Granted Patent B2

US 12,479,439 · App. 18/182,474 · Granted Nov 25, 2025

Inventors: Yan Ming Jonathan Goh (Palo Alto, CA); Franck Djeumou (Palo Alto, CA)

Assignees: Toyota Research Institute, Inc.; Toyota Jidosha Kabushiki Kaisha

B60W30/18172B60W40/101G05B13/027G05B13/048B60W2520/10B60W2520/14B60W2520/20B60W2520/28B60W2720/26

Abstract

Claims (19)

1 . A method comprising: operating a vehicle with a control framework incorporating an ExpTanh function; calculating prior slip data based on measurements obtained by the control framework; using a first Neural Network predictive model to determine ExpTanh parameters based upon the prior slip data and the measurements; and inputting a slip parameter and the ExpTanh parameters into the ExpTanh function to estimate a tire force, wherein the ExpTanh function has multiple inflection points based on the ExpTanh parameters; and adjusting a vehicle control input based on the tire force estimate.

2 . The method of claim 1 , wherein using the first Neural Network predictive model to determine the ExpTanh parameters is further based on scaling parameters; and the method further comprises: using, in parallel with the first Neural Network predictive model, a second Neural Network predictive model to determine the scaling parameters based on the prior slip data.

3 . The method of claim 2 , wherein the method further comprises: adjusting the tire force by the scaling parameters to estimate a pair of lateral and longitudinal tire forces.

4 . The method of claim 3 , wherein using the first or second Neural Network predictive models includes using a stochastic gradient descent algorithm to obtain the ExpTanh parameters.

5 . The method of claim 4 , wherein the prior slip data contains slip angle data and longitudinal slip ratio data; and the slip parameter is a current total slip.

6 . The method of claim 2 , further comprising selecting a confidence parameter and wherein using the first Neural Network predictive model to determine the ExpTanh parameters is further based upon the confidence parameter.

7 . The method of claim 6 , wherein the measurements include at least one of yaw rate data, velocity data, sideslip angle data, front axle wheel speed data, or rear axle wheel speed data.

8 . The method of claim 7 , wherein one or more parameters of the ExpTanh parameters is passed through a function to ensure the one or more parameters of the ExpTanh parameters have a nonnegative value.

9 . A system comprising: a processor; and a memory storing instructions that, when executed by the processor, cause the processor to: operate a vehicle with a control framework incorporating an ExpTanh function; calculate prior slip data based on measurements obtained by the control framework; use a first Neural Network predictive model to determine ExpTanh parameters based upon the prior slip data and the measurements; and input a slip parameter and the ExpTanh parameters into the ExpTanh function to estimate a tire force, wherein the ExpTanh function has multiple inflection points based on the ExpTanh parameters; and adjust a vehicle control input based on the tire force estimate.

10 . The system of claim 9 , wherein the instruction to use the first predictive model to determine the ExpTanh parameters is further based on scaling parameters; and the system further includes instructions to: use, in parallel with the first Neural Network predictive model, a second Neural Network predictive model to determine the scaling parameters based on the prior slip data.

11 . The system of claim 10 , wherein the system further includes instructions to: adjust the tire force by the scaling parameters to estimate a pair of lateral and longitudinal tire forces.

12 . The system of claim 11 , wherein the use of the first or second Neural Network predictive models includes the use of a stochastic gradient descent algorithm to obtain the ExpTanh parameters.

13 . The system of claim 12 , wherein the prior slip data contains slip angle data and longitudinal slip ratio data; and the slip parameter is a current total slip.

14 . The system of claim 9 , wherein the system further includes the instruction to select a confidence parameter and wherein the instruction to use the first Neural Network predictive model to determine the ExpTanh parameters is further based upon the confidence parameter.

15 . The system of claim 14 , wherein the measurements include at least one of yaw rate data, velocity data, sideslip angle data, front axle wheel speed data, or rear axle wheel speed data.

16 . The system of claim 15 , wherein the system further includes instructions to pass one or more parameters of the ExpTanh parameters through a function to ensure the one or more parameters of the ExpTanh parameters have a nonnegative value.

17 . A non-transitory computer-readable medium including instructions that when executed by one or more processors cause the one or more processors to: operate a vehicle with a control framework incorporating an ExpTanh function; calculate prior slip data based on measurements obtained by the control framework; use a first Neural Network predictive model to determine ExpTanh parameters based upon the prior slip data and the measurements; and input a slip parameter and the ExpTanh parameters into the ExpTanh function to estimate a tire force, wherein the ExpTanh function has multiple inflection points based on the ExpTanh parameters; and adjust a vehicle control input based on the tire force estimate.

18 . The non-transitory computer-readable medium of claim 17 , wherein the instructions further include to apply scaling parameters to estimate a lateral tire force estimate or a longitudinal tire force estimate from the tire force estimate.

19 . The non-transitory computer-readable medium of claim 18 , wherein the instructions further include to adjust the vehicle control input based on the lateral tire force estimate, or the longitudinal tire force estimate.

Assignments (2)

Continuity (1)

References Cited (70)