MACHINE LEARNING NAVIGATIONAL ENGINE WITH IMPOSED CONSTRAINTS

• US 20180032082A1
• Filed: 09/12/2017
• Published: 02/01/2018
• Est. Priority Date: 01/05/2016
• Status: Active Grant

• Current Assignee: MobilEye Vision Technologies Ltd. (Intel Corporation)
• Original Assignee: MobilEye Vision Technologies Ltd. (Intel Corporation)
• Inventors: Shalev-Shwartz, Shai, Shashua, Amnon, Shammah, Shaked
• Granted Patent: US 9,977,430 B2
• Time in Patent Office: Days
• CPC Class Codes: B60W 2050/0088 Adaptive recalibration
• CPC Class Codes: B60W 2420/403 Image sensing, e.g. optical camera
• CPC Class Codes: B60W 2520/16 Pitch
• CPC Class Codes: B60W 2520/18 Roll
• CPC Class Codes: B60W 2552/15 Road slope, i.e. the inclination of a road segment in the longitudinal direction
• CPC Class Codes: B60W 2552/30 Road curve radius
• CPC Class Codes: B60W 2552/53 Road markings, e.g. lane marker or crosswalk
• CPC Class Codes: B60W 2554/00 Input parameters relating to objects
• CPC Class Codes: B60W 2554/20 Static objects
• CPC Class Codes: B60W 2554/4029 Pedestrians
• CPC Class Codes: B60W 2554/4046 Behavior, e.g. aggressive or erratic
• CPC Class Codes: B60W 2554/802 Longitudinal distance
• CPC Class Codes: B60W 2554/804 Relative longitudinal speed
• CPC Class Codes: B60W 30/09 Taking automatic action to avoid collision, e.g. braking and steering
• CPC Class Codes: B60W 30/0953 the prediction being responsive to vehicle dynamic parameters
• CPC Class Codes: B60W 30/0956 the prediction being responsive to traffic or environmental parameters
• CPC Class Codes: B60W 30/18163 Lane change; Overtaking manoeuvres
• CPC Class Codes: B60W 50/0097 Predicting future conditions
• CPC Class Codes: B60W 50/045 Monitoring control system parameters
• CPC Class Codes: B60W 60/0011 involving control alternatives for a single driving scenario, e.g. planning several paths to avoid obstacles
• CPC Class Codes: B60W 60/00276 : for two or more other traffic participants
• CPC Class Codes: G01C 21/34 : Route searching; Route guidance
• CPC Class Codes: G01C 21/3453 : Special cost functions, i.e. other than distance or default speed limit of road segments
• CPC Class Codes: G01C 21/3602 : Input other than that of destination using image analysis, e.g. detection of road signs, lanes, buildings, real preceding vehicles using a camera
• CPC Class Codes: G05D 1/0055 : with safety arrangements
• CPC Class Codes: G05D 1/0231 : using optical position detecting means position-fixing by using electromagnetic waves other than radio waves, e.g. optical position detecting means G01S5/16
• CPC Class Codes: G05D 1/0246 : using a video camera in combination with image processing means
• CPC Class Codes: G05D 1/0253 : extracting relative motion information from a plurality of images taken successively, e.g. visual odometry, optical flow determining position or orientation from images G06T7/70
• CPC Class Codes: G06N 20/00 : Machine learning
• CPC Class Codes: G06N 3/00 : Computing arrangements based on biological models
• CPC Class Codes: G06N 3/006 : based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
• CPC Class Codes: G06N 3/044 : Recurrent networks, e.g. Hopfield networks
• CPC Class Codes: G06N 3/08 : Learning methods
• CPC Class Codes: G06N 5/046 : Forward inferencing; Production systems
• CPC Class Codes: G06N 7/01 : Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06V 20/56 : exterior to a vehicle by using sensors mounted on the vehicle
• CPC Class Codes: G06V 20/58 : Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
• CPC Class Codes: G06V 20/584 : of vehicle lights or traffic lights
• CPC Class Codes: G06V 20/588 : Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road

Abstract

Systems and methods are provided for navigating an autonomous vehicle using reinforcement learning techniques. In one implementation, a navigation system for a host vehicle may include at least one processing device programmed to: receive, from a camera, a plurality of images representative of an environment of the host vehicle; analyze the plurality of images to identify a navigational state associated with the host vehicle; provide the navigational state to a trained navigational system; receive, from the trained navigational system, a desired navigational action for execution by the host vehicle in response to the identified navigational state; analyze the desired navigational action relative to one or more predefined navigational constraints; determine an actual navigational action for the host vehicle, wherein the actual navigational action includes at least one modification of the desired navigational action determined based on the one or more predefined navigational constraints; and cause at least one adjustment of a navigational actuator of the host vehicle in response to the determined actual navigational action for the host vehicle.

Claims

1. A navigation system for a host vehicle, the system comprising:

at least one processing device programmed to;

receive, from a camera, a plurality of images representative of an environment of the host vehicle;

analyze the plurality of images to identify a navigational state associated with the host vehicle;

provide the navigational state to a trained navigational system;

receive, from the trained navigational system, a desired navigational action for execution by the host vehicle in response to the identified navigational state;

analyze the desired navigational action relative to one or more predefined navigational constraints;

determine an actual navigational action for the host vehicle, wherein the actual navigational action includes at least one modification of the desired navigational action determined based on the one or more predefined navigational constraints; and

cause at least one adjustment of a navigational actuator of the host vehicle in response to the determined actual navigational action for the host vehicle.

2. The navigation system of claim 1, wherein the trained navigational network is trained, through reinforcement learning, to account for the predefined navigational constraints in order to provide the desired navigational action.

3. The navigation system of claim 1, wherein the one or more predefined navigational constraints include a pedestrian envelope.

4. The navigation system of claim 3, wherein the pedestrian envelope defines a buffer zone, within which navigation of the host vehicle is prohibited, and at least a portion of the buffer zone extends a predetermined distance from a detected pedestrian.

5. The navigation system of claim 4, wherein the predetermined distance is at least one meter.

6. The navigation system of claim 1, wherein the one or more predefined navigational constraints include a target vehicle envelope.

7. The navigation system of claim 6, wherein the target vehicle envelope defines a buffer zone, within which navigation of the host vehicle is prohibited, and at least a portion of the buffer zone extends a predetermined distance from a detected target vehicle.

9. The navigation system of claim 1, wherein the one or more predefined navigational constraints include a stationary object envelope.

10. The navigation system of claim 9, wherein the stationary object envelope defines a buffer zone, within which navigation of the host vehicle is prohibited, and at least a portion of the buffer zone extends a predetermined distance from a detected stationary object.

12. The navigation system of claim 10, wherein the detected stationary object includes at least one of a tree, a pole, a road sign, or an object in a roadway.

13. The navigation system of claim 1, wherein the one or more predefined navigational constraints include a road barrier envelope.

14. The navigation system of claim 13, wherein the road barrier envelope defines a buffer zone, within which navigation of the host vehicle is prohibited, and the buffer zone extends a predetermined distance from a surface of a detected road barrier.

16. The navigation system of claim 14, wherein the road barrier includes a guard rail or a roadside barrier.

17. The navigation system of claim 1, wherein the one or more predefined navigational constraints vary based on a detected motion associated with an object identified in the plurality of images.

18. The navigation system of claim 1, wherein the predefined navigational constraint includes a maximum speed of travel when passing within an influence zone of a detected pedestrian.

19. The navigation system of claim 18, wherein the influence zone is within a predetermined distance of the detected pedestrian.

21. The navigation system of claim 19, wherein the influence zone is multi-staged, each stage being associated with a different maximum speed of travel.

22. The navigation system of claim 21, wherein a first stage extends to a distance of at least 10 meters.

23. The navigation system of claim 21, wherein a second stage extends to a distance of at least 20 meters and is associated with a maximum speed of travel greater than a maximum speed of travel associated with the first stage.

24. The navigation system of claim 1, wherein the one or more predefined navigational constraints include a maximum deceleration rate.

25. The navigation system of claim 24, wherein the maximum deceleration rate is determined based on a detected distance to a target vehicle following the host vehicle.

26. The navigation system of claim 1, wherein the one or more predefined navigational constraints include a mandatory stop at a sensed crosswalk or a railroad crossing.

27. The navigation system of claim 1, wherein the one or more predefined navigational constraints include a requirement that an exit time of the host vehicle from point at a projected intersection between trajectories of the host vehicle with at least one target vehicle be earlier than an arrival time of the at least one target vehicle.

28. The navigation system of claim 27, wherein the exit time of the host vehicle and the arrival time of the at least one target vehicle are separated by a predetermined time interval.

29. The navigation system of claim 28, wherein the predetermined time interval is at least 0.5 seconds.

30. The navigation system of claim 1, wherein the one or more predefined navigational constraints include a requirement that an arrival time of the host vehicle to a point at a projected intersection between trajectories of the host vehicle with at least one target vehicle be later than an exit time of the at least one target vehicle.

31. The navigation system of claim 30, wherein the arrival time of the host vehicle and the exit time of the at least one target vehicle are separated by a predetermined time interval.

33. The navigation system of claim 1, wherein the navigational actuator includes at least one of a steering mechanism, a brake, or an accelerator.

8. (canceled)

11. (canceled)

15. (canceled)

20. (canceled)

32. (canceled)

34. An autonomous vehicle, the autonomous vehicle comprising:

a frame;

a body attached to the frame;

a camera; and

a navigation system, the navigation system comprising;

at least one processing device programmed to;

receive, from the camera, a plurality of images representative of an environment of the vehicle;

analyze the plurality of images to identify a navigational state associated with the vehicle;

provide the navigational state to a trained navigational system;

receive, from the trained navigational system, a desired navigational action for execution by the vehicle in response to the identified navigational state;

analyze the desired navigational action relative to one or more predefined navigational constraints;

determine an actual navigational action for the vehicle, wherein the actual navigational action includes at least one modification of the desired navigational action determined based on the one or more predefined navigational constraints; and

cause at least one adjustment of a navigational actuator of the vehicle in response to the determined actual navigational action for the host vehicle.

35. A method of navigating an autonomous vehicle, the method comprising:

receiving, from a camera associated with the autonomous vehicle, a plurality of images representative of an environment of the autonomous vehicle;

analyzing the plurality of images to identify a navigational state associated with the autonomous vehicle;

providing the navigational state to a trained navigational system;

receiving, from the trained navigational system, a desired navigational action for execution by the autonomous vehicle in response to the identified navigational state;

analyzing the desired navigational action relative to one or more predefined navigational constraints;

determining an actual navigational action for the autonomous vehicle, wherein the actual navigational action includes at least one modification of the desired navigational action determined based on the one or more predefined navigational constraints; and

causing at least one adjustment of a navigational actuator of the autonomous vehicle in response to the determined actual navigational action for the host vehicle.

36-135. -135. (canceled)