Method and apparatus for monitoring unmanned ground vehicle

12103,545

17/535833

November 26, 2021

According to an embodiment, provided are a method and an apparatus for monitoring an unmanned ground vehicle (UGV), for detecting a failure of a UGV actuator in consideration of terrain information. Accordingly, the accuracy of detecting a failure of the UGV is improved.

DAEGU GYEONGBUK INSTITUTE OF SCIENCE AND TECHNOLOGY (Daegu, KR)

DAEGU GYEONGBUK INSTITUTE OF SCIENCE AND TECHNOLOGY (Daegu, KR)

1. A method for monitoring an unmanned ground vehicle (UGV) having a plurality of actuators, the method comprising: obtaining a sensor signal for a UGV state from at least one sensor disposed in the UGV; determining a friction information estimate of a ground surface at a current position of the UGV, wherein the friction information estimate is estimated by a smoothness of the ground surface from an image of an image sensor of the UGV; determining a sensor signal estimate for the UGV state by an unknown input observer designed on the basis of the friction information estimate; detecting whether there is a failure in each of the plurality of actuators of the UGV on the basis of a residue between the sensor signal and the sensor signal estimate; and sounding an alarm for an actuator where there is the failure among the plurality of actuators, wherein the determining of the sensor signal estimate comprises executing the unknown input observer designed on the basis of a UGV rated model based on the friction information estimate, and wherein the UGV rated model is a dynamic model based on a system matrix for the UGV state and a failure matrix corresponding to a failure signal of the actuator of the UGV.

2. The method of claim 1 , wherein the sensor signal corresponds to a vector comprising a global position, a yaw angle, and a yaw rate of the UGV.

3. The method of claim 1 , wherein the system matrix is a constant matrix comprising a constant defined on the basis of the friction information estimate and a mass of the UGV.

4. The method of claim 1 , wherein the failure matrix corresponds to a time varying matrix comprising a time varying parameter defined on the basis of a yaw angle of the UGV.

5. The method of claim 1 , wherein the unknown input observer comprises a plurality of unknown input observers associated one-to-one with the plurality of actuators of the UGV, and wherein the determining of the estimate for the sensor signal comprises determining a sensor signal estimate of each unknown input observer by executing the plurality of unknown input observers.

6. The method of claim 1 , wherein the unknown input observer comprises a plurality of unknown input observers associated one-to-one with the plurality of actuators of the UGV, and wherein the detecting of the failure comprises: determining a residue between the sensor signal estimate of each unknown input observer of the plurality of unknown input observers and the sensor signal; and determining whether there is the failure in the actuator associated with each unknown input observer on the basis of the residue.

7. The method of claim 6 , wherein the determining of whether there is a failure comprises determining that an actuator corresponding to an unknown input observer of which the residue exceeds a predetermined threshold is in a failure state.

8. An apparatus for monitoring a UGV having a plurality of actuators, the apparatus comprising: a memory configured to store at least one instruction; and a processor, wherein the processor is configured to, by executing the at least one instruction: obtain a sensor signal for a UGV state from at least one sensor disposed in the UGV; determine a friction information estimate of a ground surface at a current position of the UGV, wherein the friction information estimate is estimated by a smoothness of the ground surface from an image of an image sensor of the UGV; determine a sensor signal estimate for the UGV state by an unknown input observer designed on the basis of the friction information estimate; detect whether there is a failure in each of the plurality of actuators of the UGV on the basis of a residue between the sensor signal and the sensor signal estimate; and sound an alarm for an actuator where there is the failure among the plurality of actuators, wherein the processor is configured to execute the unknown input observer designed on the basis of a UGV rated model based on the friction information estimate, and wherein the UGV rated model is a dynamic model based on a system matrix for the UGV state and a failure matrix corresponding to a failure signal of the actuator of the UGV.

9. The apparatus of claim 8 , wherein the sensor signal corresponds to a vector comprising a global position, a yaw angle, and a yaw rate of the UGV.

10. The apparatus of claim 8 , wherein the system matrix is a constant matrix comprising a constant defined on the basis of the friction information estimate and a mass of the UGV.

11. The apparatus of claim 8 , wherein the failure matrix corresponds to a time varying matrix comprising a time varying parameter defined on the basis of a yaw angle of the UGV.

12. The apparatus of claim 8 , wherein the unknown input observer comprises a plurality of unknown input observers associated one-to-one with the plurality of actuators of the UGV, and wherein the processor is configured to determine a sensor signal estimate of each unknown input observer by executing the plurality of unknown input observers.

13. The apparatus of claim 8 , wherein the unknown input observer comprises a plurality of unknown input observers associated one-to-one with the plurality of actuators of the UGV, and wherein, in order to detect the failure, the processor is configured to: determine a residue between the sensor signal estimate of each unknown input observer of the plurality of unknown input observers and the sensor signal; and determine whether there is the failure in the actuator associated with each unknown input observer on the basis of the residue.

14. The apparatus of claim 13 , wherein, in order to determine whether there is a failure, the processor is configured to determine that an actuator corresponding to an unknown input observer of which the residue exceeds a predetermined threshold is in a failure state.

20190359252 November 2019 Yu
20210188252 June 2021 Lu
20220111895 April 2022 Schumann
20220144257 May 2022 Maeda
20220281456 September 2022 Giovanardi
20220410913 December 2022 Na
20220410914 December 2022 Na
20230182745 June 2023 Misu
2001-264132 September 2001
2016-113050 June 2016
10-2010-0071554 June 2010
10-2019-0045488 May 2019
10-2020-0140449 December 2020

Samia Mellah et al., “4-mecanum wheeled mobile robot actuator fault detection & isolation using unknown input observerbased approach”, European Control Conference ECC, May 2020, hal-02454947v1. cited by applicant
KIPO, Office Action of KR 10-2021-0083952 dated Jun. 10, 2022. cited by applicant
Juqi Hu et al., “Tire-Road Friction Coefficient Estimation under Constant Vehicle Speed Control,” IFAC-PapersOnLine, vol. 52, No. 8, pp. 136-141, 2019, doi: 10.1016/j.ifacol.2019.08.061. cited by applicant
KIPO, Notice of Allowance of KR 10-2021-0083952 dated Nov. 22, 2022. cited by applicant

edspgr.12103545