Differential budding: method and apparatus for path planning with moving obstacles and goals

• US 4,949,277 A
• Filed: 03/09/1988
• Issued: 08/14/1990
• Est. Priority Date: 03/09/1988
• Status: Expired due to Term

• Current Assignee: North American Philips Corp. (Koninklijke Philips N.V.)
• Original Assignee: North American Philips Corp. (Koninklijke Philips N.V.)
• Inventors: Trovato, Karen I., Dorst, Leendert
• Primary Examiner(s): Ruggiero, Joseph
• Application Number:
  US07/166,599
• Time in Patent Office: 888 Days
• Field of Search: 364/513, 364/148, 364/424.02, 364/474.2, 364/300, 364/200 MS File, 364/900 MS File, 901/1, 901/50
• US Class Current: 700/255
• CPC Class Codes: B25J 9/1666 Avoiding collision or forbidden zones
• CPC Class Codes: G05B 19/4061 Avoiding collision or forbidden zones
• CPC Class Codes: G05B 2219/35415 3-D three dimension, space input, spaceball
• CPC Class Codes: G05B 2219/40443 Conditional and iterative planning
• CPC Class Codes: G05B 2219/40476 Collision, planning for collision free path
• CPC Class Codes: G05B 2219/49143 Obstacle, collision avoiding control, move so that no collision occurs
• CPC Class Codes: G05D 1/0246 using a video camera in combination with image processing means
• CPC Class Codes: G06F 30/18 Network design, e.g. design based on topological or interconnect aspects of utility systems, piping, heating ventilation air conditioning [HVAC] or cabling circuit design at the physical level G06F30/39; network planning tools for wireless communication networks H04W16/18
• CPC Class Codes: Y10S 706/905 Vehicle or aerospace
• CPC Class Codes: Y10S 706/919 Designing, planning, programming, CAD, CASE

Abstract

A method is presented for path planning after changes in task space. In one embodiment, the method is applied to planning a path for a robot arm. The method identifies areas in the configuration space which are affected by the changes in task space. Cost waves can then be repropagated in these affected areas to allow for planning in N dimensions and using space variant metrics. The method is also adapted to use in the presence of phantom obstacles.

Claims

1. A method for planning a path for an object to follow in a task space in which there has been a change in conditions comprising the steps of:

a. starting from an initialized configuration space;

b. identifying a perimeter of a region in the configuration space which is effected by the change in conditions; and

c. propagating cost waves from the perimeter using a space variant metric to create updated direction errors values corresponding to the change in conditions,

wherein the method is part of a method for controlling the motion of the object, comprising the further steps of;

a. using the updated direction arrows values to find a least cost path a start point to a goal point in the task space;

b. providing at least one point on the path; and

c. controlling the object to travel to the at least one point.

2. The method of claim 1 wherein, during the starting step, the configuration space contains direction arrows which reflect propagation in response to a space variant metric and which configuration space corresponds to the task space as it existed prior to the change in conditions.

3. The method of claim 1 wherein the change in conditions comprises at least one added obstacle in the task space.

4. The method of claim 3 wherein said identifying step comprises storing a perimeter list of states in the configuration space surrounding the affected region.

5. The method of claim 3 wherein:

a. the starting step comprises starting from the initialized configuration space, which is a configuration space in which the direction arrows values are associated with states, with which cost to goal values are also associated;

b. the identifying step comprises the steps of;

i. locating a back edge of the obstacle; and

ii. storing a perimeter list of states surrounding states in the affected region, the affected region being those states for which the direction arrows values indicate a path through the added object via the back edge of the obstacle; and

iii. clearing cost to goal values and direction arrows values from the states in the affected region; and

c. the propagating step comprises budding the states in the perimeter list, wherein budding comprises;

(i) exploring all neighbors of a state in the configuration space;

(ii) improving neighbors of the state by assigning appropriate values of cost to goal and direction arrows to the state;

(iii) adding improved neighbors to a storage data structure; and

(iv) repeating (c) (i)-(c) (iii) for all states in the storage data structure.

6. The method of claim 5 wherein the identifying step further comprises the steps of:

a. second locating a front edge of the obstacle; and

b. including the front edge in the perimeter list.

7. The method of claim 5 wherein said storage step comprises storing said perimeter list in a sifting heap.

8. The method of claim 1 further comprising the step of detecting the change in conditions while following a precalculated path.

9. The method of claim 8 comprising the steps of:

a. first encountering a portion of the change in conditions;

b. then treating the encountered portion as the change in conditions;

c. then performing said identifying and repropagating steps;

d. then using the updated direction arrows to find a new least cost path;

e. repeating steps and until the new least cost path takes the change of conditions into account.

10. The method of claim 9 comprising the step of selectively branching around the treating, performing, using, and repeating steps when movement of the object is of higher priority than updating configuration space and when the change in conditions is not an obstacle which prevents following the precalculated path.

11. The method of claim 9 wherein the encountering, treating, performing, using and repeating steps are executed while following said precalculated path.

12. The method of claim 9 wherein said encountering step comprises sensing a single state corresponding to the encountered portion.

13. The method of claim 9 wherein:

a. the object has proximity and/or tactile sensors; and

b. said encountering step comprises the steps of;

i. sensing a pose of the object which corresponds to the encountered portion;

ii. determining a set of associated states which correspond to the encountered portion.

14. The method of claim 1 wherein the change in conditions comprises at least one goal having been removed from the task

15. The method of claim 14 wherein said identifying step comprises storing a perimeter list of states from the configuration space which states surround the affected region.

16. The method of claim 15 wherein said storing step comprises storing the perimeter list in a sifting heap.

17. The method of claim 14 wherein:

a. the starting step comprises starting from the configuration space in which the direction arrows values are associated with states, with which cost to goal values are also associated;

steps of;

b. the identifying step comprises the

i. storing a perimeter list of states surrounding states in the affected region, the affected region being those states for which the direction arrows values indicate a path to the removed goal; and

ii. clearing cost to goal values and direction arrows values from the affected states; and

c. the propagating step comprises budding the states in the perimeter list, wherein budding comprises;

(i) exploring all neighbors of a state in the configuration space;

(ii) improving neighbors of the state by assigning appropriate values of cost to goal and direction arrows to the state;

(iii) adding improved neighbors to a storage data structure; and

(iv) repeating (c) (i)-(c) (iii) for all states in the storage data structure.

18. The method of claim 17 wherein said storing step comprises storing said perimeter list in a sifting heap.

19. The method of claim 1 wherein the change in conditions comprises that at least one obstacle has been removed.

20. The method claim 19 wherein the propagating step comprises propagating cost values from a set of neighbor states of the removed obstacle.

21. The method claim 20 comprising the further step of storing each such neighbor state on a sifting heap.

22. The method of claim 21 wherein the propagating step comprises budding the states on the sifting heap, wherein budding comprises:

(i) exploring all neighbors of a state in the configuration space;

(ii) improving neighbors of the state by assigning appropriate values of cost to the state;

(iii) adding improved neighbors to a storage data structure; and

(iv) repeating (i)-(iii) for all states in the storage data structure.

23. The method of claim 20 comprising the further step of clearing cost values from the states corresponding to the removed obstacle.

24. The method of claim 1 wherein the change in conditions comprises that a goal has been added to the task space.

25. The method of claim 24 wherein said identifying step comprises the step of storing at least one state corresponding to the added goal.

26. The method of claim 25 wherein said storing step comprises the step of storing the at least one state in a sifting heap.

27. The method claim 26 wherein:

a. the starting step comprises starting from the configuration space in which the cost to goal values are associated with states and with which direction arrows values are also associated; and

b. the propagating step comprises budding from at least one state, wherein budding comprises;

(i) exploring all neighbors of a state in the configuration space;

(ii) improving neighbors of the state by assigning appropriate values of cost to goal and direction arrows to the state;

(iii) adding improved neighbors to a storage data structure; and

(iv) repeating (b) (i)-(b) (iii) for all states in the storage data structure.

28. The method of claim 1 wherein the configuration space has more than two dimensions.

29. A method for in which planning a path for an object to follow in a task space there has been a change in conditions comprising the steps of:

a. starting from a configuration space containing direction arrows values and corresponding to the task space as it existed prior to the change in conditions; and

b. differentially budding an area in the configuration space which is affected by the change in conditions to obtain updated cost values, wherein differentially budding comprises the steps of;

i. identifying a perimeter of a region configuration space which is affected by the change in conditions; and

d cost

ii. budding from the perimeter to create update values corresponding to the change in conditions, wherein budding comprises;

(A) exploring all neighbors of a state in the configuration space;

(B) improving neighbors of the state by assigning appropriate values of cost to goal and direction arrows to the state;

(C) adding improved neighbors to a storage data structure; and

(D) repeating (b) (ii) (A)-(b) (ii) (D) for all states in the storage data structure,

wherein the method is part of a method for controlling the motion of the object, comprising the further steps of;

a. using the updated cost values to find a least cost path from a starting point to a goal point in the task space;

b. providing at least one point on the path; and

c. controlling the object to move to the at least one point.

30. The method of claim 29 wherein:

a. the starting step comprises starting from the configuration space in which the cost values are associated with states, with which direction arrows are also associated;

b. the differential budding step comprises the steps

i. first checking any added obstacles and any removed goals;

ii. clearing the influence of added obstacles and removed goals, so that a perimeter list of the affected region appears on a heap;

iii. second checking removed obstacles and added goals, so that neighbor states of the removed obstacles and added goals appear on the heap; and

iv. budding all states on the heap.

31. A method for planning a path for an object to follow in a task space in which there has been a change in conditions and for controlling the object to follow the path comprising the steps of:

a. starting from an initialized configuration space;

b. identifying a perimeter of a region in the configuration space which is affected by the change in conditions;

c. repropagating cost waves from the perimeter to create an updated configuration space; and

d. controlling the object to follow a new path indicated by the updated configuration space.

32. The method of claim 31 wherein the configuration space has more than two dimensions.

33. A method for planning a path for a robot to follow in a task space in which there has been a change in conditions, comprising the steps of:

a. starting from an initialized configuration space;

b. identifying a perimeter of a region in the configuration space which is affected by the change in conditions; and

c. repropagating cost waves from the perimeter to create an updated configuration space

wherein

(a) the robot is a robot arm;

(b) the method is part of a method for controlling the motion of the robot arm;

(c) during the starting step, the configuration space contains direction arrows values which reflect propagation in response to a metric which minimizes motion of an end effector of the robot; and

(d) the method further comprises the step of controlling the robot arm to follow a new path indicated by the configuration space.

34. Apparatus for controlling motion of a robot, comprising:

(a) a configuration space comprising a plurality of states filled with direction arrows values and post to goal values and representing a task space of the robot according to a plurality of perimeters of the robot;

(b) means for signifying in the configuration space a change in conditions in the task space;

(c) means for identifying a perimeter of a region in the configuration space which is affected by the change in conditions;

(d) means for propagating cost waves from the perimeter using a space variant metric to create updated direction arrows values and cost to goal values corresponding to the change in conditions; and

(e) means for controlling the robot to follow a path indicated by the updated direction arrows values.

35. The apparatus of claim 34 wherein

(a) the change in conditions comprises addition of an obstacle to the task space,

(b) the means for identifying comprises

(i) means for locating a back edge of the obstacle,

(ii) means for storing a perimeter list of states surrounding states in the affected region, the affected region being those states for which the direction arrows values indicate a path through the added object via the back edge of the obstacle;

(iii) means for clearing cost to goal values and direction arrows values from states in the affected region;

(iv) means for locating a front edge of the obstacle; and

(v) means for adding the front edge of the obstacle to the perimeter list; and

(c) the propagating means buds the states in the perimeter list, wherein budding comprises;

(i) exploring all neighbors of a state in the configuration space;

(ii) improving neighbors of the state by assigning appropriate values of cost to goal and direction arrows to the state;

(iii) adding improved neighbors to a storage data structure; and

(iv) repeating (c) (i)-(c) (iii) for all states in the storage data structure.

36. The apparatus of claim 34 wherein

(a) the change in conditions comprises removal of a goal from the task space;

(b) the means for identifying comprises

(i) means for storing a perimeter list of states surrounding states in the affected region, the affected region being those states for which the direction arrows values indicate a path to the removal goal; and

(ii) means for clearing cost to goal values and direction arrows values from the affected states; and

(c) the propagating means buds the states in the perimeter list, wherein budding comprises;

(i) exploring all neighbors of a state in the configuration space;

(ii) improving neighbors of the state by assigning appropriate values of cost to goal and direction arrows to the state;

(iii) adding improved neighbors to a storage data structure; and

(iv) repeating (c) (i)-(c) (iii) for all states in the storage data structure.

37. The apparatus of claim 34 wherein

(a) the change in conditions comprises removal of an obstacle from the task space;

and

(b) the propagating means buds a set of neighbor states of the removed obstacle, wherein building comprises;

(i) exploring all neighbors of a state in the configuration space;

(ii) improving neighbors of the state by assigning appropriate values of cost to goal and direction arrows to the state;

(iii) adding improved neighbors to a storage data structure; and

(iv) repeating (b) (i)-(b) (iii) for all states in the storage data structure.

38. The apparatus of claim 34 wherein

(a) the change in conditions comprises addition of a goal to the task space;

and

(b) the repropagating means buds the state corresponding to the added goal, wherein budding comprises;

(i) exploring all neighbors of a state in the configuration space;

(ii) improving neighbors of the state by assigning appropriate values of cost to goal and direction arrows to the state;

(iii) adding improved neighbors to a storage data structure; and

(iv) repeating (b) (i)-(b) (iii) for all states in the storage data structure.

39. The apparatus of claim 34 wherein the configuration space has more than two dimensions.

40. The apparatus of claim 34 wherein the robot is a robot arm, the perimeters are joint angles of the robot arm, and the configuration space has a dimension for each joint angle.