Autonomous driving apparatus and method

1. An autonomous driving apparatus for an ego vehicle that autonomously travels, the autonomous driving apparatus comprising:

a first sensor to detect a vehicle nearby the ego vehicle;

a memory to store map information; and

a processor to control autonomous driving of the ego vehicle based on the nearby vehicle detected by the first sensor and the map information stored in the memory,

wherein the processor is configured to:

generate an actual driving trajectory of the nearby vehicle based on driving information of the nearby vehicle detected by the first sensor and generate an expected driving trajectory of the nearby vehicle based on the map information stored in the memory;

perform a first reliability diagnosis of autonomous driving control over the ego vehicle by comparing the generated actual driving trajectory and the generated expected driving trajectory;

perform a second reliability diagnosis of the autonomous driving control over the ego vehicle by applying, to a predefined classification model, a first sensing value obtained through the first sensor, the map information stored in the memory, a second sensing value obtained through a second sensor mounted on the nearby vehicle, and map information applied to the nearby vehicle, and

maintain or stop the autonomous driving control over the ego vehicle based on results of the first and second reliability diagnoses,

wherein the processor is configured to perform the first reliability diagnosis based on an amount of a trajectory error between the actual driving trajectory and expected driving trajectory of the nearby vehicle or a cumulative addition of the trajectory errors,

wherein the processor is configured to determine that the autonomous driving control over the ego vehicle is unreliable, when the amount of the trajectory error reaches or exceeds a preset first threshold value within a preset first critical time, and

wherein the processor is further configured to additionally perform the first reliability diagnosis based on the cumulative addition of the trajectory errors when the amount of the trajectory error is less than the first threshold value for the first critical time.

2. The autonomous driving apparatus of claim 1 , wherein the processor is further configured to determine that the autonomous driving control over the ego vehicle is unreliable, when a cumulative addition of accumulated and added trajectory errors reaches or exceeds a preset second threshold value within a second critical time, the second critical time being preset as a value greater than the first critical time, when the amount of the trajectory error is less than the first threshold value for the first critical time.

3. The autonomous driving apparatus of claim 1 , wherein the predefined classification model used in the second reliability diagnosis is defined to diagnose reliability of the autonomous driving control over the ego vehicle by taking into consideration a first deviation between the first and second sensing values for an identical reference object, and a second deviation between first and second location values of the identical reference object, the first location value being present in the map information stored in the memory and the second location value being present in the map information applied to the nearby vehicle.

4. The autonomous driving apparatus of claim 1 , wherein the processor is configured to:

maintain the autonomous driving control over the ego vehicle when both the first and second reliability diagnoses are satisfied, and

stop the autonomous driving control over the ego vehicle when at least one of the first and second reliability diagnoses are not satisfied.

5. The autonomous driving apparatus of claim 1 , wherein the nearby vehicle comprises at least two surrounding vehicles.

6. An autonomous driving method for an ego vehicle, the autonomous driving method comprising the steps of:

controlling autonomous driving of the ego vehicle based on a nearby vehicle detected by a first sensor and map information stored in a memory;

generating an actual driving trajectory of the nearby vehicle based on driving information of the nearby vehicle detected by the first sensor and generating an expected driving trajectory of the nearby vehicle based on the map information stored in the memory,

performing a first reliability diagnosis of autonomous driving control over the ego vehicle by comparing the generated actual driving trajectory and the generated expected driving trajectory,

performing a second reliability diagnosis of the autonomous driving control over the ego vehicle by applying, to a predefined classification model, a first sensing value obtained through the first sensor, the map information stored in the memory, a second sensing value obtained through a second sensor mounted on the nearby vehicle, and map information applied to the nearby vehicle, and

maintaining or stopping the autonomous driving control over the ego vehicle based on results of the first and second reliability diagnoses

wherein the step of performing of the first reliability diagnosis comprises the steps of:

performing a first reliability diagnosis based on an amount of a trajectory error between the actual driving trajectory and expected driving trajectory of the nearby vehicle or a cumulative addition of the trajectory errors,

determining that the autonomous driving control over the ego vehicle is unreliable, when the amount of the trajectory error reaches or exceeds a preset first threshold value within a preset first critical time, and

performing additionally the first reliability diagnosis based on the cumulative addition of the trajectory errors when the amount of the trajectory error is less than the first threshold value for the first critical time.

7. The autonomous driving method of claim 6 , wherein the step of performing the first reliability diagnosis further comprises the step of determining that the autonomous driving control over the ego vehicle is unreliable, when a cumulative addition of accumulated and added trajectory errors reaches or exceeds a preset second threshold value within a second critical time, the second critical time being preset as a value greater than the first critical time, when the amount of the trajectory error is less than the first threshold value for the first critical time.

8. The autonomous driving method of claim 6 , wherein the predefined classification model used in the second reliability diagnosis is defined to diagnose reliability of the autonomous driving control over the ego vehicle by taking into consideration a first deviation between the first and second sensing values for an identical reference object, and a second deviation between first and second location values of the identical reference object, the first location value being present in the map information stored in the memory and the second location value being present in the map information applied to the nearby vehicle.

9. The autonomous driving method of claim 6 , wherein the step of maintaining or stopping the autonomous driving control comprises the steps of maintaining the autonomous driving control over the ego vehicle when both the first and second reliability diagnoses are satisfied, and stopping the autonomous driving control over the ego vehicle when at least one of the first and second reliability diagnoses are not satisfied.