Model-based control with uncertain motion model

Patent metadata

Field	Value
Publication number	US-12078972-B2
Application number	US-202217654580-A
Country	US
Kind code	B2
Filing date	Mar 12, 2022
Priority date	Mar 12, 2022
Publication date	Sep 3, 2024
Grant date	Sep 3, 2024

Abstract

A probabilistic feedback controller for controlling an operation of a robotic system using a probabilistic filter subject to a structural constraint on an operation of the robotic system is configured to execute a probabilistic filter estimates a distribution of a current state of the robotic system given a previous state of the robotic system based on a motion model of the robotic system perturbed by stochastic process noise and a measurement model of the robotic system perturbed by stochastic measurement noise having an uncertainty modeled as a time-varying Gaussian process represented as a weighted combination of time-varying basis functions with weights defined by corresponding Gaussian distributions. The probabilistic filter recursively updates both the distribution of the current state of the robotic system and the Gaussian distributions of the weights of the basis functions selected to satisfy the structural constraint indicated by measurements of the state of a robotic system.

First claim

The invention claimed is:

1. A probabilistic feedback controller for controlling an operation of a robotic system using a probabilistic filter subject to a structural constraint on an operation of the robotic system, comprising: at least one processor; and a memory having instructions stored thereon that, when executed by the at least one processor, cause the controller to:
collect digital representations of a sequence of measurements of a state of a robotic system at different instances of time indicative of a structural constraint on a shape of a function of a parameter affecting the operation of the robotic system;
execute a probabilistic filter configured to recursively estimate a distribution of a current state of the robotic system given a previous state of the robotic system based on a motion model of the robotic system perturbed by stochastic process noise and a measurement model of the robotic system perturbed by stochastic measurement noise, wherein one or a combination of the motion model, the process noise, the measurement model, and the measurement noise includes the parameter having an uncertainty modeled as a time-varying Gaussian process represented as a weighted combination of time-varying basis functions with weights defined by corresponding Gaussian distributions, such that a time-varying mean of the Gaussian process is a function of the basis functions modified with means of the corresponding Gaussian distributions, and a time-varying variance of the Gaussian process is a function of the basis function modified with variances of the corresponding Gaussian distributions, wherein the probabilistic filter recursively updates both the distribution of the current state of the robotic system and the Gaussian distributions of the weights of the basis functions, and wherein each of the basis function is selected to satisfy the structural constraint indicated by the sequence of measurements of the state of a robotic system; and
execute a control action determined based on an estimate of the distribution of the current state of the robotic system to change the current state of the robotic system according to a control objective.

2. The probabilistic feedback controller of claim 1 , wherein each measurement in the sequence of measurements of the state of the robotic system relates to a value of the parameter through one or a combination of the motion model and the measurement model, such that the sequence of measurements of the state of the robotic system relates to a sequence of values of the parameters statistically conserving the structural constraint on the shape of the function of the parameter.

3. The probabilistic feedback controller of claim 2 , wherein smoothed representation of data fitted into the sequence of values of the parameters includes a function having a shape conserving the structural constraint.

4. The probabilistic feedback controller of claim 2 , wherein statistical properties of the sequence of values of the parameters relies on the structural constraint.

5. The probabilistic feedback controller of claim 2 , wherein imposing the structural constraint of the shape of the function of the parameter reduces a difference between the state of the robotic system predicted using the motion model having the parameter and the measured state of the robotic system.

6. The probabilistic feedback controller of claim 1 , wherein the structural constraint is determined offline based on experimental or simulation data of the operation of the robotic system, such that the basis functions are selected offline in response to determining the structural constraint.

7. The probabilistic feedback controller of claim 1 , wherein the structural constraint is determined online during the control of the robotic system, such that the basis functions are selected during the operation of the robotic system in response to determining the structural constraint.

8. The probabilistic feedback controller of claim 7 , wherein to select the basis functions, the processor is configured to
actuate the robotic system using a range of control inputs to collect training measurements of the state of the robotic system;
estimate the state of the robotic system caused by the actuation with the range of control inputs using different probabilistic filters with different structural constraints on the shape of the function of the parameter affecting the operation of the robotic system;
select the structural constraint from the probabilistic filter that estimates the training measurements best according to a cost function; and
select the basis functions conforming to the selected structural constraint.

9. The probabilistic feedback controller of claim 7 , wherein to select the basis functions, the processor is configured to
actuate the robotic system using a range of control inputs to collect training measurements of the state of the robotic system;
estimate the state of the robotic system caused by the actuation with the range of control inputs using a probabilistic filter without a nominal function of the parameter;
determine a pattern of distribution of data points of differences between the training measurements and corresponding estimation of the state of the robotic system;
transform the pattern into the structural constraint.

10. The probabilistic feedback controller of claim 9 , wherein the pattern is represented as a system of equations dependent on the parameter, such that the basis functions enabling a solution of the system of equations satisfy the structural constraint.

11. The probabilistic feedback controller of claim 1 , wherein the probabilistic filter is a Kalman filter with the motion model including the parameter inserted into the Kalman filter allowing the probabilistic filter to perform a dual update of the state of the robotic system and the time-varying Gaussian process.

12. The probabilistic feedback controller of claim 1 , wherein the probabilistic filter is a particle filter with the motion model including the parameter associated with each particle filter allowing the probabilistic filter to perform a dual update of the state of the robotic system and the time-varying Gaussian process.

13. The probabilistic feedback controller of claim 1 , wherein to satisfy the structural constraint each of the basis functions is selected to be an odd function.

14. The probabilistic feedback controller of claim 1 , wherein to satisfy the structural constraint each of the basis functions is selected to be an even function.

15. The probabilistic feedback controller of claim 1 , wherein the robotic system includes a vehicle traveling on a road having an uncertain road-surface condition modeled as the time-varying Gaussian process represented by a weighted combination of odd basis functions.

16. The probabilistic feedback controller of claim 15 , wherein the function of the parameter is a friction function, and the shape of the friction function is asymmetric.

17. The probabilistic feedback controller of claim 1 , wherein the robotic system includes multiple vehicles forming an uncertain traffic flow modeled as the time-varying Gaussian process across different segments, wherein each segment is represented by a weighted combination of odd or even basis functions.

18. The probabilistic feedback controller of claim 1 , wherein the robotic system includes a set of drones controlled by a magnetic field represented by a nonlinear function having a structural constraint of being curl-free, such that the basis functions are selected as a combination of sine and cosine basis f

Assignees

• Mitsubishi Electric Res Laboratories Inc

Inventors

• Berntorp Karl
• Menner Marcel

Classifications

• G05B13/042

  in which a parameter or coefficient is automatically adjusted to optimise the performance · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G05B13/048 Primary

  using a predictor · CPC title

• G05B13/045

  using a perturbation signal · CPC title