Object avoidance methods for automated aerial vehicles

Patent metadata

Field	Value
Publication number	US-10762797-B2
Application number	US-201715726330-A
Country	US
Kind code	B2
Filing date	Oct 5, 2017
Priority date	Jun 25, 2014
Publication date	Sep 1, 2020
Grant date	Sep 1, 2020

Abstract

This disclosure describes an automated aerial vehicle that includes one or more object detection elements configured to detect the presence of objects and an avoidance determining element configured to cause the automated aerial vehicle to automatically determine and execute an avoidance maneuver to avoid the objects. For example, an object may be detected and an avoidance maneuver determined based on a position of the object and an object vector representative of a direction and a magnitude of velocity of the object.

First claim

What is claimed is:

1. A computer implemented method to alter a path of an automated aerial vehicle, comprising:
detecting a presence of an object;
determining an object vector representative of a distance and a direction of the object relative to the automated aerial vehicle;
determining a pushing function representative of the distance between the detected object and the automated aerial vehicle;
determining, based at least in part on the object vector and the pushing function, a retreat vector;
determining, based at least in part in on the object vector, a plurality of available defensive directions, wherein each of the plurality of available defensive directions corresponds to a direction that is different than the direction of the object vector;
randomly selecting an available defensive direction from the plurality of available defensive directions;
generating an avoidance vector based at least in part on the retreat vector and the randomly selected defensive direction, wherein the avoidance vector is a sum of the retreat vector and the randomly selected defensive direction; and
executing the avoidance vector to avoid the object.

2. The computer implemented method of claim 1 , wherein generating the avoidance vector is based at least in part on a magnitude of the retreat vector and the randomly selected defensive direction.

3. The computer implemented method of claim 1 , wherein the pushing function is based at least in part on an inverse of the distance between the object and the automated aerial vehicle.

4. The computer implemented method of claim 1 , wherein:
the object vector and the retreat vector are in a first plane;
the defensive direction is substantially perpendicular to the first plane; and
the avoidance vector is not in the first plane.

5. The computer implemented method of claim 1 , wherein the defensive direction is selected based at least in part on an object type or a location of the automated aerial vehicle.

6. An automated aerial vehicle comprising:
a body;
a plurality of motors;
an object detection element configured to detect a presence of an object within a defined distance of the body;
a memory coupled to one or more processors and storing program instructions that when executed by the one or more processors cause the one or more processors to at least:
determine an object vector representative of a detected object;
determine, based at least in part on the object vector, a retreat vector representative of a first direction and a first magnitude for navigating the automated aerial vehicle to avoid the detected object;
determine, based at least in part on the object vector, a defensive direction representative of a second direction for navigating the automated aerial vehicle to avoid the detected object;
generate an avoidance vector based at least in part on the retreat vector and the defensive direction, wherein the avoidance vector is a sum of the retreat vector and the defensive direction; and
provide instructions to navigate the automated aerial vehicle according to the avoidance vector.

7. The automated aerial vehicle of claim 6 , wherein the object vector representative of the detected object includes a direction of the detected object relative to the automated aerial vehicle and a distance of the detected object relative to the automated aerial vehicle.

8. The automated aerial vehicle of claim 6 , wherein:
the object vector is in the first direction, and
the defensive direction is in the second direction that intersects the first direction.

9. The automated aerial vehicle of claim 8 , wherein the avoidance vector is in a third direction that intersects each of the first direction and the second direction.

10. The automated aerial vehicle of claim 6 , wherein the defensive direction is randomly selected from a plurality of directions that are different than a direction of the detected object.

11. The automated aerial vehicle of claim 6 , wherein:
the object detection element is further configured to detect a presence of a second object within the defined distance of the body; and
the avoidance vector is further generated based at least in part on a time to impact between the second detected object and the automated aerial vehicle.

12. The automated aerial vehicle of claim 11 , wherein the avoidance vector is further generated based at least in part on a sum of the object vector representative of a distance and a direction of the detected object relative to the automated aerial vehicle and a second object vector representative of a second distance and a second direction of the second detected object relative to the automated aerial vehicle.

13. The automated aerial vehicle of claim 12 , wherein the avoidance vector is further generated based at least in part on a first pushing function determined for the detected object and a second pushing function determined for the second detected object.

14. The automated aerial vehicle of claim 13 , wherein the first pushing function is based at least in part on the distance between the detected object and the automated aerial vehicle or a time to impact between the detected object and the automated aerial vehicle, and the second pushing function is based at least in part on the second distance between the second detected object and the automated aerial vehicle or a time to impact between the second detected object and the automated aerial vehicle.

15. The automated aerial vehicle of claim 12 , wherein the object vector and the second object vector lie in a first plane with respect to the automated aerial vehicle and the defensive direction lies in a second plane that is different than the first plane.

16. The automated aerial vehicle of claim 6 , wherein the instructions when executed by the one or more processors further cause the one or more processors to at least:
transmit, via a communication element of the automated aerial vehicle, at least one of a location of the automated aerial vehicle when the object is detected, a location of the detected object, an object type of the detected object, the object vector of the detected object, the defensive direction, or the avoidance vector.

17. A computer implemented method to avoid an object by an automated aerial vehicle, comprising:
determining an object vector of a detected object;
determining, based at least in part on the object vector, a retreat vector including a first direction to avoid the detected object;
determining, based at least in part on the object vector, a defensive direction including a second direction to avoid the detected object;
generating an avoidance vector including a third direction based at least in part on the retreat vector and the defensive direction, wherein the avoidance vector is a sum of the retreat vector and the defensive direction; and
executing the avoidance vector to avoid the detected object.

18. The computer implemented method of claim 17 , wherein:
the second direction of the defensive direction intersects the first direction of the retreat vector; and
the third direction of the avoidance vector intersects each of the first direction and the second direction.

19. The computer implemented method of claim 17 , wherein the defensive direction is selected from a plurality of directions that are different than the first direction; and
wherein the plurality of directions are determined based at least in part on at least one of a direction of movement of the detected object, an object type of the detected object, or a location of the automated

Assignees

• Amazon Tech Inc

Inventors

• Navot Amir
• Kimchi Gur
• Beckman Brian C
• Schaffalitzky Frederik
• Buchmueller Daniel
• Anderson Robert John

Classifications

• B64U2201/104

  using satellite radio beacon positioning systems, e.g. GPS · CPC title

• G08G5/723

  from the aircraft · CPC title

• G08G5/57

  for unmanned aircraft · CPC title

• G08G5/55

  for a single aircraft · CPC title

• G08G5/21

  located onboard the aircraft · CPC title