Trajectory correction system and method therefor
A trajectory correction method includes: determining whether there is a collision zone according to an initial path and a predicted moving path of a dynamic object; forming a first repulsive object at the corresponding collision zone of an original potential field map in response to determining that there is the collision zone, to obtain a modified potential field map; and obtaining an obstacle avoidance path according to the modified potential field map.
1 . A trajectory correction system, comprising:
a memory, for storing a program code; and
a processor, coupled to the memory, for accessing and executing the program code to:
determine whether there is a collision zone according to an initial path of a mobile device and a predicted moving path of a dynamic object, wherein the collision zone is an intersecting area between the initial path and the predicted moving path;
form a first repulsive object at a collision zone position corresponding to the collision zone in an original potential field map in response to determining that there is the collision zone and form a second repulsive object corresponding to the dynamic object that moves along the predicted moving path over time in the original potential field map, to obtain a modified potential field map, wherein a strength of a repulsive field of the second repulsive object is the same as a strength of a repulsive field of the first repulsive object at a collision time point corresponding to the collision zone;
generate an obstacle avoidance path according to the modified potential field map; and
control a driving device of the mobile device to move along the obstacle avoidance path.
2 . The trajectory correction system according to claim 1 , wherein the strength of the repulsive field of the first repulsive object or the second repulsive object corresponds to a type of the dynamic object.
3 . The trajectory correction system according to claim 1 , wherein the processor inputs the modified potential field map into a neural network to obtain the obstacle avoidance path.
4 . The trajectory correction system according to claim 3 , wherein the neural network comprises a convolutional neural network and a liquid time-constant network (LTC).
5 . The trajectory correction system according to claim 1 , wherein the initial path comprises a plurality of first coordinate parameters, the predicted moving path comprises a plurality of second coordinate parameters, and the processor determines whether there is the collision zone according to the first coordinate parameters, the second coordinate parameters, and a distance threshold.
6 . The trajectory correction system according to claim 5 , wherein each of the first coordinate parameters comprises a first coordinate and a first time point, each of the second coordinate parameters comprises a second coordinate and a second time point, and the processor obtains a distance value according to the first coordinate corresponding to the first time point and the second coordinate corresponding to the second time point, determines that there is the collision zone in response to the distance value is not greater than the distance threshold, and determines the first coordinates and the second coordinates corresponding to the distance value as the collision zone.
7 . The trajectory correction system according to claim 5 , wherein the distance threshold is a size corresponding to the dynamic object.
8 . The trajectory correction system according to claim 1 , wherein the strength of the repulsive field of the first repulsive object corresponding to the collision zone changes with time, and the strength of the repulsive field of the first repulsive object is at a maximum value at a collision time point corresponding to the collision zone.
9 . A trajectory correction method, comprising:
determining whether there is a collision zone according to an initial path of a mobile device and a predicted moving path of a dynamic object, wherein the collision zone is an intersecting area between the initial path and the predicted moving path;
forming a first repulsive object at a collision zone position corresponding to the collision zone in an original potential field map in response to determining that there is the collision zone and forming a second repulsive object corresponding to the dynamic object that moves along the predicted moving path over time in the original potential field map, to obtain a modified potential field map, wherein a strength of a repulsive field of the second repulsive object is the same as a strength of a repulsive field of the first repulsive object at a collision time point corresponding to the collision zone;
generating an obstacle avoidance path according to the modified potential field map; and
controlling a driving device of the mobile device to move along the obstacle avoidance path.
10 . The trajectory correction method according to claim 9 , wherein the strength of the repulsive field of the first repulsive object or the second repulsive object corresponds to a type of the dynamic object.
11 . The trajectory correction method according to claim 9 , wherein the modified potential field map is inputted into a neural network to obtain the obstacle avoidance path.
12 . The trajectory correction method according to claim 11 , wherein the neural network comprises a convolutional neural network and a liquid time-constant network (LTC).
13 . The trajectory correction method according to claim 9 , wherein the initial path comprises a plurality of first coordinate parameters, the predicted moving path comprises a plurality of second coordinate parameters, and the collision zone is determined according to the first coordinate parameters, the second coordinate parameters, and a distance threshold.
14 . The trajectory correction method according to claim 13 , wherein each of the first coordinate parameters comprises a first coordinate and a first time point, each of the second coordinate parameters comprises a second coordinate and a second time point, obtaining a distance value according to the first coordinate corresponding to the first time point and the second coordinate corresponding to the second time point, the collision zone is determined in response to the distance value is not greater than the distance threshold, and the first coordinates and the second coordinates corresponding to the distance value are determined as the collision zone.
15 . The trajectory correction method according to claim 13 , wherein the distance threshold is a size corresponding to the dynamic object.
16 . The trajectory correction method according to claim 9 , wherein the strength of the repulsive field of the first repulsive object corresponding to the collision zone changes with time, and the strength of the repulsive field of the first repulsive object is at a maximum value at a collision time point corresponding to the collision zone.