of an initial gesture, to change in at least one of space and time at least one of the trajectory or a speed of the at least one of upper and lower members of the robot.

Assignees

• Softbank Robotics Europe

Inventors

• Dalibard Sébastien
• Garcia Aldenis
• Collette Cyrille
• Garcia Nicolas
• Souchet Lucas

Classifications

• B25J11/0035

  Dancing, executing a choreography · CPC title

• B25J9/1694

  characterised by use of sensors other than normal servo-feedback from position, speed or acceleration sensors, perception control, multi-sensor controlled systems, sensor fusion · CPC title

• B25J9/1676 Primary

  Avoiding collision or forbidden zones · CPC title

• B25J9/1666

  Avoiding collision or forbidden zones · CPC title

• G05B2219/39091

  Avoid collision with moving obstacles · CPC title