WO2022233332A1 - 一种路径规划方法 - Google Patents

一种路径规划方法 Download PDF

Info

Publication number
WO2022233332A1
WO2022233332A1 PCT/CN2022/091415 CN2022091415W WO2022233332A1 WO 2022233332 A1 WO2022233332 A1 WO 2022233332A1 CN 2022091415 W CN2022091415 W CN 2022091415W WO 2022233332 A1 WO2022233332 A1 WO 2022233332A1
Authority
WO
WIPO (PCT)
Prior art keywords
map
path
algorithm
path planning
episodic
Prior art date
Application number
PCT/CN2022/091415
Other languages
English (en)
French (fr)
Inventor
孙荣川
吴俊毅
郁树梅
陈国栋
孙立宁
Original Assignee
苏州大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 苏州大学 filed Critical 苏州大学
Priority to US17/910,842 priority Critical patent/US11906977B2/en
Publication of WO2022233332A1 publication Critical patent/WO2022233332A1/zh

Links

Classifications

    • G05D1/246—
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
    • G05D1/0214—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory in accordance with safety or protection criteria, e.g. avoiding hazardous areas
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
    • G05D1/0246—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means
    • G05D1/0253—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means extracting relative motion information from a plurality of images taken successively, e.g. visual odometry, optical flow
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
    • G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0268—Control of position or course in two dimensions specially adapted to land vehicles using internal positioning means
    • G05D1/0274—Control of position or course in two dimensions specially adapted to land vehicles using internal positioning means using mapping information stored in a memory device
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0276—Control of position or course in two dimensions specially adapted to land vehicles using signals provided by a source external to the vehicle
    • G05D1/22—
    • G05D1/24—
    • G05D1/2437—

Definitions

  • the invention belongs to the field of mobile robot path planning, in particular to a situational memory path planning method based on memory fusion.
  • the path planning algorithm based on episodic memory can plan the optimal path based on the existing cognitive map, but the path planning method lacks the understanding of the map and may not be the shortest globally.
  • the path planning method In order to improve the navigation efficiency of robots, it is necessary to increase their understanding of cognitive maps and explore potential safe paths in the environment. After fusing the detected potential path with the original map, a better path can be planned for the mobile robot.
  • the purpose of the present invention is to provide a new episodic memory path planning method based on memory fusion, so as to plan a better path.
  • the present invention can adopt following technical scheme:
  • a path planning method comprising the following steps:
  • the episodic memory model is a path planning algorithm
  • the situational cognition map is a two-dimensional incremental matrix composed of discrete limited event spaces and event transition sets;
  • the slope of the road edge makes a preliminary judgment on the connectivity, including: when the absolute value of the difference between the slopes of the road edge is less than a set threshold, judging that there is a possibility of connectivity between two points;
  • the potential path detection network is a two-dimensional network structure inspired by the continuous attractor network in the RatSLAM model, which is used to simulate the process of biological judgment of whether the road is connected;
  • the updated context awareness map refers to a new map obtained by revising the event transition weights in the original context awareness map according to the detected potential path.
  • the RatSLAM algorithm is a bionic navigation algorithm, and the use of the RatSLAM algorithm to establish an experience map includes: establishing a two-dimensional experience map through RGB image information collected by a monocular camera.
  • the Canny operator is an edge extraction algorithm.
  • the above technical solution of the present invention uses the potential path detection network to find potential safe paths in the environment. Compared with the original episodic memory model that only includes the trajectories experienced by the mobile robot in the past time and space, the episodic memory model after integrating the potential path detection network , a better path can be planned for the mobile robot.
  • FIG. 1 is a schematic flowchart of a path planning method according to an embodiment of the present invention.
  • FIG. 2 is a schematic diagram of the motion trajectory of the mobile robot.
  • Figure 3 is a schematic diagram of an empirical map created using the RatSLAM algorithm.
  • FIG. 4 is a schematic diagram of the effect of road edge detection based on the Canny operator.
  • Figure 5 is a schematic diagram of a potential path detection network.
  • FIG. 6 is a schematic diagram of the original episodic cognition map obtained based on the episodic memory model.
  • FIG. 7 is a schematic diagram of an updated situational awareness map obtained by fusing the latent path detection network.
  • Figure 8 is a schematic diagram of the updated experience map generated by the RatSLAM algorithm based on the latent path detection network.
  • a path planning method is provided in one of the embodiments of the present invention.
  • the path planning method is a situational memory path planning method based on memory fusion, and the path planning method includes the following steps :
  • S3 will be converted from the pixel coordinate system to the world coordinate system based on the road edge, and make a preliminary judgment on connectivity according to the slope of the road edge;
  • S6 Perform path planning based on the updated situational awareness map.
  • the RatSLAM algorithm is a bionic navigation algorithm; using the RatSLAM algorithm to establish an experience map includes: establishing a two-dimensional experience map through RGB image information collected by a monocular camera.
  • the episodic memory model is a path planning algorithm; the episodic cognitive map is a two-dimensional incremental matrix composed of discrete finite event spaces and event transition sets.
  • FIG. 2 is a schematic diagram of the motion trajectory 100 of the mobile robot, wherein the dotted line represents the potential path 200 .
  • Figure 3 is a schematic diagram of the empirical map established using the RatSLAM algorithm.
  • the Canny operator is an edge extraction algorithm.
  • the road edge 300 of the corridor is successfully detected based on the Canny operator.
  • the Canny operator has strong anti-interference ability to noise, can adapt to different environments, and has good detection performance, so the Canny operator is selected as the edge detection algorithm.
  • the detection results based on the Canny operator have rich environmental structure edge information, but only the road edge information is needed for path planning, so other redundant edges need to be eliminated. By setting an appropriate detection area, the two longest edges in the image are detected to detect road edge information.
  • step S3 the slope of the road edge 300 makes a preliminary judgment on the connectivity, including: when the absolute value of the difference between the slopes of the road edge 300 is less than a set threshold, preliminary judgment is made that there is a possibility of connectivity between two points.
  • the potential path 200 refers to a potential safe path
  • the potential path detection network is a two-dimensional network structure inspired by the continuous attractor network in the RatSLAM model, which is used to simulate the process of biological judgment of whether the road is connected.
  • Figure 5 it is a schematic diagram of the latent path detection network.
  • the updated context awareness map refers to a new map obtained by modifying the event transfer weights in the original context awareness map according to the detected potential path 200 .
  • FIG. 6 it is a schematic diagram of the original episodic cognition map obtained based on the episodic memory model in S1.
  • FIG. 7 it is a schematic diagram of the updated contextual cognitive map obtained after the original contextual cognitive map is fused with the potential path detection network.
  • FIG. 8 it is a schematic diagram of the updated experience map generated by the RatSLAM algorithm based on the latent path detection network.
  • the potential path detection network is used to find potential safe paths in the environment. Compared with the original episodic memory model that only contains the trajectories experienced by the mobile robot in the past time and space, the episodic memory after fusion of the potential path detection network is used. The model can plan a better path for the mobile robot.

Abstract

一种路径规划方法，其包括如下步骤：S1,基于情景记忆模型，使用RatSLAM算法建立经验地图和相应的情景认知地图；S2,通过Canny算子提取历史记忆图像中的道路边缘；S3，将基于所述道路边缘从像素坐标系转换到世界坐标系下，根据所述道路边缘的斜率对连通性做初步判断；S4，根据对潜在路径的持续观测，不断向所述潜在路径检测网络注入能量，以对道路连通性做进一步判断；S5，将检测到的所述潜在路径与原有的所述情景认知地图融合，并对经验地图进行相应的更新；S6，基于更新后的所述情景认知地图进行路径规划。路径规划方法能够检测环境中潜在的安全路径，基于更新后的情景记忆模型规划出更优的路径。

Description

一种路径规划方法 技术领域
本发明属于移动机器人路径规划领域，具体涉及一种基于记忆融合的情景记忆路径规划方法。
 背景技术
移动机器人在陌生环境中对环境进行建模时，需要规划出一条通往目标的安全路径以开展相关导航任务。由于现实中实际环境的复杂性，需要算法对环境中的动态信息有一定的适应能力，传统的路径规划算法如A*(A-Star，A*)、人工势场法以及快速搜索随机树(Rapidly-exploring Random Trees，RRT)等算法的性能会受到较大的限制，在复杂环境中的路径规划效果并不理想。其中，A*算法规划出的路径不够平滑，人工势场法易陷入局部最优，RRT算法搜索效率较低，且规划出的并不是最优路径。
许多自然界中的生物都可以对环境进行认知学习，并在复杂的动态场景中高效地完成导航任务，这种生物拥有的特殊导航本领极大地提高了研究人员对于认知启发的兴趣。研究表明，在规划路径时海马体内的情景记忆机制起到了极大的作用。在进行环境认知的过程中，海马体对过去经历的各种情景进行记忆；当给定导航任务时，通过提取与任务相关的记忆片段规划出一条最优路径。
 技术问题
基于情景记忆的路径规划算法可以基于已有的认知地图规划出最优路径，但该路径规划方法缺少对地图的理解，在全局上可能不是最短的。为了提高机器人的导航效率，需要增加其对认知地图的理解，探索环境中潜在的安全路径。将检测到的潜在路径与原地图融合后，可以为移动机器人规划出一条更优的路径。
因此，研究从环境中寻找潜在的安全路径从而完善认知地图，对于优化基于情景记忆模型规划的路径具有着重要意义。
 技术解决方案
本发明目的是：提供一种新的基于记忆融合的情景记忆路径规划方法，以规划出一条更优的路径。
本发明可采用如下技术方案：
一种路径规划方法，其包括如下步骤：
基于情景记忆模型，使用RatSLAM算法建立经验地图和相应的情景认知地图；
通过Canny算子提取历史记忆图像中的道路边缘；
将基于所述道路边缘从像素坐标系转换到世界坐标系下，根据所述道路边缘的斜率对连通性做初步判断；
根据对潜在路径的持续观测，不断向所述潜在路径检测网络注入能量，以对道路连通性做进一步判断；
将检测到的所述潜在路径与原有的所述情景认知地图融合，并对经验地图进行相应的更新；
基于更新后的所述情景认知地图进行路径规划；
所述情景记忆模型是一种路径规划算法；
所述情景认知地图是一种二维增量式矩阵，由离散有限的事件空间和事件转移集合组成；
所述道路边缘的斜率对连通性做初步判断，包括：当所述道路边缘的斜率之差的绝对值小于设定的阈值时，判断两点之间存在连通的可能性；
所述潜在路径检测网络是受RatSLAM模型中连续吸引子网络启发提出的一种二维网络结构，用于模拟生物判断道路是否连通的过程；
所述更新后的情景认知地图是指根据检测到的所述潜在路径对原有的所述情景认知地图中事件转移权值进行修正后得到的新地图。
进一步的，所述RatSLAM算法是一种仿生导航算法，所述使用RatSLAM算法建立经验地图包括：通过单目相机采集的RGB图像信息建立二维经验地图。
进一步的，所述Canny算子是一种边缘提取算法。
 有益效果
本发明上述技术方案，使用潜在路径检测网络寻找环境中潜在的安全路径，相比于原来的仅包含过去的时空中移动机器人经历过的轨迹的情景记忆模型，融合潜在路径检测网络后的情景记忆模型，可以为移动机器人规划出一条更优的路径。
 附图说明
下面结合附图及实施例对本发明作进一步描述：
图1为本发明一实施例的路径规划方法的流程示意图。
图2为移动机器人运动轨迹示意图。
图3为使用RatSLAM算法建立的经验地图示意图。
图4为基于Canny算子的道路边缘检测效果示意图。
图5为潜在路径检测网络示意图。
图6为基于情景记忆模型得到的原有的情景认知地图示意图。
图7为融合潜在路径检测网络后得到的更新后的情景认知地图示意图。
图8为RatSLAM算法基于潜在路径检测网络生成的更新后的经验地图示意图。
 本发明的最佳实施方式
下面将结合附图以及实施例，对本公开做进一步说明。
如图1所示，在本发明的其中一种实施方式中提供一种路径规划方法，本实施例中，该路径规划方法是基于记忆融合的情景记忆路径规划方法，该路径规划方法包括如下步骤：
S1:基于情景记忆模型，使用RatSLAM算法建立经验地图和相应的情景认知地图；
S2:通过Canny算子提取历史记忆图像中的道路边缘；
S3:将基于所述道路边缘从像素坐标系转换到世界坐标系下，根据所述道路边缘的斜率对连通性做初步判断；
S4:根据对潜在路径的持续观测，不断向所述潜在路径检测网络注入能量，以对道路连通性做进一步判断；
S5:将检测到的所述潜在路径与原有的所述情景认知地图融合，并对经验地图进行相应的更新；
S6:基于更新后的所述情景认知地图进行路径规划。
步骤S1中，RatSLAM算法是一种仿生导航算法；使用RatSLAM算法建立经验地图包括：通过单目相机采集的RGB图像信息建立二维经验地图。情景记忆模型是一种路径规划算法；情景认知地图是一种二维增量式矩阵，由离散有限的事件空间和事件转移集合组成。如图2所示为移动机器人运动轨迹100示意图，其中虚线代表潜在路径200。如图3所示是使用RatSLAM算法建立的经验地图示意图。
步骤S2中，Canny算子是一种边缘提取算法。如图4所示为基于Canny算子成功检测出走廊的道路边缘300。上文中，Canny算子对噪声有较强的抗干扰能力，可以适应不同的环境，并具有良好的检测性能，所以选择Canny算子作为边缘检测算法。基于Canny算子的检测结果具有丰富的环境结构边缘信息，但在开展路径规划时只需要道路边缘信息，所以需要对其它冗余的边缘进行剔除。通过设定合适的检测区域对图像中最长的两条边缘进行检测，从而检测道路边缘信息。
步骤S3中，道路边缘300的斜率对连通性做初步判断，包括：当所述道路边缘300的斜率之差的绝对值小于设定的阈值时，初步判断两点之间存在连通的可能性。
步骤S4中，潜在路径200是指潜在的安全路径，潜在路径检测网络是受RatSLAM模型中连续吸引子网络启发提出的一种二维网络结构，用于模拟生物判断道路是否连通的过程。如图5所示，是潜在路径检测网络的示意图。
步骤S6中，所述更新后的情景认知地图是指根据检测到的所述潜在路径200对原有的所述情景认知地图中事件转移权值进行修正后得到的新地图。
如图6所示，是S1中基于情景记忆模型得到的原有的情景认知地图示意图。如图7所示，是原有的情景认知地图融合潜在路径检测网络后得到的更新后的情景认知地图示意图。如图8所示，是RatSLAM算法基于潜在路径检测网络生成的更新后的经验地图示意图。
本发明的上述实施例，使用潜在路径检测网络寻找环境中潜在的安全路径，相比于原来的仅包含过去的时空中移动机器人经历过的轨迹的情景记忆模型，融合潜在路径检测网络后的情景记忆模型，可以为移动机器人规划出一条更优的路径。
当然上述实施例只为说明本发明的技术构思及特点，其目的在于让熟悉此项技术的人能够了解本发明的内容并据以实施，并不能以此限制本发明的保护范围。凡根据本发明主要技术方案的精神实质所做的修饰，都应涵盖在本发明的保护范围之内。

Claims (3)

  1. 一种路径规划方法，其特征在于，包括如下步骤：
    基于情景记忆模型，使用RatSLAM算法建立经验地图和相应的情景认知地图；
    通过Canny算子提取历史记忆图像中的道路边缘；
    将基于所述道路边缘从像素坐标系转换到世界坐标系下，根据所述道路边缘的斜率对连通性做初步判断；
    根据对潜在路径的持续观测，不断向所述潜在路径检测网络注入能量，以对道路连通性做进一步判断；
    将检测到的所述潜在路径与原有的所述情景认知地图融合，并对经验地图进行相应的更新；
    基于更新后的所述情景认知地图进行路径规划；
    所述情景记忆模型是一种路径规划算法；
    所述情景认知地图是一种二维增量式矩阵，由离散有限的事件空间和事件转移集合组成；
    所述道路边缘的斜率对连通性做初步判断，包括：当所述道路边缘的斜率之差的绝对值小于设定的阈值时，判断两点之间存在连通的可能性；
    所述潜在路径检测网络是受RatSLAM模型中连续吸引子网络启发提出的一种二维网络结构，用于模拟生物判断道路是否连通的过程；
    所述更新后的情景认知地图是指根据检测到的所述潜在路径对原有的所述情景认知地图中事件转移权值进行修正后得到的新地图。
  2. 根据权利要求1所述的路径规划方法，其特征在于：所述RatSLAM算法是一种仿生导航算法，所述使用RatSLAM算法建立经验地图包括：通过单目相机采集的RGB图像信息建立二维经验地图。
  3. 根据权利要求1所述的路径规划方法，其特征在于：所述Canny算子是一种边缘提取算法。
PCT/CN2022/091415 2021-05-07 2022-05-07 一种路径规划方法 WO2022233332A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/910,842 US11906977B2 (en) 2021-05-07 2022-05-07 Path planning method

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110494874.7A CN113110492B (zh) 2021-05-07 2021-05-07 一种路径规划方法
CN202110494874.7 2021-05-07

Publications (1)

Publication Number Publication Date
WO2022233332A1 true WO2022233332A1 (zh) 2022-11-10

Family

ID=76721227

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/091415 WO2022233332A1 (zh) 2021-05-07 2022-05-07 一种路径规划方法

Country Status (3)

Country Link
US (1) US11906977B2 (zh)
CN (1) CN113110492B (zh)
WO (1) WO2022233332A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113110492B (zh) 2021-05-07 2022-06-21 苏州大学 一种路径规划方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060129506A1 (en) * 2004-07-15 2006-06-15 Neurosciences Research Foundation, Inc. Mobile brain-based device having a simulated nervous system based on the hippocampus
CN106814737A (zh) * 2017-01-20 2017-06-09 安徽工程大学 一种基于啮齿类动物模型和RTAB‑Map闭环检测算法的SLAM方法
CN106949896A (zh) * 2017-05-14 2017-07-14 北京工业大学 一种基于鼠脑海马的情景认知地图构建及导航方法
US20180046153A1 (en) * 2016-07-10 2018-02-15 Beijing University Of Technology Method of Constructing Navigation Map by Robot using Mouse Hippocampal Place Cell Model
CN109668566A (zh) * 2018-12-05 2019-04-23 大连理工大学 一种基于鼠脑定位细胞的机器人情景认知地图的构建与导航方法
CN113110492A (zh) * 2021-05-07 2021-07-13 苏州大学 一种路径规划方法

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7765029B2 (en) * 2005-09-13 2010-07-27 Neurosciences Research Foundation, Inc. Hybrid control device
US8103602B2 (en) * 2006-12-29 2012-01-24 Neurosciences Research Foundation, Inc. Solving the distal reward problem through linkage of STDP and dopamine signaling
CN101701830A (zh) * 2009-11-30 2010-05-05 四川虹微技术有限公司 一种导航终端的导航方法
CN103699125B (zh) * 2013-12-09 2016-04-06 北京工业大学 一种基于鼠脑海马导航的机器人仿真导航方法
CN107363834B (zh) * 2017-07-20 2020-09-29 电子科技大学 一种基于认知地图的机械臂抓取方法
CN208126153U (zh) * 2018-04-23 2018-11-20 安徽工程大学 一种基于仿生slam的移动机器人特能分析实验平台
CN111474932B (zh) * 2020-04-23 2021-05-11 大连理工大学 一种集成情景经验的移动机器人建图与导航方法
CN111376273B (zh) * 2020-04-23 2022-04-12 大连理工大学 一种类脑启发的机器人认知地图构建方法
CN111813113B (zh) * 2020-07-06 2021-07-02 安徽工程大学 仿生视觉自运动感知地图绘制方法、存储介质及设备
CN111860643B (zh) * 2020-07-20 2023-10-03 苏州大学 基于调频模型的视觉模板匹配的鲁棒性提升方法
CN112097769B (zh) * 2020-08-05 2022-06-10 北京航空航天大学 仿信鸽脑-海马的无人机同时定位与建图导航系统及方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060129506A1 (en) * 2004-07-15 2006-06-15 Neurosciences Research Foundation, Inc. Mobile brain-based device having a simulated nervous system based on the hippocampus
US20180046153A1 (en) * 2016-07-10 2018-02-15 Beijing University Of Technology Method of Constructing Navigation Map by Robot using Mouse Hippocampal Place Cell Model
CN106814737A (zh) * 2017-01-20 2017-06-09 安徽工程大学 一种基于啮齿类动物模型和RTAB‑Map闭环检测算法的SLAM方法
CN106949896A (zh) * 2017-05-14 2017-07-14 北京工业大学 一种基于鼠脑海马的情景认知地图构建及导航方法
CN109668566A (zh) * 2018-12-05 2019-04-23 大连理工大学 一种基于鼠脑定位细胞的机器人情景认知地图的构建与导航方法
CN113110492A (zh) * 2021-05-07 2021-07-13 苏州大学 一种路径规划方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
JIANG, XIAOJUN: "Research on the Episodic Cognitive Mapping and Robot Navigation Method Based on the Cognitive Mechanism of Rat Hippocampus", CHINESE MASTER’S THESES FULL-TEXT DATABASE (ELECTRONIC JOURNAL), 1 June 2018 (2018-06-01), pages 1 - 87, XP093000052 *

Also Published As

Publication number Publication date
US11906977B2 (en) 2024-02-20
US20230195134A1 (en) 2023-06-22
CN113110492A (zh) 2021-07-13
CN113110492B (zh) 2022-06-21

Similar Documents

Publication Publication Date Title
CN112132893B (zh) 一种适用于室内动态环境的视觉slam方法
Ruan et al. Mobile robot navigation based on deep reinforcement learning
CN103389699B (zh) 基于分布式智能监测控制节点的机器人监控及自主移动系统的运行方法
US20200372822A1 (en) Training system for autonomous driving control policy
CN106096729A (zh) 一种面向大规模环境中复杂任务的深度策略学习方法
CN111098301B (zh) 一种基于场景知识图谱任务型机器人的控制方法
CN106949893A (zh) 一种三维避障的室内机器人导航方法和系统
CN111645073B (zh) 一种机器人视觉语义导航方法、装置及系统
Chen et al. Driving with llms: Fusing object-level vector modality for explainable autonomous driving
CN110084307A (zh) 一种基于深度强化学习的移动机器人视觉跟随方法
Nguyen et al. Autonomous navigation in complex environments with deep multimodal fusion network
CN104834308A (zh) 满足复杂需求的最优巡回控制方法
CN110728751A (zh) 一种室内3d点云语义地图的构建方法
CN111814823B (zh) 一种基于场景模板生成的迁移学习方法
WO2022233332A1 (zh) 一种路径规划方法
Li et al. Inspection robot based on offline digital twin synchronization architecture
Liu et al. A hierarchical reinforcement learning algorithm based on attention mechanism for uav autonomous navigation
CN113961013A (zh) 一种基于rgb-d slam的无人机路径规划方法
CN108921852B (zh) 基于视差与平面拟合的双分支室外非结构化地形分割网络
Magid Simulation Tools for Urban Search and Rescue Robotics
CN114594768B (zh) 一种基于视觉特征图重构的移动机器人导航决策方法
CN116852347A (zh) 一种面向非合作目标自主抓取的状态估计与决策控制方法
Li et al. Vision-based obstacle avoidance algorithm for mobile robot
Lv et al. A deep safe reinforcement learning approach for mapless navigation
CN112857373B (zh) 一种最小化无用动作的节能性无人车路径导航方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22798672

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 22798672

Country of ref document: EP

Kind code of ref document: A1