Apparatus and method for failure diagnosis and calibration of sensors for advanced driver assistance systems

1. An apparatus for failure diagnosis and calibration of sensors for advanced driver assistance systems (ADAS), the apparatus comprising:

a measuring unit including two or more sensors mounted in a vehicle;

a storage configured to store characteristic data of each of the two or more sensors; and

a processor configured to select one of the two or more sensors as a reference sensor and configured to determine whether or not a failure of the other sensor to be diagnosed is detected and whether calibration thereof is required, on the basis of error covariance information of the reference sensor wherein the characteristic data includes a mounting position, a measuring range, and error covariance information of the reference sensor.

2. The apparatus according to claim 1 , wherein the two or more sensors measure homogeneous data with respect to an object positioned in a common measurement area of the two or more sensors.

3. The apparatus according to claim 2 , wherein the reference sensor is any one of a global positioning system (GPS) sensor, an ultrasonic sensor, a radio detection and ranging (RADAR) sensor, a light detection and ranging (LiDAR) sensor, and a camera.

4. The apparatus according to claim 1 , wherein the processor is further configured to

select the reference sensor by analyzing reliability of the two or more sensors;

compare measured data of the reference sensor with measured data of the sensor to be diagnosed, determine whether or not the degree of error is out of an allowable error range, and determine whether or not the failure of the sensor to be diagnosed is detected and the calibration of the sensor to be diagnosed is required; and

calibrate a calibration parameter of the sensor to be diagnosed on the basis of the measured data of the reference sensor, when the calibration of the sensor to be diagnosed is determined to be required.

5. The apparatus according to claim 4 , wherein the processor is further configured to analyze the reliability of the two or more sensors by using accuracy with respect to the measured data of the two or more sensors.

6. The apparatus according to claim 4 , wherein the processor is further configured to analyze the reliability of the two or more sensors by using accuracy according to specifications of respective sensors.

7. The apparatus according to claim 4 , wherein the processor is further configured to calculate covariance intersection of the error covariance information of the reference sensor and error covariance information of the sensor to be diagnosed to calculate an allowable error range for determining whether or not the calibration of the sensor to be diagnosed is required.

8. The apparatus according to claim 4 , wherein the processor is further configured to calculate covariance union of the error covariance information of the reference sensor and error covariance information of the sensor to be diagnosed to calculate an allowable error range for determining whether or not the failure of the sensor to be diagnosed is detected.

9. The apparatus according to claim 4 , wherein the processor is further configured to calculate a Mahalanobis distance between the measured data of the reference sensor and the measured data of the sensor to be diagnosed to calculate measurement errors of the reference sensor and the sensor to be diagnosed.

10. The apparatus according to claim 4 , wherein the processor is further configured to calculate the calibration parameter that minimizes a difference between the measured data of the reference sensor and the measured data of the sensor to be diagnosed by using a cost function.

11. A method for failure diagnosis and calibration of sensors for advanced driver assistance systems (ADAS), the method comprising:

determining whether or not diagnostic conditions for initiating diagnosis of two or more sensors mounted in a vehicle are satisfied;

selecting one of the sensors as a reference sensor by analyzing a reliability of the two or more sensors;

determining whether or not failure of the other sensor to be diagnosed is detected and whether calibration thereof is required, on the basis of error covariance information included in characteristic data of the reference sensor; and

performing a process according to diagnostic results of the sensor to be diagnosed,

wherein the characteristic data includes a mounting position, a measuring range, and error covariance information of the reference sensor.

12. The method according to claim 11 , wherein, in the step of determining whether or not the diagnostic conditions are satisfied, it is determined whether or not an object is present in a common measurement area of the two or more sensors.

13. The method according to claim 11 , wherein the step of determining whether or not the failure of the sensor to be diagnosed is detected and whether the calibration thereof is required comprises:

calculating covariance intersection of the error covariance information of the reference sensor and error covariance information of the sensor to be diagnosed;

verifying whether or not the degree of error between measured data of the reference sensor and measured data of the sensor to be diagnosed is out of an allowable error range set by the covariance intersection to determine whether or not the calibration is required;

calculating covariance union of the error covariance information of the reference sensor and error covariance information of the sensor to be diagnosed, when the calibration of the sensor to be diagnosed is determined to be required;

verifying whether or not the degree of error between the measured data of the reference sensor and the measured data of the sensor to be diagnosed is out of an allowable error range set by the covariance union to determine whether or not there is probability of failure;

counting the cumulative number of failure diagnoses of the sensor to be diagnosed, when there is the probability of failure of the sensor to be diagnosed; and

verifying whether or not the number of failure diagnoses of the sensor to be diagnosed exceeds a threshold value to determine the failure of the sensor to be diagnosed.

14. The method according to claim 13 , wherein the step of determining whether or not there is the probability of failure further comprises:

counting the number of requests for calibration of the sensor to be diagnosed unless there is the probability of failure of the sensor to be diagnosed; and

verifying whether or not the number of requests for calibration of the sensor to be diagnosed exceeds a threshold value to determine that the calibration of the corresponding sensor is required.

15. The method according to claim 14 , further comprising calculating a calibration parameter that minimizes a difference between the measured data of the reference sensor and the measured data of the sensor to be diagnosed by using a cost function, when it is determined that the calibration of the corresponding sensor is required.

16. The method according to claim 13 , further comprising outputting information notifying the failure of the sensor to be diagnosed, when it is determined that the failure of the corresponding sensor is detected.