Method for the improved detection of the collision of a robot with its environment, system and computer program product implementing said method

Patent metadata

Field	Value
Publication number	US-10478969-B2
Application number	US-201414766568-A
Country	US
Kind code	B2
Filing date	Feb 7, 2014
Priority date	Feb 14, 2013
Publication date	Nov 19, 2019
Grant date	Nov 19, 2019

Abstract

A method of detecting collision between a robot and its environment comprises: generating, on the basis of a dynamic model of the robot, a signal representative of the collisions between the robot and environment, the signal being a residual r; carrying out an adaptive high-pass filtering of the residual r; determining, in a recursive manner, an adaptive threshold T composed of at least one first dynamic term TΔ1 equal to an item of information regarding the parametric uncertainties between the model and real behavior of the robot, the parametric uncertainties being related to a first variable characteristic of the state of the robot among the following variables: position, velocity or acceleration of an articulation of the robot, or a linear or nonlinear function of one of these variables or of a combination; comparing the filtered residual with the adaptive threshold T to deduce the existence or otherwise of a collision.

First claim

The invention claimed is:

1. A method for detecting a collision between a robot composed of a plurality of bodies linked together by at least one articulation and the environment of the robot, said method being executed by a system for detecting a collision comprising a processor, a control member, and a physical interface, the method comprising the steps of:
moving at least one articulation of the robot by manipulating, via a motor of the robot, a torque applied to the at least one articulation of the robot,
receiving, from a measurement device of the control member, at least a measure of a variable characteristic of the state of the robot among a position, a velocity or an acceleration of the at least one articulation of the robot,
executing, with the processor, a computer program stored on a non-transitory computer readable medium, the computer program stored on the non-transitory computer readable medium including instructions that when executed by the processor cause the processor to automatically perform the steps of:
generating, from a dynamic model of the robot, a signal representative of the collisions between the robot and the environment of the robot, said signal being called a residual r and comprising as many components as articulations of the robot,
adaptive high-pass filtering of the residual r so as to render the residual r independent of parametric or non-parametric uncertainties related to low-frequency phenomena,
determining recursively adaptive thresholds T composed of at least one first dynamic term T Δ1 dependent on the parametric uncertainties between said model and the real behavior of the robot, said parametric uncertainties being related to a first variable e 1 characteristic of the state of said robot from among the following variables: the position, the velocity or the acceleration of the at least one articulation of the robot, or a linear or nonlinear function of one of these variables or of a combination of these variables, wherein determining recursively an adaptive threshold T of the adaptive thresholds T at a given instant is based upon a previously determined adaptive threshold of the adaptive thresholds T at an instant previous to the given instant, so that the adaptive threshold T is dependent on said parametric uncertainties between said model and the real behavior of the robot,
comparing the filtered residual with the adaptive threshold T so as to deduce therefrom the existence or otherwise of the collision,
deciding, as a result of the comparison of the filtered residual and the adaptive threshold T, if a collision between the robot and the environment of the robot occurs,
providing the decision to the interface, and
utilizing the decision provided to the interface, wherein

the adaptive high-pass filtering is implemented via a recursive least squares algorithm and comprises:
estimating, in a recursive manner, the coefficients of the transfer function of the high-pass filter G 0 −1 , and

filtering the residual r with the estimated high-pass filter.

2. The method of detecting collision as claimed in claim 1 , wherein said adaptive threshold T is composed of the sum of several dynamic terms T Δ1 , T Δ2 , T Δ3 , each equal to an item of information regarding the parametric uncertainties between said model and the real behavior of the robot, said parametric uncertainties being related to a different variable, characteristic of the state of said robot from among the following variables: the position, the velocity or the acceleration of a fixed point of the robot, or a linear or nonlinear function of one of these variables or of a combination of these variables.

3. The method of detecting collision as claimed in claim 1 , in which said adaptive threshold T furthermore comprises a static term T static configured so as to be greater than a measurement noise level.

4. The method of detecting collision as claimed in claim 1 , in which an additional step of temporal filtering is applied to the filtered residual for each of the components.

5. The method of detecting collision as claimed in claim 4 , in which the additional step of temporal filtering is a step of root mean square (RMS) calculation.

6. The method of detecting collision as claimed in claim 1 , in which the step of determining at least one first dynamic term T Δ1 , T Δ2 , T Δ3 of the adaptive threshold T is carried out by means of a recursive least squares algorithm.

7. The method of detecting collision as claimed in claim 6 , in which an additional step of temporal filtering is applied to said dynamic term T Δ1 , T Δ2 , T Δ3 of the adaptive threshold T for each of its components.

8. The method of detecting collision as claimed in claim 7 , in which the additional step of temporal filtering is a step of root mean square calculation.

9. The method of detecting collision as claimed in claim 1 , in which the step of determining at least one first dynamic term T Δ1 , T Δ2 , T Δ3 of the adaptive threshold T comprises the following sub-steps:
estimating, in a recursive manner, the coefficients of the transfer function Δ i modeling the parametric uncertainties related to said first variable e i characteristic of the state of said robot, and
filtering said first variable e i characteristic of the state of said robot with a filter of transfer function Δ i estimated in the previous step so as to obtain a dynamic term T Δ1 , T Δ2 , T Δ3 of the adaptive threshold T.

10. The method of detecting collision as claimed in claim 1 , in which step of comparing the filtered residual with the adaptive threshold T so as to deduce therefrom the existence or otherwise of a collision comprises the following sub-steps:
comparing, for each articular component, the filtered residual with the adaptive threshold T, and
concluding the existence of a collision if, for at least K components, the filtered residual is greater than the adaptive threshold T, K being a strictly positive predetermined integer less than or equal to the number of articulations of the robot.

11. The method of detecting collision as claimed in claim 1 , in which step of generating the residual consists of the following sub-steps:
determining, on the basis of an item of information regarding the state of the robot and by way of a dynamic model, an estimation of the articular torques of the robot,
performing a measurement of the state of the robot, for example a measurement of the articular torques, and
calculating the residual as the difference between the estimation and the measurement of the state of the robot.

12. The method of detecting collision as claimed in claim 1 , in which the parametric uncertainties related to the articular acceleration of the robot are uncertainties regarding the inertia matrix of the robot.

13. The method of detecting collision as claimed in claim 1 , in which the parametric uncertainties related to the articular velocity of the robot are uncertainties regarding the matrix of the centrifugal and Coriolis terms of the robot and/or regarding the viscous frictions.

14. The method of detecting collision as claimed in claim 1 , in which a nonlinear function is the sign function or the exponential function or the absolute value function.

15. The method of detecting collision as claimed in claim 14 , wherein the parametric uncertainties related to the sign of the articular velocity of the robot are uncertainties regarding the dry frictions.

16. A system for command of a robot comprising:
a control member for the manipulation of the robot;
an interface for the exchange of information regarding the state of the robo

Assignees

• Commissariat Energie Atomique

Inventors

• Caldas Alex
• Makarov Maria
• Grossard Mathieu

Classifications

• B25J9/1605

  Simulation of manipulator lay-out, design, modelling of manipulator · CPC title

• Y10S901/09

  Closed loop, sensor feedback controls arm movement · CPC title

• Y10S901/46

  Sensing device · CPC title

• B25J9/1674 Primary

  characterised by safety, monitoring, diagnostic · CPC title

• B25J9/06

  characterised by multi-articulated arms · CPC title