Method and system for decomposing cross-domain path planning of amphibious vehicle

Patent metadata

Field	Value
Publication number	US-2024111288-A1
Application number	US-202318096588-A
Country	US
Kind code	A1
Filing date	Jan 13, 2023
Priority date	Sep 30, 2022
Publication date	Apr 4, 2024
Grant date	—

Abstract

A method and system for decomposing cross-domain path planning of an amphibious vehicle. The method includes: defining a starting point and a target point of cross-domain path planning by acquiring an amphibious map containing coastline information, constructing a candidate set of land-water transition points; setting a multilevel gradient search range on a coastline and a set number of iterations; determining a locally-optimal land-water transition point in each search range in a level-wise search mode; setting a heuristic search termination condition, and determining, based on the heuristic search termination condition, whether a current locally-optimal land-water transition point can be globally optimal; if yes, stopping searching, and determining a globally-optimal land-water transition point; if no, determining whether the range is completely traversed; if yes, stopping searching, and determining a globally-optimal land-water transition point; if no, continuing to traverse the range; finally performing global path planning based on the globally-optimal land-water transition point.

First claim

What is claimed is:

1 . A method for decomposing cross-domain path planning of an amphibious vehicle, comprising:
acquiring an amphibious map containing coastline information, determining a starting point and a target point of a cross-domain path planning task of the amphibious vehicle based on the amphibious map, and constructing a candidate set of land-water transition points, wherein the coastline information comprises several pieces of feasible land-water transition region position information; setting a multilevel gradient search range on a coastline, and setting a set number of iterations for each search range; searching from each search range based on the set number of iterations in a level-wise search mode, determining a locally-optimal land-water transition point in each search range, and adding the locally-optimal land-water transition point into the candidate set of land-water transition points; setting a heuristic search termination condition, and determining, based on the heuristic search termination condition, whether there is a possibility that a current locally-optimal land-water transition point is globally optimal; if yes, stopping searching, and selecting a globally-optimal land-water transition point from the candidate set of land-water transition points; or if no, determining whether the multilevel gradient search range is completely traversed; if yes, stopping searching, and selecting a globally-optimal land-water transition point from the candidate set of land-water transition points; or if no, continuing to traverse the multilevel gradient search range until the globally-optimal land-water transition point is determined; and redetermining the starting point and the target point of the cross-domain path planning task based on the globally-optimal land-water transition point, and performing global path planning.

2 . The method for decomposing cross-domain path planning of an amphibious vehicle according to claim 1 , wherein the starting point and the target point are located in different geographical regions; and the geographical regions comprise land and water areas.

3 . The method for decomposing cross-domain path planning of an amphibious vehicle according to claim 1 , wherein the setting a multilevel gradient search range on a coastline, and setting a set number of iterations for each search range specifically comprises:
selecting an intersection of a line connecting the starting point to the target point and the coastline as an intermediate node; extending leftward and rightward for a set distance along the coastline with the intermediate node as a center, determining a left first-level node and a right first-level node, and taking a range formed by the left first-level node and the right first-level node on the coastline as a first-level search range; continuing to extend leftward from the left first-level node and extend rightward from the right first-level node to obtain a left second-level node and a right second-level node, and taking a range formed by the left first-level node and the left second-level node on the coastline and a range formed by the right first-level node and the right second-level node on the coastline as a two-level search range; continuing to determine the search range until the entire coastline between the starting point and the target point is traversed, to complete setting of the multilevel gradient search range, wherein when the search range is determined, if a coastline end point is reached in advance during extension in one of the leftward direction and the rightward direction, range length compensation is performed in the opposite direction to ensure that a length of each level of search range is kept the same; when the next level of search range is determined, unidirectional extension is performed in a direction opposite to the original extension direction, and an extended distance is twice an extended distance in the original extension direction when the previous level of search range is determined; and allocating, after the multilevel gradient search range is obtained, the number of range iterations based on a size of a feasible land-water transition region comprised in each range.

4 . The method for decomposing cross-domain path planning of an amphibious vehicle according to claim 1 , wherein the determining a locally-optimal land-water transition point in each search range specifically comprises:
randomly sampling in a feasible land-water transition region comprised in a current search range, and determining a total cost value of a current land-water transition point by using formula f=h water +h land +h transition for land-water transition points obtained by sampling, wherein f is the total cost value of the current land-water transition point, h water is a cost value of reaching a starting point or a target point on water from the current land-water transition point, h land is a cost value of reaching a starting point or a target point on land from the current land-water transition point, and h transition is a cost value of the current land-water transition point; and selecting, within a range of the set number of iterations, a land-water transition point with a minimum total cost value as the locally-optimal land-water transition point in the current search range.

5 . The method for decomposing cross-domain path planning of an amphibious vehicle according to claim 4 , wherein the setting a heuristic search termination condition, and determining, based on the heuristic search termination condition, whether there is a possibility that a current locally-optimal land-water transition point is globally optimal specifically comprises:
determining a heuristic cost value of the current search range based on a positional relationship between the starting point, the target point, and the locally-optimal land-water transition point in the current search range, wherein the heuristic cost value represents a minimum path length from the starting point to the target point through the optimal land-water transition point without considering any obstacles; determining whether the total cost value of the locally-optimal land-water transition point in the current search range and the heuristic cost value of the current search range meet a condition f<a*G wherein a is an optimality proportional coefficient, and G is the heuristic cost value; and determining, if the condition is met, that there is no locally-optimal land-water transition point in the subsequent search range that has a total cost value less than that of the current locally-optimal land-water transition point, and terminating the searching.

6 . The method for decomposing cross-domain path planning of an amphibious vehicle according to claim 1 , wherein the redetermining the starting point and the target point of the cross-domain path planning task based on the globally-optimal land-water transition point, and performing global path planning specifically comprises:
dividing the cross-domain path planning task into a land planning stage and a water planning stage based on the globally-optimal land-water transition point, wherein starting points of the land planning stage and the water planning stage are the globally-optimal land-water transition point, and target points of the land planning stage and the water planning stage are separately the starting point or the target point of the cross-domain path planning task; and performing path search based on a global path planning algorithm corresponding to the land planning stage and the water planning stage, and merging a found water path and land path into a final path of the cross-domain path planning task to complete the global path planning.

7 . A system for decomposing cross-domain path pl

Assignees

• Univ Chongqing
• Univ Shanghai

Inventors

• Pu Huayan
• Ding Lele
• Zheng Xuyang
• Luo Jun
• Ma Jie
• Huang Jing
• Chen Yongbing
• Liu Hongliang
• Shen Anming
• Sun Haonan

Classifications

• G05D1/43

  Control of position or course in two dimensions [2D] · CPC title

• B60W60/001

  Planning or execution of driving tasks · CPC title

• G05D1/0212 Primary

  with means for defining a desired trajectory (involving a plurality of land vehicles G05D1/0287) · CPC title

• G05D1/0206

  specially adapted to water vehicles · CPC title

• G05D1/0214 Primary

  in accordance with safety or protection criteria, e.g. avoiding hazardous areas (monitoring the location of vehicles within a certain area, e.g. forbidden or allowed areas, in traffic control systems for road vehicles G08G1/13) · CPC title