Determining sensor parameters and model parameters of a robot

Patent metadata

Field	Value
Publication number	US-9844872-B1
Application number	US-201514797912-A
Country	US
Kind code	B1
Filing date	Jul 13, 2015
Priority date	Jul 13, 2015
Publication date	Dec 19, 2017
Grant date	Dec 19, 2017

Abstract

Methods, apparatus, systems, and computer readable media are provided for determining: 1) sensor parameters for sensors of a robot and 2) model parameters of a dynamic model of the robot. The sensor parameters and model parameters are determined based on applying, as values for known variables of a system equation of the robot, sensor readings and position values for each of a plurality of instances of a traversal of the robot along a trajectory. The system equation of the robot is a dynamic model for the robot that includes sensor models substituted for one or more corresponding variables of the dynamic model. The system equation includes unknown variables representing unknown sensor biases for the sensors of the robot and unknown model parameters of the dynamic model of the robot. Solutions to the unknown variables are generated and utilized to determine the sensor parameters and the model parameters.

First claim

What is claimed is:

1. A computer implemented method, comprising:
determining, by one or more processors for each of a plurality of instances of a traversal of a robot along a trajectory:
sensor readings for a plurality of sensors of the robot, each of the sensor readings corresponding to a respective one of the sensors, and the sensors each measuring a corresponding parameter that is in addition to position,
position values for a plurality of actuators of the robot, each of the position values corresponding to a respective one of the actuators,
velocity values and acceleration values that are based on the position values, and
gain values for the sensors of the robot;

applying the sensor readings, the position values, the velocity values, the acceleration values, and the gain values for each of the instances of a group of the instances as values for known variables of a system equation of the robot, the system equation being an equation of motion that includes sensor models substituted for a plurality of variables of the equation of motion, and the system equation including the known variables and unknown variables, the unknown variables representing unknown sensor biases for the sensors of the robot and unknown model parameters of a dynamic model of the robot;
generating a solution to each of a plurality of the unknown variables based on the applied sensor readings, position values, velocity values, acceleration values, and gain values for the group of the instances;
determining sensor biases based on the solutions to the unknown variables representing the unknown sensor biases, wherein the sensor biases each define a bias for a respective one of the sensors;
determining model parameters based on the solutions to the unknown variables representing the unknown model parameters, wherein the model parameters are for the dynamic model of the robot;
applying a correction to at least one of the sensors of the robot based on the determined sensor bias for the sensor; and
applying the model parameters to the dynamic model of the robot.

2. The method of claim 1 , wherein the sensor models each include sensor model known variables of the known variables and a sensor model unknown variable of the unknown variables;
wherein the sensor model unknown variable represents the unknown sensor biases; and
wherein applying the sensor readings, the position values, the velocity values, the acceleration values, and the gain values for each of the instances of the group as values for the known variables comprises applying the sensor readings and the gain values as the sensor model known variables.

3. The method of claim 2 , wherein the sensor models comprise torque sensor models for joint torque sensors of the sensors of the robot.

4. The method of claim 1 , wherein applying the sensor readings, the position values, the velocity values, the acceleration values, and the gain values for each of the instances of the group as the values for the known variables for the system equation of the robot comprises:
generating a plurality of system equations by applying the sensor readings, the position values, the velocity values, the acceleration values, and the gain values for each of the instances of the group as the known variables of the system equation of the robot;
wherein generating the solution to each of the plurality of the unknown variables based on the applied sensor readings, position values, velocity values, acceleration values, and the gain values comprises:
generating the solution to each of the plurality of the unknown variables based on the system equations.

5. The method of claim 4 , wherein generating the solution to each of the plurality of the unknown variables based on the system equations comprises applying a fitting procedure to generate the solutions.

6. The method of claim 5 , wherein the fitting procedure is a least squares fitting procedure.

7. The method of claim 4 , wherein generating the solution to each of the plurality of the unknown variables based on the system equations comprises:
stacking the system equations; and
generating the solutions based on applying a least squares fitting procedure to the stacked system equations.

8. The method of claim 1 , wherein applying the correction to at least one of the sensors of the robot based on the determined sensor biases for the sensor comprises:
applying a digital correction to future sensor readings of the sensor.

9. The method of claim 1 , further comprising:
controlling the actuators of the robot based on the dynamic model of the robot with the applied model parameters.

10. The method of claim 1 , further comprising:
traversing the robot along the trajectory.

11. A robotic system comprising:
a plurality of actuators of a robot;
sensors of the robot, each of the sensors measuring at least one property associated with a corresponding one of the actuators;
logic providing control commands to the actuators to traverse the robot along a trajectory;
a group determination engine identifying, for each of a plurality of instances of the traversal of the robot along the trajectory:
sensor readings for the sensors, each of the sensor readings corresponding to a respective one of the sensors, and the sensor readings each being for a corresponding parameter that is in addition to position,
position values for the actuators, each of the position values corresponding to a respective one of the actuators,
velocity values and acceleration values that are based on the position values, and
gain values for the sensors of the robot;
the group determination engine further determining a group of the instances;

a system equations generation engine configured to apply the sensor readings, the position values, the velocity values, the acceleration values, and the gain values for each of the instances of the group of the instances as values for known variables of a system equation of the robot, the system equation being an equation of motion that includes sensor models substituted for a plurality of variables of the equation of motion, and the system equation including the known variables and unknown variables, the unknown variables representing unknown sensor biases for the sensors of the robot and unknown model parameters of a dynamic model of the robot; and
a sensor parameters and model parameters calculation engine configured to:
generate a solution to each of a plurality of the unknown variables based on the applied sensor readings, position values, velocity values, acceleration values, and gain values for the group of the instances,
determine sensor biases for the sensors based on the solutions to the unknown variables that represent the unknown sensor biases, and
determine model parameters for the dynamic model of the robot based on the solutions to the unknown variables that represent the unknown model parameters.

12. The system of claim 11 , wherein the sensor models each include sensor model known variables of the known variables and a sensor model unknown variable of the unknown variables;
wherein the sensor model unknown variable represents the unknown sensor biases; and
wherein the system equations generation engine applies the sensor readings and the gain values as the sensor model known variables.

13. The system of claim 12 , wherein the sensor models comprise torque sensor models for joint torque sensors of the sensors of the robot.

14. The system of claim 11 , wherein the system equations generation engine is configured to:
generate a plurality of system equations by applying the sensor readings, the position values, the velocity values, the ac

Assignees

• X Dev Llc

Inventors

• Buschmann Thomas

Classifications

• B25J9/163 Primary

  learning, adaptive, model based, rule based expert control · CPC title

• B25J9/162

  Mobile manipulator, movable base with manipulator arm mounted on it · CPC title

• B25J9/1633

  compliant, force, torque control, e.g. combined with position control · CPC title

• G05B19/406

  characterised by monitoring or safety (G05B19/19 takes precedence) · CPC title

• B25J9/1653

  parameters identification, estimation, stiffness, accuracy, error analysis · CPC title