Using topological structure for path planning in semi-structured environments
원문보기
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
G01C-021/00
G01C-021/34
출원번호
US-0471079
(2009-05-22)
등록번호
US-8392117
(2013-03-05)
발명자
/ 주소
Dolgov, Dmitri A.
Thrun, Sebastian
출원인 / 주소
Toyota Motor Engineering & Manufacturing North America, Inc.
인용정보
피인용 횟수 :
8인용 특허 :
7
초록▼
A method of creating a lane network of a semi-structured environment for a vehicle, assigning a corresponding cost function to each of a plurality of coordinates of the semi-structured environment using the lane network and a state of the vehicle, the state of the vehicle corresponding to a coordina
A method of creating a lane network of a semi-structured environment for a vehicle, assigning a corresponding cost function to each of a plurality of coordinates of the semi-structured environment using the lane network and a state of the vehicle, the state of the vehicle corresponding to a coordinate location of the vehicle and an angular orientation of the vehicle, and determining an obstacle-free path on the semi-structured environment using the cost function of each of the plurality of coordinates.
대표청구항▼
1. A method, comprising: creating a lane network of an environment for a vehicle using a processor of a computing device and storing the lane network in electronic memory of the computing device;calculating a cost function for each of a plurality of coordinates of the environment using the processor
1. A method, comprising: creating a lane network of an environment for a vehicle using a processor of a computing device and storing the lane network in electronic memory of the computing device;calculating a cost function for each of a plurality of coordinates of the environment using the processor from the lane network, a goal position of the vehicle and a state of the vehicle, the state of the vehicle including a location and an angular orientation of the vehicle; anddetermining an obstacle-free path for the vehicle from the location to the goal position by minimizing a calculated cost of the obstacle-free path, the calculated cost being a sum of the cost functions of the coordinates that the obstacle-free path intersects. 2. The method recited in claim 1, wherein the calculating includes calculating a first cost function for each of the plurality of coordinates based on a distance between the each of the plurality of coordinates and the goal position, and then modifying the first cost function with a modifying function corresponding to the lane network and the state of the vehicle. 3. The method recited in claim 2, wherein the calculated cost of the obstacle-free path is defined as ρ(s)=ρ0(s)+wG∑iD(〈xi,θi〉,G), where ρ0(s) is the first cost function created using a free-space planner, wG∑iD(〈xi,θi〉,G) is the modifying function and modifies the calculated cost if the obstacle-free path deviates from the lane network defined as G, where s={xi,θi} is a trajectory of the vehicle at a state xi,θi, wG is a predetermined weight factor, and D(xi,θi,G) is a distance between the vehicle at a state xi,θi and a nearest edge of the lane network G. 4. The method recited in claim 3, wherein the minimizing the calculated cost ρ(s)=ρ0(s)+wG∑iD(〈xi,θi〉,G) uses a conjugate-gradient descent. 5. The method recited in claim 1, wherein the determining includes determining a plurality of obstacle-free paths between the location of the vehicle and the goal position,each of the plurality of obstacle-free paths has a calculated cost, andan obstacle-free path of the plurality of obstacle-free paths having the lowest calculated cost is recorded onto the electronic memory. 6. The method recited in claim 1, wherein the determining includes smoothing the obstacle-free path. 7. The method recited in claim 1, wherein the environment is a semi-structured environment. 8. A method of planning a path for a vehicle, comprising: accessing a lane network of an environment stored in electronic memory of a computing device, and determining a location and a goal location of the vehicle within the environment using a processor of the computing device;calculating a cost function for a plurality of coordinates of the environment between the location and the goal location using the processor, the cost function based on the lane network, the goal location and a state of the vehicle, the state of the vehicle including the location and an angular orientation of the vehicle; anddetermining an obstacle-free path of the environment by minimizing a sum of the cost functions corresponding to the coordinates which intersect the obstacle-free path between the location and the goal location using the processor and recording the determined obstacle-free path onto the electronic memory. 9. A computer readable medium including computer executable instructions, wherein the instructions, when executed by a processor of a computing device, cause the processor to perform a method comprising: creating a lane network of an environment for a vehicle using the processor of the computing device and storing the lane network in electronic memory of the computing device;calculating a cost function for each of a plurality of coordinates of the environment using the processor from the lane network, a goal position of the vehicle and a state of the vehicle, the state of the vehicle including a location and an angular orientation of the vehicle; anddetermining an obstacle-free path for the vehicle from the location to the goal position by minimizing a calculated cost of the obstacle-free path, the calculated cost being a sum of the cost functions of the coordinates that the obstacle-free path intersects.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (7)
Eberhardt Nikolai (Bethlehem PA), Automatic guided vehicle system.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.