Local path planning method and system based on spatial memory neural mechanism

Abstract

The invention discloses a local path planning method and system based on a spatial memory neural mechanism, and aims to improve obstacle avoidance and shortcut discovery capabilities of a robot in a complex environment. The algorithm is inspired by a hippocampus navigation neural circuit, and efficient memory search and path planning are realized by simulating functions of grid cells, position cells and boundary cells. Firstly, a robot explores a feasible path and records memory nodes including grid cell characterization, reward values and cognitive values; secondly, constructing a hierarchical memory module, realizing fine-to-sparse information abstraction from the back side to the ventral side, and realizing information inheritance and memory search through connection at fixed intervals; and finally, performing path planning by utilizing direction decoding and boundary cell neural representation, and dynamically selecting an optimal sub-target point. The algorithm does not need a large amount of training, can be quickly deployed in various robots, and realizes intelligent autonomous navigation.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence and path planning technology, and in particular to a local path planning method and system based on a spatial memory neural mechanism. Background Art

[0002] In recent years, the field of neurobiology has made significant progress in the study of biological navigation mechanisms. As one of the core theories to explain biological navigation, cognitive map theory has been widely recognized by researchers. The theory holds that animals form internal representations of the environment through brain centers (especially the hippocampus and entorhinal cortex) to achieve spatial positioning and path planning. In this process, specific nerve cells in the hippocampus and entorhinal cortex of the brain, such as place cells, grid cells, and boundary cells, play a vital role. The prefrontal cortex and basal ganglia are involved in cognitive functions such as strategic planning, decision-making, and action selection. This complex collaboration of neural circuits enables organisms to quickly adapt and navigate in dynamic and complex environments.

[0003] On this basis, researchers have proposed a brain-like navigation algorithm to improve the robot's autonomous navigation ability in complex environments by simulating this biological navigation mechanism. The advantage of the brain-like navigation algorithm is that it can use biologically inspired mechanisms to perform more flexible and intelligent path planning and decision-making in unknown or dynamic environments. Therefore, the development of brain-like navigation algorithms based on neural mechanisms can significantly improve the robot's performance in obstacle avoidance, path optimization, and task execution. However, although this research direction is of great significance to promoting the practical application of robot brain-like navigation technology, it still faces many challenges. Summary of the invention

[0004] Based on this, the embodiments of the present application provide a local path planning method and system based on a spatial memory neural mechanism, combined with an efficient memory search strategy, to improve the obstacle avoidance and shortcut discovery capabilities in complex environments, and achieve a more intelligent and autonomous navigation effect.

[0005] In a first aspect, a local path planning method based on a spatial memory neural mechanism is provided, the method comprising:

[0006] Step 1: The robot is manually controlled to explore a feasible path from the current position to the target point in the environment, and the robot senses the surrounding environment through the laser radar carried by the robot, and records the information of the path points as the robot's memory information source;

[0007] Step 2, after obtaining the information of the path points, the information of the recorded path points is processed according to the organizational hierarchical memory network to realize dynamic path planning; wherein the organizational hierarchical memory network includes grid cells, place cells and boundary cells;

[0008] Step 3, initialize parameters, calculate the motion information from the current position to the target position, obtain the target direction and the indication signal value; when the signal threshold is reached, it is considered that the robot has reached the vicinity of the target and the planning is completed; otherwise, select the motion direction.

[0009] Optionally, the robot can sense the surrounding environment through a laser radar and record the path points, including:

[0010] The information of each waypoint is represented as a tuple, which contains the grid cell representation pattern, reward value and cognitive value; the grid cell representation pattern is the spatial coordinate of the waypoint, the reward value is generated based on the reverse memory playback of the target point, and the cognitive value represents the level of the memory node. The higher the level, the greater the cognitive value.

[0011] Optionally, the hierarchical memory network includes grid cells, place cells, and border cells, specifically including:

[0012] Generate an organized hierarchical memory network based on the hippocampal navigation neural circuit structure;

[0013] Grid cells are distributed in the medial entorhinal cortex and are used to provide a spatial coordinate system and perform vector navigation functions through the motor output module.

[0014] Place cells are distributed in the CA region of the hippocampus and are used to connect synapses to support the storage and recall functions of memory, as well as to support the hierarchical organization and efficient search of memory at different spatial sparsities;

[0015] Boundary cells are distributed in the lateral entorhinal cortex and are used to perceive local obstacle information and indicate whether a direction is navigable.

[0016] Optionally, in step 2, after obtaining the information of the path points, the information of the recorded path points is processed according to the organized hierarchical memory network to implement dynamic path planning, including:

[0017] From dorsal to ventral direction, at fixed intervals o l The upper layer inherits the memory information of the exploration nodes in the lower layer and completes the abstract representation from fine to sparse. Among them, the tuple of the jlth memory node in the memory pool of the lth layer is:

[0018]

[0019] From the ventral to the dorsal direction, at fixed intervals o l Connect the exploration nodes from the upper layer to the lower layer to complete the memory search from abstract to concrete. The connection weight from layer l to layer l-1 is:

[0020]

[0021] Optionally, step 3, initializing parameters, calculating motion information from the current position to the target position, and obtaining the target direction and the reaching indication signal value, includes:

[0022] Step 3.1, use the direction decoding module to calculate the motion information from the current position to the target position, obtain the target direction and the indication signal value. When the signal threshold is reached, it is considered that the robot has reached the vicinity of the target and the planning is completed; otherwise, select the motion direction; otherwise, go to step 3.1;

[0023] Step 3.2, select the direction of movement, first determine whether the target direction is passable:

[0024] s=B*T(θ target )

[0025] If s=1, the target direction θ is considered target It is passable, a=θ target , this strategy is called vector navigation strategy; conversely, if s = 0, the target direction θ target It is not passable, a = θ, and the strategy is called integrated navigation strategy;

[0026] Step 3.3: Update the current position p according to the direction in step 3.2, and then go to step 3.1.

[0027] Step 3.4: Calculate the direction of movement using the integrated navigation strategy.

[0028] Optionally, step 3.4: calculating the direction of motion using an integrated navigation strategy, specifically includes:

[0029] Step 3.4.1: Calculate the direction of the memory node. For the memory pool of the first layer, calculate the direction of the memory node involved in the search, and use the search set I l Provides indication. When l = 1, I l =[1,2,...,n E ]; otherwise, it is provided by step 3.4.3. The direction of the jlth node is obtained by the direction decoding module:

[0030]

[0031] Step 3.4.2: Modulation of reward excitation-boundary inhibition to obtain the discharge state of the memory node in the lth layer

[0032]

[0033] Then we get the memory node search vector

[0034]

[0035] Step 3.4.3: If l = 1, output the optimal direction Otherwise, perform a low-level search, searching set I l-1 for:

[0036]

[0037] In a second aspect, a local path planning system based on a spatial memory neural mechanism is provided, the system comprising:

[0038] The recording unit controls the robot to explore a feasible path from the current position to the target point in the environment, and uses the laser radar carried by the robot to perceive the surrounding environment and record the information of the path points as the robot's memory information source;

[0039] A memory network unit is used to process the recorded information of the path points according to the organization hierarchical memory network to realize dynamic path planning after obtaining the information of the path points; wherein the organization hierarchical memory network includes grid cells, place cells and boundary cells;

[0040] The processing unit is used to initialize parameters, calculate the motion information from the current position to the target position, obtain the target direction and the reaching indication signal value; when the signal threshold is reached, it is considered that the robot has reached the vicinity of the target and the planning is completed; otherwise, the motion direction is selected.

[0041] In a third aspect, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the local path planning method described in any one of the first aspects is implemented.

[0042] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the local path planning method described in any one of the first aspects is implemented.

[0043] In a fifth aspect, a computer program product is provided, comprising a computer program / instruction, which, when executed by a processor, implements the local path planning method described in any one of the first aspects.

[0044] The beneficial effects brought by the technical solution provided by the embodiment of the present application include at least:

[0045] (1) Flexible obstacle avoidance and shortcut discovery capabilities: Based on the navigation neural computing mechanism of mammals, the present invention integrates external obstacle perception information and internal contextual memory to give the robot more flexible obstacle avoidance capabilities and the ability to discover unexplored shortcuts. Compared with the prior art, the present invention adopts a strategy of multiple neuron models (such as boundary cells, place cells, and grid cells) to support the dynamic switching of vector navigation and integrated navigation, thereby achieving more intelligent path planning decisions.

[0046] (2) Multi-scenario path planning that reduces the computational cost of mapping: Compared with traditional path planning technology, the present invention only needs to manually control a trajectory to reach the target to provide global path guidance, without the need to fully explore and build a detailed map of the scene in advance. This advantage is mainly due to the fact that the present invention combines an efficient memory search strategy on the basis of global path and local obstacle avoidance, avoiding the complete modeling of the entire environment and greatly saving the computational cost of mapping. Therefore, the present invention can implement an efficient path planning method in multiple scenarios and multiple tasks.

[0047] (3) Hierarchical memory structure to improve computational efficiency and interpretability: Compared with existing brain-like navigation technologies, the present invention is not only limited to the role of the representation pattern of a single neuron in spatial navigation, but also focuses on the computational function of the hippocampal circuit. By introducing a multi-level brain-like memory structure (such as ventral sparse memory and dorsal fine memory), the computational efficiency and interpretability of the path planning algorithm are improved. This hierarchical structure eliminates the need for the algorithm to traverse all memory information during the path search process. Instead, it uses a search method similar to a "skip table" to reduce the computational complexity from linear to logarithmic level while maintaining the optimal path length. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0049] Figure 1 Flowchart of the local path planning method based on spatial memory neural mechanism

[0050] Figure 2 This is a map of the neural circuits and spatial cell distribution in the hippocampus;

[0051] Figure 3 This is a schematic diagram of the control process of the hippocampus model;

[0052] Figure 4 A schematic diagram of the information inheritance and weight connection in the ventral-dorsal axis of the hippocampus;

[0053] Figure 5 Schematic diagram for integrated navigation strategy calculation;

[0054] Figure 6 Provide a flow chart for the strategy organization of the path planning algorithm;

[0055] Figure 7 Schematic diagram of the simulation environment (A. obstacle environment; B. hairpin maze);

[0056] Figure 8 Results of path planning in an obstacle environment (A. Visualization of hierarchical memory pool; B. Visualization of connection weights; C. Result of strategy switching on two paths: sky blue for vector navigation strategy, purple for integrated navigation strategy; D. Result of path planning for 50 groups with different starting points);

[0057] Fig. 9 Results of path planning in the hairpin maze (A. Visualization of the hierarchical memory pool; B. Strategy switching results of the original maze; C. Strategy switching results of three sets of new mazes; sky blue represents the vector navigation strategy, and purple represents the integrated navigation strategy);

[0058] Fig.10 Statistical graphs of path planning results under single-layer and multi-layer memory structures (A. Path length statistics; B. Computation time statistics). DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0060] In the description of the present invention, the terms "comprises", "has" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may also include other steps or units that are not explicitly listed but are inherent to these processes, methods, products or apparatuses, or steps or units added based on further optimization schemes conceived by the present invention.

[0061] The purpose of this invention is to provide a local path planning algorithm based on the working mechanism of the biological brain, combined with an efficient memory search strategy, to improve the obstacle avoidance and shortcut discovery capabilities in complex environments, and achieve a more intelligent and autonomous navigation effect.

[0062] Please refer to Figure 1, which shows a flow chart of a local path planning method based on a spatial memory neural mechanism provided by an embodiment of the present application. The method may include the following steps:

[0063] Step 1: The robot is manually controlled to explore a feasible path from the current position to the target point in the environment, and the robot senses the surrounding environment through the laser radar carried by the robot, and records the information of the path points as the robot's memory information source.

[0064] Among them, the information of each waypoint is represented as a tuple, which contains the grid cell representation pattern, reward value and cognitive value; the grid cell representation pattern is the spatial coordinates of the waypoint, the reward value is generated based on the reverse memory playback of the target point, and the cognitive value represents the level of the memory node. The higher the level, the greater the cognitive value.

[0065] Step 2, after obtaining the information of the path points, the information of the recorded path points is processed according to the organizational hierarchical memory network to realize dynamic path planning; wherein the organizational hierarchical memory network includes grid cells, place cells and boundary cells;

[0066] Step 3, initialize parameters, calculate the motion information from the current position to the target position, obtain the target direction and the indication signal value; when the signal threshold is reached, it is considered that the robot has reached the vicinity of the target and the planning is completed; otherwise, select the motion direction.

[0067] The local path planning algorithm in this invention is based on the working mechanism of the biological brain. By combining an efficient memory search strategy, it improves the robot's obstacle avoidance and shortcut discovery capabilities in complex environments, achieving a more intelligent and autonomous navigation effect. The design of the entire algorithm is inspired by the structure and function of the hippocampal navigation neural circuit. Its specific structure is as follows Figure 2 The core of the algorithm consists of the following three key neuron types and their corresponding functional modules:

[0068] Grid cells: Distributed in the medial entorhinal cortex (MEC), they are responsible for providing a spatial coordinate system and performing vector navigation functions through the motor output module. Grid cells decode the motor output pattern to further obtain the direction of movement and the signal strength when approaching the target.

[0069] Place cells: Located in the CA region of the hippocampus, especially in the CA1 and CA3 regions, and have different functions in these two regions. The recurrently connected synapses in the CA3 region support the storage and recall functions of memory, while the changes in the position signals along the ventral-dorsal axis in the CA1 region support the hierarchical organization and efficient search of memory at different spatial sparsities.

[0070] Boundary cells: distributed in the lateral entorhinal cortex (LEC), they are mainly used to perceive local obstacle information and indicate whether the direction is passable.

[0071] Boundary cells, place cells, and grid cells work together to dynamically select the optimal sub-target points during real-time navigation, and support the flexible selection and integration of multiple strategies to achieve dynamic path planning and obstacle avoidance. In order to achieve simple and efficient deployment, the internal weights and parameters of each functional module in the present invention are pre-configured without relying on a large amount of training process. Therefore, the algorithm can be easily migrated and deployed to various robots and automated equipment to achieve instant navigation functions. In addition, other advanced control logic implemented in the prefrontal cortex or basal ganglia is represented by explicit rules to ensure the robustness and adaptability of the system.

[0072] Figure 3 A control diagram inspired by the hippocampal circuit is shown, and the specific steps are as follows:

[0073] Step 1: Human-controlled robot explores a feasible path from the current position to the target point This is used as the memory information source of the robot. The information of the memory node is represented by a tuple:

[0074] Ω j =(g j , r j , v j ), j = 1, 2, ..., n E .

[0075] Among them, g j is the path point p j The grid cell representation pattern, r j and v j The reward value and cognitive value of the corresponding point. The reward value is generated by the reverse memory playback after the target point is determined, that is, The cognitive value is determined by the layer where the memory node is located. Simply put, the higher the layer, the greater the cognitive value.

[0076] Step 2: Organize the hierarchical memory modules in the ventral and dorsal direction of the hippocampus, such as Figure 4 Specifically, from the dorsal to the ventral direction, at a fixed interval o l The upper layer inherits the memory information of the lower layer exploration node and completes the abstract representation from fine to sparse. Then, the jth node in the lth layer memory pool l The tuple of memory nodes is:

[0077]

[0078] From the ventral to the dorsal direction, at fixed intervals o l Connect the exploration nodes from the upper layer to the lower layer to complete the memory search from abstract to concrete. Then, the connection weight from layer l to layer l-1 is:

[0079]

[0080] Step 3: Enter path planning and first initialize the parameters: initial position (current position) p, initial position and target position grid cell pattern g, exploration trajectory and multi-layer memory pool Ω obtained by steps 1-2, arrival signal threshold U thre The method of the present application provides a flexible choice of navigation strategies, firstly introducing a direction decoding module and a boundary cell neural representation method.

[0081] Direction decoding module: Execute vector decoding module D through grid cell representation of current position and target position:

[0082]

[0083] This module includes n D direction detectors, each with its neural activity uk and preferred direction θ k , uniformly sampled from [0, 2π]. By decoding the motion signal Θ, the target direction θ and the reaching indication signal value u are obtained:

[0084]

[0085] u=||Θ||2

[0086] Boundary cell neural activity: Through the onboard laser radar (the full-circle scanning rays are numbered i=1, 2, ..., n b ) and the preset distance threshold, we can get the binary direction feasible indication information (also the boundary cell discharge pattern) B, that is, when the detection direction If the obstacle distance is higher than the threshold, it is considered passable, and if it is lower than the threshold, it is considered impassable. Using the angle fine-tuning method T(·), the angle θ to be examined is adjusted to the detection direction θ. i Align and get Φ:

[0087]

[0088] in, If and only if i = argmax(cos(θ-θ i )) is activated, which is essentially still a binary discharge mode.

[0089] The detailed process of step 3 is as follows:

[0090] Step 3.1: Use the direction decoding module to calculate the motion information from the current position to the target position and obtain the target direction θ target and reaches the indication signal value u. If u thre , the robot is considered to have reached the vicinity of the target, and the planning stops; otherwise, it goes to step 3.2;

[0091] Step 3.2: Select the movement direction a. First determine whether the target direction is passable:

[0092] s=B*T(θ target )

[0093] If s=1, the target direction θ is considered targe t is passable, a=θ target , this strategy is called vector navigation strategy; conversely, if s = 0, the target direction θ target No access. This strategy is called the integrated navigation strategy, see step 3.5;

[0094] Step 3.3: Update the current position p according to the direction in step 3.2, and then go to step 3.1.

[0095] Step 3.4: Calculate the direction of movement using the integrated navigation strategy like Figure 5 As shown, the specific steps are as follows:

[0096] Step 3.4.1: Calculate the direction of the memory node. For the memory pool of the first layer, calculate the direction of the memory node involved in the search, and use the search set I l Provides indication. When l = 1, I l =[1, 2, ..., n E ]; otherwise, it is provided by step 3.4.3. The direction of the jlth node is obtained by the direction decoding module:

[0097]

[0098] Step 3.4.2: Modulation of reward excitation-boundary inhibition to obtain the discharge state of the memory node in the lth layer

[0099]

[0100] According to the "winner takes all" rule, the memory node search vector is obtained

[0101]

[0102] Step 3.4.3: If l = 1, output the optimal direction ​Otherwise, perform a low-level search, searching set I l-1 for:

[0103]

[0104] like Figure 6 A flowchart showing the technical implementation details is provided. In order to verify the usability and efficiency of the algorithm of the present invention, the present application is Figure 7 The two different environments shown were tested and the resulting red tracks were based on the method described in step 1.

[0105] Example 1: Flexible obstacle avoidance in an open environment with distributed obstacles

[0106] exist Figure 7 In the open environment shown in the figure, the exploration trajectory is first organized into a three-layer memory structure, such as Figure 8 A shows the visualization of the connection matrix. Figure 8 B. Next, the path planning process is performed. The algorithm of the present invention can flexibly switch between the vector navigation strategy (sky blue track) and the integrated navigation strategy (blue track), such as Figure 8 As shown in C. Figure 8 D shows 50 sets of planned trajectories generated from different starting points, all of which successfully avoided obstacles and reached the vicinity of the target, demonstrating the algorithm's reliable obstacle avoidance capability.

[0107] Example 2: Finding shortcuts in a complex maze

[0108] exist Figure 7 In the complex maze environment shown in B, the exploration trajectory is first organized into a three-layer memory structure, such as Fig. 9 A. Path planning is then performed. Fig. 9 B shows the trajectory planned on the original maze map. Fig. 9 In the three different shortcut finding tasks shown in C, the algorithm successfully found shortcuts, demonstrating its path optimization capability in complex environments.

[0109] Based on the results of Example 1 and Example 2, the path planning algorithm of the present invention not only achieves flexible obstacle avoidance by combining the memory search strategy, but also can find shortcuts in unexplored paths, which reflects the efficiency and practicality of the algorithm.

[0110] Example 3: Efficient Search of Multi-layer Memory Mechanism

[0111] In this embodiment, the present application removes the multi-level memory structure and only uses a single-layer memory module directly generated based on the exploration trajectory to re-test the embodiments 1 and 2. Fig.10As shown in the figure, the difference between the planning path length and the calculation time is compared. The results show that the search using a single-layer memory module does not significantly change the path length, but the calculation time increases by about 5-7 times. This further proves the advantage of the multi-level memory search mechanism in improving calculation efficiency. The computer configuration used for the test is Ubuntu18.04. W-2195CPU@2.30GHz.

[0112] The experimental results show that: in Examples 1 and 2, the boundary cells indicate the passability of the target direction by perceiving the distance of external obstacles, and serve as a direct indicator for switching between vector navigation and integrated navigation strategies; grid cells support the implementation of vector navigation, combining past memories with current perception information to achieve integrated navigation. In Example 3, the hierarchical search method from ventral sparse memory to dorsal fine memory not only avoids access to all memory information, but also significantly reduces the computational complexity, while ensuring that the length of the planned path does not increase. Therefore, the present invention can efficiently and effectively complete the robot's obstacle avoidance task in an obstacle distribution environment and the shortcut discovery task in a maze environment.

[0113] The embodiment of the present application also provides a local path planning system based on a spatial memory neural mechanism, including:

[0114] The recording unit controls the robot to explore a feasible path from the current position to the target point in the environment, and uses the laser radar carried by the robot to perceive the surrounding environment and record the information of the path points as the robot's memory information source;

[0115] A memory network unit is used to process the recorded information of the path points according to the organization hierarchical memory network to realize dynamic path planning after obtaining the information of the path points; wherein the organization hierarchical memory network includes grid cells, place cells and boundary cells;

[0116] The processing unit is used to initialize parameters, calculate the motion information from the current position to the target position, obtain the target direction and the reaching indication signal value; when the signal threshold is reached, it is considered that the robot has reached the vicinity of the target and the planning is completed; otherwise, the motion direction is selected.

[0117] For the specific limitations of the local path planning system based on the spatial memory neural mechanism, please refer to the limitations of the local path planning method based on the spatial memory neural mechanism above, which will not be repeated here. Each module in the above-mentioned local path planning system based on the spatial memory neural mechanism can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0118] In one embodiment, an electronic device is provided, which may be a computer. The electronic device includes a processor, a memory and a network interface connected via a system bus. The processor of the device is used to provide computing and control capabilities. The memory of the device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used for local path planning data based on a spatial memory neural mechanism. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a local path planning method based on a spatial memory neural mechanism is implemented.

[0119] In one embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned local path planning method are implemented.

[0120] In one embodiment of the present application, a computer program product is provided, including a computer program / instruction, and when the computer program / instruction is executed by a processor, the steps of the above-mentioned local path planning method are implemented.

[0121] The computer-readable storage medium and computer program product provided in this embodiment have similar implementation principles and technical effects to those of the above method embodiments, and are not described in detail here.

[0122] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in M ​​forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SyMchliMk) DRAM (SLDRAM), memory bus (RaMbus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0123] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0124] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.

Claims

1. A local path planning method based on spatial memory neural mechanism, characterized in that: The method comprises: Step 1: The robot is manually controlled to explore a feasible path from the current position to the target point in the environment, and the robot senses the surrounding environment through the laser radar carried by the robot, and records the information of the path points as the robot's memory information source; Step 2, after obtaining the information of the path points, the information of the recorded path points is processed according to the organizational hierarchical memory network to realize dynamic path planning; wherein the organizational hierarchical memory network includes grid cells, place cells and boundary cells; Step 3, initialize parameters, calculate the motion information from the current position to the target position, obtain the target direction and the indication signal value; when the signal threshold is reached, it is considered that the robot has reached the vicinity of the target and the planning is completed; otherwise, select the motion direction.

2. The local path planning method according to claim 1, characterized in that: The robot uses the laser radar to sense the surrounding environment and record the following information about the path points: The information of each waypoint is represented as a tuple, which contains the grid cell representation pattern, reward value and cognitive value; the grid cell representation pattern is the spatial coordinate of the waypoint, the reward value is generated based on the reverse memory playback of the target point, and the cognitive value represents the level of the memory node. The higher the level, the greater the cognitive value.

3. The local path planning method according to claim 1, characterized in that: The hierarchical memory network consists of grid cells, place cells, and border cells, including: Generate an organized hierarchical memory network based on the hippocampal navigation neural circuit structure; Grid cells are distributed in the medial entorhinal cortex and are used to provide a spatial coordinate system and perform vector navigation functions through the motor output module. Place cells are distributed in the CA region of the hippocampus and are used to connect synapses to support the storage and recall functions of memory, as well as to support the hierarchical organization and efficient search of memory at different spatial sparsities; Boundary cells are distributed in the lateral entorhinal cortex and are used to perceive local obstacle information and indicate whether a direction is navigable.

4. The local path planning method according to claim 3, characterized in that: Step 2, after obtaining the information of the path points, the information of the recorded path points is processed according to the organized hierarchical memory network to realize dynamic path planning, including: From dorsal to ventral direction, at fixed intervals o l The upper layer inherits the memory information of the exploration nodes in the lower layer and completes the abstract representation from fine to sparse. l The tuple of memory nodes is: From the ventral to the dorsal direction, at fixed intervals o l Connect the exploration nodes from the upper layer to the lower layer to complete the memory search from abstract to concrete. The connection weight from layer l to layer l-1 is:

5. The local path planning method according to claim 1, characterized in that: Step 3, initializing parameters, calculating the motion information from the current position to the target position, obtaining the target direction and the reaching indication signal value, including: Step 3.1, use the direction decoding module to calculate the motion information from the current position to the target position, obtain the target direction and the indication signal value. When the signal threshold is reached, it is considered that the robot has reached the vicinity of the target and the planning is completed; otherwise, select the motion direction; otherwise, go to step 3.1; Step 3.2, select the direction of movement, first determine whether the target direction is passable: s=B*T(θ target ) If s=1, the target direction θ is considered target It is passable, α=θ target , this strategy is called vector navigation strategy; conversely, if s = 0, the target direction θ target No access. The strategy is called the integrated navigation strategy; Step 3.3: Update the current position p according to the direction in step 3.2, and then go to step 3.1; Step 3.4: Calculate the direction of movement using the integrated navigation strategy.

6. The local path planning method according to claim 5, characterized in that: Step 3.4: Calculate the direction of movement using the integrated navigation strategy, including: Step 3.4.1: Calculate the direction of the memory node. For the memory pool of the first layer, calculate the direction of the memory node involved in the search, and use the search set I l Provides an indication that when l = 1, I l =[1,2,...,n E ]; otherwise, provided by step 3.4.3, the jth l The direction of each node is obtained by the direction decoding module: Step 3.4.2: Modulation of reward excitation-boundary inhibition to obtain the discharge state of the memory node in the first layer Then we get the memory node search vector Step 3.4.3: If l = 1, output the optimal direction Otherwise, perform a low-level search, searching set I l-1 for:

7. A local path planning system based on spatial memory neural mechanism, characterized in that: The system comprises: The recording unit controls the robot to explore a feasible path from the current position to the target point in the environment, and uses the laser radar carried by the robot to perceive the surrounding environment and record the information of the path points as the robot's memory information source; A memory network unit is used to process the recorded information of the path points according to the organization hierarchical memory network to realize dynamic path planning after obtaining the information of the path points; wherein the organization hierarchical memory network includes grid cells, place cells and boundary cells; The processing unit is used to initialize parameters, calculate the motion information from the current position to the target position, obtain the target direction and the reaching indication signal value; when the signal threshold is reached, it is considered that the robot has reached the vicinity of the target and the planning is completed; otherwise, the motion direction is selected.

8. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the local path planning method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the local path planning method as described in any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the local path planning method as described in any one of claims 1 to 6 is implemented.

Images