Indoor navigation and path planning method, device and equipment

Abstract

The invention provides an indoor navigation and path planning method, device and equipment. Comprising the steps that path planning request parameters are acquired, and the request parameters at least comprise a starting point, an ending point, a time context, a user portrait and a scene template of path planning; generating a runtime snapshot reflecting the traffic state at the current moment in combination with a pre-constructed static grid system, a knowledge graph and a dynamic grid system generated in real time; according to an indoor structure mask and guide identification information in the static grid system, generating a search domain limit of path planning on the runtime snapshot; judging whether the starting point and the ending point are on the same floor; performing path search by adopting a same-layer search strategy or a cross-layer search strategy; and dynamically injecting constraints to the result of the path search and calculating the path cost to generate a path plan. According to the embodiment of the invention, the indoor space can be uniformly expressed, the multi-source dynamic information can be fused in real time, and the path planning scheme can be quickly generated, so that the navigation requirement in a modern complex indoor scene can be met.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent navigation, in particular to an indoor navigation and path planning method, device and equipment suitable for multi-layer complex buildings. BACKGROUND

[0002] With the complication of large-scale shopping malls, hospitals, airports, smart factories and other buildings, indoor path planning has become a key technology to improve operational efficiency and user experience. Existing methods mainly include graph-based methods (such as Dijkstra, A* algorithm), grid method, topological semantic method and image analysis-based path planning method.

[0003] However, the above methods have the following limitations: the graph theory method can find the shortest path, but the graph construction is complex and difficult to update dynamically; the grid method consumes a lot of memory and has poor path smoothness; the topological semantic method is difficult to adapt to the flexible path demand in open space; the image analysis-based method is heavily dependent on the quality of the base map and lacks dynamic expansion capability. These methods generally have the problems of non-uniform space expression, difficulty in dynamic information fusion, high computational complexity and output path not meeting human walking habits. SUMMARY

[0004] The purpose of the present application is to provide an indoor navigation and path planning method, device and equipment suitable for multi-layer complex buildings, which can uniformly express indoor space, real-time fuse multi-source dynamic information, and quickly generate path planning scheme to meet the navigation needs in modern complex indoor scenes.

[0005] To solve the above technical problems, the present application provides an indoor navigation and path planning method, which comprises the following steps: step S1, obtaining path planning request parameters, the request parameters at least including the starting point, the ending point, the time context, the user portrait and the scene template of path planning; Step S2, according to the request parameters, combining the pre-constructed static grid system and knowledge graph, and the real-time generated dynamic grid system, generating a runtime snapshot reflecting the current traffic state; the dynamic grid system is generated by the data of the multi-source sensing device acquired in real time, the multi-source sensing device at least including: Internet of Things device, camera and radar; Step S3, generating the search domain limit of path planning on the runtime snapshot according to the indoor structure mask and guide sign information in the static grid system; Step S4, judging whether the starting point and the ending point are on the same floor; Step S5, if yes, a same-floor search strategy is adopted to perform path search within the search domain limit; otherwise, a floor connection graph containing elevators, escalators, stairs and ramps is established within the search domain limit, and a cross-floor search strategy is adopted to perform path search on the floor connection graph to obtain an initial path; Step S6, constraints are dynamically injected into the initial path according to the user portrait, time context and knowledge graph, and path cost is calculated to generate the path planning.

[0006] Another aspect of the embodiments of the present application also provides an indoor navigation and path planning device, comprising: A request parameter acquisition module is configured to acquire path planning request parameters, the request parameters at least including a starting point, an ending point, a time context, a user portrait and a scene template of path planning; A runtime snapshot generation module is configured to generate a runtime snapshot reflecting a current traffic state according to the request parameters, in combination with a pre-constructed static grid system and a knowledge graph, and a dynamically generated dynamic grid system; the dynamic grid system is generated from data of a plurality of source sensing devices acquired in real time, the plurality of source sensing devices at least including Internet of Things devices, cameras and radars; A search domain limit generation module is configured to generate a search domain limit for path planning on the runtime snapshot according to an indoor structure mask and guide sign information in the static grid system; A judgment module is configured to judge whether the starting point and the ending point are on the same floor; A path search module is configured to, when the starting point and the ending point are on the same floor, adopt a same-floor search strategy to perform path search within the search domain limit; when the starting point and the ending point are not on the same floor, establish a floor connection graph containing elevators, escalators, stairs and ramps within the search domain limit, and adopt a cross-floor search strategy to perform path search on the floor connection graph to obtain an initial path; A path planning generation module is configured to dynamically inject constraints into the initial path according to the user portrait, the time context and the knowledge graph, and calculate path cost to generate the path planning.

[0007] The embodiments of the present application also provide an electronic device, comprising a processor, a memory, and a computer program stored on the memory and capable of running on the processor, the processor implementing the steps of the indoor navigation and path planning method described above when executing the computer program.

[0008] Compared with the prior art, the embodiment of the application can respond to environmental changes in seconds by fusing static, dynamic and knowledge graph data in real time through a "runtime snapshot" mechanism and using an incremental re-planning algorithm. Combined with a pre-set knowledge graph and a user portrait, the path planning can understand business rules, user preferences and real-time status, and automatically select the optimal traffic strategy. The intelligent path planning realizes cross-floor integration, real-time dynamic response, and the fusion of user preferences and business rules. BRIEF DESCRIPTION OF DRAWINGS

[0009] One or more embodiments are illustrated by way of example in the figures that are part of this document, and which illustrate by way of example, not limitation, the principles of the embodiments. Similar reference numbers in different drawings represent similar elements, unless otherwise indicated. The drawings in which:

[0010] Figure 1 is a flowchart of an indoor navigation and path planning method according to an embodiment of the application; Figure 2 is a structural schematic diagram of an indoor navigation and path planning device according to an embodiment of the application; Figure 3 is a structural schematic diagram of an electronic device according to an embodiment of the application. DETAILED DESCRIPTION

[0011] The embodiments of the present application will be described in detail with specific reference to particular embodiments, some examples of which are illustrated in the appended drawings. Obviously, the embodiments described are only a part of the embodiments of the present application, but not all the embodiments. The present application can be implemented or applied in other different specific embodiments, and the details in the specification can be modified or changed based on different views and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0012] It should be noted that the following description covers various aspects of embodiments within the scope of the appended claims. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this application, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number and aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0013] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0014] Additionally, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that practice can be carried out without these specific details.

[0015] like Figure 1 As shown in the figure, this application provides an indoor navigation and route planning method, the specific process of which is as follows: In step 101, path planning request parameters are obtained. These parameters include at least the starting point, ending point, time context, user profile, and scene template for path planning. Specifically, the path planning request parameters in this embodiment can be input by the user. They mainly include: Start and end points: given in the building’s local coordinate system, for example: the number of rooms on the number of floors (FloorIDg,xg,yg). Time context: Used to determine whether access control time period, escalator direction, and temporary closure are valid at the current time, such as t_now; User profiles: These can be matched with user preferences and used to translate them into hard constraints or cost weights in subsequent steps, such as: avoid taking stairs / disable stairs / prefer elevators / wheelchairs, etc., profile={avoid_stairs,prefer_elevator,wheelchair,max_walk, …}; Scene templates: used to select the corresponding soft constraint strategy from different default rule groups, such as shopping malls, hospitals, airports, factories, etc. Coordinate validity check: According to the floor boundary recorded in the floor mapping table, check whether the starting point / endpoint falls within the corresponding floor range; if not, it will be adsorbed to the nearest passable unit in the subsequent domain limit generation step.

[0016] In step 102, according to the request parameters, combined with the pre-constructed static grid system and knowledge graph, and the real-time generated dynamic grid system, the runtime snapshot reflecting the current time traffic state is generated; the dynamic grid system is generated by the data of the multi-source sensing device acquired in real time, and the multi-source sensing device at least includes: Internet of Things device, camera and radar.

[0017] Specifically, the construction method of the static grid system and the dynamic grid system in the embodiment includes: Obtain static data and dynamic data; wherein, the static data at least includes indoor structure (such as wall, door, room, corridor), traffic facilities (such as elevator, escalator, staircase, ramp) and floor mapping FloorMapping={FloorID,bounds,level}; the static data usually comes from CAD / BIM / IFC or indoor GIS, which is relatively fixed spatial information. Dynamic data at least includes: access control period TW_door=[t_open,t_close), escalator direction and period, crowd density p(x,y,t), temporary closed / construction area C_block(t) and Internet of Things (IoT) alarm, etc., which is the running data changing with time.

[0018] Based on two-dimensional space-time grid and floor identification, a unified coding key is generated for indoor grid unit; according to the unified coding key, the static data and dynamic data are encoded respectively to obtain the static grid system and the dynamic grid system; by generating the unified coding key, the static data and dynamic data with different sources, different coordinate systems and different semantic levels can be uniformly converted into the unique identification key of indoor two-dimensional grid+floor, which can open up the data path of subsequent steps. The specific method is as follows: (1) According to the floor mapping table of the building, determine the space boundary and level parameter of each floor; according to the coordinate range [x_min, x_max] x [y_min, y_max] and the target level L given in the floor mapping table, the local coordinates (x, y) are quantized to integer grid index according to the predetermined resolution:

[0019] (2) The two-dimensional Morton curve interleaving method is used to combine (i, j) into a plane grid code CellID=interleave2(i, j, L); (3) A unified encoding key is formed by the floor identifier FloorID, the plane grid code CellID, and the level L: Key = (FloorID, CellID, L). All static grids (walls, rooms, doors, corridors), all dynamic grids (temporary closures, crowded areas), and all connected facilities (elevators, escalators, stairs, ramps) are stored and accessed by the key.

[0020] (4) Through the above method, the "correspondence between two-dimensional space-time coding and floor mapping" is realized. The floor mapping provides the coding boundary, and the two-dimensional space-time coding provides the spatial locality coding. After binding, the unified space-time positioning of multi-source heterogeneous data (IOT, camera, CAD, maintenance record, etc.) can be realized, and the spatial landing point of the semantic result of the knowledge graph is also provided. After the coding is completed, the system checks whether each grid is located within the floor range, and according to the building structure (room, corridor, door, obstacle), it gives the grid a passing or shielding attribute.

[0021] The embodiment of the present application fundamentally solves the fusion and interoperability problem of multi-source heterogeneous data by mapping the static and dynamic data of different sources and formats to the same spatial reference system based on the unified encoding key of the two-dimensional space-time coding grid and the floor identifier. Abstract entities such as walls, doors, and pedestrian flow are converted into grid units with clear spatial positions and attributes, providing a standardized data foundation for subsequent automated path search and cost calculation.

[0022] The runtime snapshot of the embodiment of the present application includes: a set of non-passable grid units BlockedCells Cells, a set of non-passable or direction-restricted connection edges BlockedEdges Edges, and grid units that need to increase the passing cost ExtraCostCell:Cell and connection edges ExtraCostEdge:(u,v).

[0023] The knowledge graph of the embodiment of the present application can be pre-constructed by an enterprise or management party in relation to the scene. The knowledge graph has pre-set semantic relationships related to doors, rooms, corridors, access control rules, open periods, emergency evacuation strategies, barrier-free strategies, and device operating states. The semantic relationships are bound to the corresponding grid unit codes or connection edges. This step triggers graph reasoning according to t_now and dynamic data, and binds the reasoning result to the known grid unit encoding key value (FloorID, CellID, L) or connection edge.

[0024] By pre-setting the knowledge graph and binding the semantic relationship to the grid cell, the system can not only understand the geometric structure of the space, but also understand the business rules behind it (such as access time, facility purpose), thereby realizing the leap from "geometric accessibility" to "semantic feasibility". The runtime snapshot integrates static blocking, dynamic blocking and semantic constraints to generate a comprehensive, accurate and real-time "passability map", providing a unique and reliable data source for search algorithms and ensuring the accuracy of decision-making.

[0025] In step 103, according to the indoor structure mask and guide sign information in the static grid system, the search domain limit for path planning is generated on the runtime snapshot. The indoor structure mask is a binary map used to distinguish the passable and impassable areas of the indoor space. Specifically, the specific steps of generating the search domain limit in this step include: Perform morphological dilation on the binary mask M_struct (binary passable / non-passable area) of the indoor structure mask according to the corridor width w_c to obtain the walkable band M_walk. Grid the indoor guide information (direction signs, evacuation arrows, information that must pass through stairs / elevators) into a guide mask M_signage. Intersect the walkable band and the guide mask in the same encoding grid to obtain the search domain limit. For example: Ω = M_walk∩M_signage If M_signage is sparse, it can be degenerated to Ω = M_walk.

[0026] In step 104, it is judged whether the starting point and the ending point are on the same floor. In step 105, a same-floor search strategy is used to search for a path in the search domain limit to obtain an initial path. Specifically, if the starting point and the ending point are on the same floor, this step is executed. The same-floor search strategy includes coarse skeleton search and local refinement search.

[0027] Coarse skeleton search: Project the starting point and the ending point to a lower level grid and search in the search domain limit to obtain a skeleton path. For example: In the same floor and the same search domain limit Ω, project the encoding of the starting point and the ending point from the target high level L to a lower level L_c (for example, L_c = L-3). Search using A* / Dijkstra on this level to obtain a skeleton path that covers the main corridors and the direction of travel.

[0028] Local refinement search: the skeleton path is divided into several segments, and a fine search is performed within the search domain limit of the target high-level grid for each segment. Meanwhile, constraints such as doors, passages, congestion, and obstacles can also be added to obtain an accurate grid path. For example, two adjacent coarse grid segments on the skeleton path are regarded as a segment, and the A* search is performed again in the small segment range to obtain an accurate grid path, i.e., the initial path.

[0029] In this step, whether it is a coarse skeleton search or a local refinement search, the search range is limited within the search domain limit Ω; if the area passed by the coarse search path is dynamically shielded in the refinement layer, only the local re-search is performed for the segment; after the above two stages are completed, the output is a sequence of grid cells expressed in level L, which has been aligned with the search domain limit. This sequence will be directly used as the input of the path planning. The hierarchical search strategy of "coarse first and then fine" is adopted, that is, the global shortest path trend control (skeleton path) is quickly determined at low resolution, and then fine carving is performed at high resolution to ensure the passability and local optimality of the path. Finally, the two layers of results are combined to output the complete path balancing the "shortest distance + optimal experience". At the same time, the problem of large search space and time-consuming calculation in high-precision grid is effectively solved.

[0030] In step 106, a floor connection graph containing elevators, escalators, stairs, and ramps is established within the search domain limit, and a cross-layer search strategy is used to search the floor connection graph to obtain an initial path. Specifically, if the starting point and the ending point are on different floors, this step is executed. The connection edges of the floor connection graph in this step at least include directionality, waiting time, reachable floor set, and non-obstacle attribute.

[0031] When constructing the floor connection graph GF=(V,E), each node in the graph represents a floor or a local area of a floor, and the edge represents a vertical connection that can be passed. Each connection edge e(u,v) has a direction flag dir∈{one-way, two-way}. Facilities such as escalators, one-way passages, and turnstiles are marked as one-way edges in the static data stage; if the escalator has a fixed reverse direction (e.g., every hour), the direction attribute is updated by the time context t_now in the dynamic data. During the search process, only edges with consistent directions are allowed to be expanded to prevent reverse travel or illegal travel.

[0032] The modeling method of waiting time is as follows: Facilities such as elevators, security checkpoints, and turnstiles have a waiting time attribute wait_time in the knowledge graph, which comes from historical statistics or real-time queue length.

[0033] In the runtime snapshot stage, this attribute is mapped to the weight wtw_twt of the connection edge, and the distance cost wdw_dwd is included in the comprehensive cost function.

[0034] When a connection with too long waiting time is encountered during search, it can be adaptively avoided by cost function.

[0035] The modeling method of accessible floors is as follows: Each connection facility contains a set of AccessibleFloors, indicating the range of accessible floors.

[0036] The cross-layer edge set Elift is automatically generated during modeling, and only edge relationships between adjacent and accessible floors are established; direct access to unauthorized floors by elevators is prohibited.

[0037] The modeling method of accessibility attributes is as follows: The connection facility has an attribute Accessible={True, False} in the knowledge graph. For wheelchair or mobility-impaired users, the system will exclude edges with Accessible=False from the candidate set during the constraint injection stage. At the same time, if the accessible path has shorter waiting time or the elevator is available, the algorithm will automatically prefer this path.

[0038] The cross-layer search strategy in this step is as follows: (1) Based on the floor connection graph GF, the Dijkstra algorithm is used to calculate the optimal floor sequence; (2) For each pair of adjacent floors, determine the connection point (such as elevator entrance, stair entrance), and perform detailed path planning in the respective coding domain; (3) After completing the local path for each floor, update the starting point of the next floor to the exit unit of the corresponding connection point; (4) Until the planning reaches the target floor end point, the final global path is formed by merging each floor segment path, which is the initial path.

[0039] Through the above mechanism, directionality, waiting time, accessible floors, and accessibility attributes are unified into the spatial connection model (i.e., the floor connection graph), and through the mapping link of knowledge graph→coding domain→search domain, consistency and dynamic updating ability are maintained throughout the path planning process.

[0040] In step 107, constraints are dynamically injected into the initial path according to the user profile, time context and knowledge graph, and the path cost is calculated to generate a path plan. Specifically, the constraint injection in this embodiment includes two ways of modeling period constraint injection and running period constraint injection: the modeling period constraint injection includes constraints that have been determined, which are independent of time and users (such as permanently closed areas, structural impassable areas, inter-floor connections without stairs, and fixed one-way escalators), which are directly written into the form of BlockedCells / BlockedEdges in the runtime snapshot or search domain limit, so that subsequent searches will not access these cells. The running period constraint injection is used in the node expansion stage of the same layer search, and the connection edges that do not meet the direction, the inaccessible floor connection, or the connection edges that do not have the barrier-free attribute are removed or increased in cost according to the user profile, the time context t_now and the knowledge graph reasoning result. For example: for wheelchair users, remove the stairs edges that do not have the barrier-free attribute; for cells with high congestion, increase the ExtraCostCell; for doors that will be closed soon, increase the time cost or directly close; set different distance, time and comfort weights according to the scene template, etc.

[0041] In the embodiment of the present application, the function of calculating the path cost can be represented as: Cost = w_d dist + w_t time + w_p crowd + w_u user_pref + w_s signage Wherein, dist is the distance cost (i.e. the Euclidean or Manhattan distance between adjacent grid center points), time is the time cost (including the estimated time of waiting time), crowd is the congestion degree cost (the congestion cost obtained by mapping the crowd density), user_pref is the user preference cost (the preference cost obtained by the user profile), signage is the guide sign cost (reward / punishment item when preferentially walking along the guide corridor), and w_* is the non-negative weight corresponding to each cost item (non-negative weight, which can be obtained according to the scene template or learning method).

[0042] The embodiments of the present application introduce "runtime snapshot" and "user portrait / scene template", so that the path planning can closely combine the dynamic environment at the current time and the personalized needs of the user, output the optimal path that is truly feasible and personalized for a specific user in a specific scene at a specific time, and overcome the rigid shortcomings of static planning models. Through the explicit same-layer and cross-layer search strategy, and the establishment of a floor connection graph containing multiple types of facilities, the physical connectivity of the path in a complex multi-story building is ensured, and in particular, the actual situation of switching floors through vertical facilities can be effectively handled. Moreover, the system architecture of the embodiments of the present application is modularized, easy to integrate new data sources, rules or algorithms, has good scalability, and can adapt to various complex indoor scenes such as shopping malls, hospitals, airports and factories.

[0043] In a preferred embodiment, before dynamically injecting constraints to the initial path and generating the path planning, the initial path can also be smoothed and checked for avoiding obstacles and complying with traffic rules. For example, by detecting the angle and length of adjacent path segments, collinear points can be eliminated, angle points can be compressed, and turning angles can be limited to comply with the natural characteristics of human walking or wheelchair turning.

[0044] Specifically, the angle between the vectors formed by three consecutive points is calculated, if the angle is close to 180°, the middle point is deleted; if the angle is too small and the traffic area allows, a circular arc or a gentle bend is generated by interpolation. At the same time, the length of the path segment is detected to avoid meaningless short return segments. The optimized path is then checked for feasibility to ensure that it does not cross walls or forbidden areas, and meets the hard constraints such as access control and direction.

[0045] More preferably, the path planning can also be converted into semantic navigation instructions containing turning and floor switching. For example, first, identify the key nodes in the path, such as floor change points, turning points, doors or passageway openings; then calculate the direction change and distance according to the current orientation, and generate semantic instructions such as "go straight for x meters", "turn left", "take the elevator to x floor", "pass through x door", etc.; if a floor change is detected, generate a corresponding floor prompt and estimated time of arrival (ETA). Finally, output the multi-modal navigation results in the form of voice, text or graphics.

[0046] Through path smoothing and semantic navigation instruction generation, the path results that comply with human walking habits and are easy to understand are output, greatly improving the user experience.

[0047] In another preferred embodiment, the environment changes can also be monitored in real time during the path planning process, and an incremental search algorithm can be used to locally re-plan the path segments affected. The incremental search algorithm can be LPA or DLite algorithm, which maintains the actual cost and lookahead cost of nodes, and updates the cost of affected nodes when the environment changes to achieve the local re-planning. Specifically, the environment changes (such as construction, sudden congestion, access control switching, escalator direction change, etc.) and user deviation can be monitored in real time during the path planning process. When the event is triggered, the incremental algorithm such as LPA / DLite is used to update only the affected area to achieve local re-planning. The specific method is as follows: G(n): the current optimal actual cost from the starting point to node n; RHS(n): one-step lookahead cost to reach n based on the current adjacency relationship; key(n): key value for priority queue sorting, usually defined as: key(n)=(min(G(n),RHS(n))+h(n,goal),min(G(n),RHS(n))), where h(n,goal) is a heuristic function that satisfies the consistency condition, which can be "same layer Manhattan distance + cross layer connection cost"; km: incremental offset when the environment changes, used to maintain the correctness of the queue order; UpdateVertex(n): called when the passability / cost of node n or its adjacent nodes changes, used to recalculate RHS(n) and re-insert n into the priority queue; Termination condition: when the G value of the target node is consistent with the RHS value, and the minimum key in the priority queue is not better than the target node, the local repair is completed.

[0048] The embodiment can quickly correct the path when the user deviates or the environment changes (such as construction, congestion) by monitoring the environment changes in real time during the planning process and triggering local re-planning, avoiding the calculation delay caused by the need to completely re-plan in the traditional method, and ensuring the continuity and real-time performance of navigation. By maintaining and updating the cost information of nodes, the incremental search algorithm such as LPA or DLite can efficiently repair only the affected area, and ensure that the re-planned path is still the optimal solution under the current environment, achieving the best balance in efficiency and effect.

[0049] The following illustrates a specific scenario.

[0050] In an application scenario of shopping mall navigation : First, collect the CAD plan of the mall and the information of elevators, escalators, and stairs, etc. to construct the static data. At the same time, connect the IOT access control system and people flow counting equipment to obtain dynamic data. The knowledge graph contains information such as shops on each floor, directional signs, barrier-free passages, and opening hours of the mall.

[0051] Then generate the floor mapping table and two-dimensional space-time coding according to the CAD map, and generate the walkable zone by morphological dilation. Combine the directional signs to form the domain limit. Write the elevator directionality, waiting time, and accessible floor attributes into the connection graph through the knowledge graph.

[0052] The user selects the starting point as the parking lot entrance and the end point as a certain brand counter. The system identifies the user's preference as "avoid escalators". In the runtime snapshot, the escalator edges are shielded, and only elevator and stair connections are retained.

[0053] Layering first determines the main corridor skeleton in the coarse layer, and then searches for specific passages in the fine layer.

[0054] After path smoothing, generate semantic instructions: "straight ahead 20 meters - take the elevator to the 2nd floor - turn right 10 meters - arrive at the destination".

[0055] In an application scenario of hospital navigation : Collect the BIM map of the hospital building and IOT sensor data. Static data includes wards, outpatient areas, nurse stations, stairs, elevators, etc.

[0056] Dynamic data includes patient flow, construction closed areas, and emergency passage status.

[0057] The user profile is a wheelchair patient, and the system excludes stair connections with a barrier-free attribute of False during the constraint injection phase.

[0058] Perform layering search within the domain limit, and prefer barrier-free elevators.

[0059] Output the path and voice prompt: "Take the barrier-free elevator to the third floor, turn left into the ward corridor, and the end point is at the door of the clinic on the left." In an application scenario of airport navigation Collect the structure of the terminal building, boarding gate location, and security checkpoint layout. Dynamic data includes real-time flight information, security checkpoint queue length, and escalator direction.

[0060] The system records the security rules and time period restrictions of each passage in the knowledge graph.

[0061] According to the passenger's flight time, the system automatically selects the optimal boarding path: if a security checkpoint is crowded, increase the crowd weight in the cost function to avoid it.

[0062] Generated instruction: "Go straight for 100 meters - pass through security gate A - take the elevator to the boarding gate layer - turn right to B12 boarding gate, expected time 8 minutes." In an application scenario of factory navigation (emergency escape) Static data comes from factory CAD and workshop layout; dynamic data comes from gas sensors, cameras and access control systems.

[0063] When a workshop detects a harmful gas leak, the system updates the runtime snapshot in real time, marking the area unit as BlockedCells.

[0064] Perform incremental re-planning of this method for all personnel in the factory, and only recalculate the path for the affected area.

[0065] Output voice prompt: "There is a leak in the current workshop, please go straight along the north side of the channel for 50 meters, and turn right into the safety exit B." The whole process is updated within a few seconds to realize real-time escape path planning.

[0066] The embodiments of the present application adopt system architecture modularization, which is easy to integrate new data sources, rules or algorithms, has good scalability, and can adapt to various complex indoor scenes such as shopping malls, hospitals, airports and factories.

[0067] Based on the same inventive concept, the embodiments of the present application also provide an indoor navigation and path planning device. It should be noted that the device exemplified below is an example of the device corresponding to the above-mentioned one method embodiment, and in other device embodiments, the functions and number of module settings in the unit module can be set according to the above-mentioned method embodiments.

[0068] As shown in Figure 2 The indoor navigation and path planning device comprises: A request parameter acquisition module 1 is configured to acquire path planning request parameters, which at least include the starting point, end point, time context, user portrait and scene template of path planning; A runtime snapshot generation module 2 is configured to generate a runtime snapshot reflecting the current time traffic state according to the request parameters, in combination with a pre-constructed static grid system and knowledge graph, and a dynamically generated dynamic grid system; the dynamic grid system is generated by data of a plurality of source sensing devices acquired in real time, and the plurality of source sensing devices at least include Internet of Things devices, cameras and radars; A search domain limit generation module 3 is configured to generate a search domain limit for path planning on the runtime snapshot according to the indoor structure mask and guide sign information in the static grid system; A judgment module 4 is configured to judge whether the starting point and the end point are on the same floor. The path searching module 5 is configured to, when the start point and the end point are on the same floor, search for a path within the search domain limit by using a same-floor searching strategy; when the start point and the end point are not on the same floor, establish a floor connection graph containing elevators, escalators, stairs and ramps within the search domain limit, and search for a path on the floor connection graph by using a cross-floor searching strategy to obtain an initial path. The path planning generation module 6 is configured to dynamically inject constraints into the initial path according to the user portrait, the time context and the knowledge graph, and calculate a path cost to generate a path plan.

[0069] As shown in Figure 3 The electronic device provided in the embodiments of the present application includes a processor 12, a memory 13, and a computer program stored in the memory 13 and capable of running on the processor 12, and the processor 12 implements the steps of the indoor navigation and path planning method described above when executing the computer program.

[0070] The electronic device in the embodiments of the present application includes at least one processor 12 and a memory 13 communicatively connected to the at least one processor 12, Figure 3 The processor 12 is taken as an example for illustration.

[0071] The electronic device can further include an input device 14 and an output device 15.

[0072] The processor 12, the memory 13, the input device 14 and the output device 15 can be connected by a bus or other means, Figure 3 The connection by the bus is taken as an example.

[0073] The memory 13 as a non-transitory computer readable storage medium can be used to store non-transitory software programs, non-transitory computer executable programs and modules, such as program instructions / modules (for example, the request parameter acquisition module 1, the runtime snapshot generation module 2, the search domain limit generation module 3, etc. Figure 2 The processor 12 executes various functional applications and data processing of the server by running the non-transitory software programs, instructions and modules stored in the memory 13, that is, implements the indoor navigation and path planning method in the method embodiments.

[0074] The memory 13 can include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application required by a function; and the data storage area can store data created by the processing device according to the list operation and the like. In addition, the memory 13 can include a high-speed random access memory, and can also include a non-transitory memory, such as at least one disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory 13 can optionally include a memory disposed remotely relative to the processor 12, and these remote memories can be connected to the processor 12 for the indoor navigation and path planning method through a network. Examples of the above network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0075] The input device 14 can receive inputted digital or character information, and generate key signal inputs related to user settings and function controls of the processing device according to the list operation. The output device 15 can include a display device such as a display screen.

[0076] In the embodiment, when one or more modules stored in the memory 13 are executed by the one or more processors 12, the processor 12 executes the indoor navigation and path planning method in any method embodiment described above.

[0077] The above product can execute the method provided in the embodiments of the present application, and has the corresponding function modules and beneficial effects of executing the method. Technical details not described in detail in the embodiments can be referred to the method provided in the embodiments of the present application.

[0078] The electronic device of the embodiments of the present application exists in various forms, including but not limited to: (1) Mobile communication device: the feature of this kind of device is to have mobile communication function, and to provide voice and data communication as the main target. This kind of terminal includes: smart phone (such as iPhone), multimedia phone, functional phone, and low-end phone, etc.

[0079] (2) Ultra-mobile personal computer device: this kind of device belongs to the category of personal computer, has computing and processing functions, and generally also has the feature of mobile Internet. This kind of terminal includes: PDA, MID and UMPC device, etc., such as iPad.

[0080] (3) Portable entertainment device: this kind of device can display and play multimedia content. This kind of device includes: audio and video player (such as iPod), handheld game console, electronic book, and smart toy and portable car navigation device.

[0081] Any technical features in the above-described embodiments can be combined in any manner, and for the sake of brevity, not all possible combinations are described, however, as long as there is no conflict in combining the technical features, it should be considered that the combination is within the scope of the present disclosure.

[0082] The above-described embodiments only express several implementation manners of the present application, and the description is specific and detailed, but it should not be understood as a limitation on the patent scope of the application. It should be pointed out that for ordinary skilled persons in the art, a number of modifications and improvements can be made without departing from the concept of the present application, and these are within the protection scope of the present application. Therefore, the patent protection scope of the present application should be subject to the appended claims.

[0083] Ordinary skilled persons in the art can understand that the above-described embodiments are specific embodiments for implementing the present application, and in actual application, various changes can be made in form and details without departing from the spirit and scope of the present application.

Claims

1. A method of indoor navigation and path planning, characterized in that, The method comprises the following steps: Step S1, obtaining path planning request parameters, the request parameters at least including a starting point, an ending point, a time context, a user portrait and a scene template of path planning; Step S2, generating a runtime snapshot reflecting a current traffic state according to the request parameters, combining a pre-constructed static grid system and a knowledge graph, and a dynamically generated dynamic grid system; The dynamic grid system is generated from data of a plurality of real-time acquisition of multi-source sensing devices, the multi-source sensing devices at least including Internet of Things devices, cameras and radars; Step S3, generating a search domain limit of path planning on the runtime snapshot according to indoor structure masks and guide sign information in the static grid system; Step S4, judging whether the starting point and the ending point are on the same floor; Step S5, if yes, adopting a same-floor search strategy to search for a path in the search domain limit; otherwise, establishing a floor connection graph containing elevators, escalators, stairs and ramps in the search domain limit, and adopting a cross-floor search strategy to search for a path on the floor connection graph to obtain an initial path; Step S6, dynamically injecting constraints into the initial path and calculating a path cost according to the user portrait, the time context and the knowledge graph to generate the path planning.

2. The method of claim 1, wherein, Before the initial path is dynamically injected with constraints and the path planning is generated, the method further comprises: Smoothing the initial path and checking whether the initial path avoids obstacles and conforms to traffic rules.

3. The method of claim 1, wherein, The method further comprises: monitoring environmental changes in real time during the path planning process, and adopting an incremental search algorithm to locally re-plan path segments affected by the environmental changes.

4. The method of claim 3, wherein, The incremental search algorithm is an LPA or DLite algorithm, which updates the cost of affected nodes to achieve the local re-planning by maintaining the actual cost and the lookahead cost of nodes when the environment changes.

5. The method of claim 1, wherein, The construction method of the static grid system and the dynamic grid system comprises: Obtaining static data and dynamic data; Generating a uniform coding key for indoor grid cells based on a two-dimensional space-time grid and floor identification; Encoding the static data and the dynamic data according to the uniform coding key to obtain the static grid system and the dynamic grid system; The static data at least includes walls, doors, rooms, corridors, elevators, escalators, stairs, ramps and floor mappings; The dynamic data at least includes access control time periods, escalator directions and periods, crowd density, temporary closed / construction areas and Internet of Things device alarms.

6. The method of claim 1, wherein, The runtime snapshot in step S2 includes a set of non-passable grid cells, a set of connection edges with limited directions or directions, and grid cells and connection edges that need to increase the traffic cost; The knowledge graph is pre-provisioned with semantic relationships related to doors, rooms, corridors, access control rules, open time periods, emergency evacuation strategies, barrier-free strategies and device operating states, which are bound to the codes of corresponding grid cells or connection edges.

7. The method of claim 1, wherein, The specific steps of generating the search domain limit in step S3 include: Performing morphological dilation on a binary mask of the indoor structure mask according to the corridor width to obtain a walkable band; Rasterizing indoor guidance information into a guidance mask; Intersecting the walkable area with the guidance mask to obtain a search domain limit.

8. The method of claim 1, wherein, The same-floor search strategy in step S5 includes: Projecting the start point and the end point to a lower-level grid, searching within the search domain limit to obtain a skeleton path; Segmenting the skeleton path into several segments, and performing refined search on each segment within the search domain limit of the target high-level grid to obtain an accurate grid path.

9. The method of claim 1, wherein, The function for calculating the path cost in step S6 is: Cost = w_d dist + w_t time + w_ρ crowd + w_u user_pref + w_s signage where dist is the distance cost, time is the time cost, crowd is the congestion degree cost, user_pref is the user preference cost, signage is the guidance sign cost, and w_* is the non-negative weight corresponding to each cost item.

10. The method of claim 1, wherein, The connection edges of the floor connection graph in step S5 at least include directionality, waiting time, reachable floor set, and barrier-free attribute.

11. The method of claim 1, wherein, When calculating the path cost in step S6, it further includes: according to the user portrait, time context, and knowledge graph, removing or increasing the cost of the connection edges that do not conform to the direction, the floor connections that are not reachable, or the connection edges that do not have the barrier-free attribute.

12. The method of claim 1, wherein, After step S6, it further includes: converting the path planning into semantic navigation instructions containing turning and floor switching.

13. An indoor navigation and path planning apparatus, characterized by, It includes: A request parameter acquisition module for acquiring path planning request parameters, the request parameters at least including the start point and the end point of path planning, time context, user portrait, and scene template; A runtime snapshot generation module for generating a runtime snapshot reflecting the current traffic state according to the request parameters, combining the pre-constructed static grid system and knowledge graph, and the real-time generated dynamic grid system; The dynamic grid system is generated from data of multiple source sensing devices acquired in real time, the multiple source sensing devices at least including: Internet of Things devices, cameras, and radars; A search domain limit generation module for generating a search domain limit for path planning on the runtime snapshot according to the indoor structure mask and guidance sign information in the static grid system; A judgment module for judging whether the start point and the end point are on the same floor; A path search module for, when the start point and the end point are on the same floor, performing path search within the search domain limit by using a same-floor search strategy; and for, when the start point and the end point are not on the same floor, establishing a floor connection graph containing elevators, escalators, stairs, and ramps within the search domain limit, and performing path search on the floor connection graph by using a cross-floor search strategy to obtain an initial path; A path planning generation module for dynamically injecting constraints into the initial path and calculating the path cost according to the user portrait, time context, and knowledge graph to generate the path planning.

14. An electronic device, comprising a processor, a memory, and a computer program stored on the memory and capable of running on the processor, characterized in that, The processor implements the steps of the indoor navigation and path planning method as claimed in any one of claims 1-12 when executing the computer program.

Images