Indoor light field adaptive tracking control method and system

Abstract

The invention provides an indoor light field adaptive tracking control method. The method comprises the following steps: acquiring real-time track data of a target user; determining an inertial behavior mode of the target user according to the historical behavior data of the target user; based on the real-time trajectory data, the inertial behavior pattern and the spatial topology distribution diagram of each indoor area, predicting a target indoor area where the target user is located in the next time period, and determining the target indoor area as a condensation area; and generating a light field control instruction so as to control the illumination unit corresponding to the condensation area to be improved to the working brightness level. According to the method, the target indoor area where the target user is located in the next time period can be actively predicted by combining the real-time trajectory data, the inertial behavior pattern and the spatial topology information of the target user, so that the problem of illumination delay from entering a dark area to lighting of a lamp is solved, seamless and continuous light advance experience is provided for the user, and the user experience is improved. And the operation convenience is obviously improved.

Description

Technical Field

[0001] This invention relates to the field of smart home technology, and in particular to an indoor light field adaptive tracking control method and system. Background Technology

[0002] With the intelligent development of the furniture industry, smart indoor lighting has been gradually applied to various hotel scenarios, becoming an important support for improving hotel operational efficiency, optimizing service experience, and reducing energy consumption. In practical applications, hotels, especially larger ones, typically have multiple independent rooms, and hotel managers need to regularly conduct cleaning, security checks, and other routine tasks in each room. These routine tasks are characterized by high mobility and dispersed work areas, and require sufficient lighting to ensure cleaning accuracy and comprehensive security checks.

[0003] Currently, the indoor smart lighting fixtures used in hotel settings are mainly controlled by sound and light sensors. The control principle of these sensors is to trigger the light switch by detecting the sound intensity in the environment (such as the footsteps of management personnel or cleaning sounds) in low light conditions, thereby turning on the lights for illumination. When the sound intensity in the environment is low for a period of time, the lights will automatically turn off.

[0004] However, the sound and light sensor control method is a passive response trigger, and its lamp lighting has a significant lag. For example, it needs to wait for the sound intensity in the environment to reach a certain intensity threshold before it can be triggered. This causes staff to experience a brief period of no lighting or insufficient lighting when entering a poorly lit room (for example, when stepping from a bright corridor into the door of a dark room), which brings inconvenience to operation. Summary of the Invention

[0005] This invention provides a multi-terminal training method, electronic device, and storage medium for an acid poisoning identification model, in order to solve the problem of low model recognition accuracy caused by the centralized training mode in the prior art.

[0006] On one hand, the present invention provides an indoor light field adaptive tracking control method, wherein multiple indoor areas are respectively equipped with lighting units, each lighting unit having a base brightness level and a working brightness level, wherein the lighting brightness of the working brightness level is greater than that of the base brightness level, and the method includes: Obtain real-time trajectory data of the target user; Based on the target user's historical behavior data, determine the target user's habitual behavior pattern, which includes at least a habitual sequence and the typical duration of stay in various types of rooms; Based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the target indoor area where the target user will be located in the next time period is predicted, and the target indoor area is identified as a spotlight area, while each indoor area outside the target indoor area is identified as a non-spotlight area. Generate light field control commands to control the lighting units corresponding to the focused area to increase to the working brightness level, and control the lighting units corresponding to the non-focused area to decrease to the basic brightness level.

[0007] Preferably, based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the target indoor area where the target user will be located in the next time period is predicted, specifically including: Under preset conditions, based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the target indoor area where the target user will be located in the next time period is predicted; wherein, the preset conditions specifically include any one or more of the following: based on the real-time trajectory data, it is determined that the actual movement direction of the target user points to the exit of the current indoor area, and the actual movement speed is greater than a first preset threshold; based on the real-time trajectory data, the actual stay time of the target user in the current indoor area is calculated, and the actual stay time is greater than a second preset threshold, wherein the second preset threshold is determined based on the typical stay time.

[0008] Preferably, based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the target indoor area where the target user will be located in the next time period is predicted, specifically including: The current indoor area of ​​the target user is determined based on the real-time trajectory data; Taking the exit of the current indoor area as the starting point of the path, and combining the inertial order in the inertial behavior pattern, determine M candidate indoor areas, where M is a positive integer greater than or equal to 2; Based on the spatial topology distribution map, calculate the path cost from the starting point of the path to each of the M candidate indoor areas, wherein the path cost includes at least one of the following: the physical distance from the starting point of the path to the candidate indoor area, and the number of turns in the path; Based on the path cost of M candidate indoor areas, N candidate indoor areas with the lowest path cost are selected from the M candidate indoor areas as the predicted target indoor areas, where N is greater than or equal to 1 and less than M.

[0009] Preferably, the method further includes: combining the service occupancy status of each indoor area, removing target indoor areas with a service occupancy status of "occupied" from the predicted target indoor areas, wherein the service occupancy status includes "unoccupied" and "occupied".

[0010] Preferably, after generating the light field control command, the method further includes: Using the spatial topology distribution map, predictive trajectory data from the starting point of the path to each target indoor area is generated; Continuously acquire the target user's subsequent real-time trajectory data at the starting point of the path; Determine whether the subsequent real-time trajectory data deviates from all predicted trajectory data; If deviations occur, a new target indoor area is re-predicted based on the subsequent real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map, and a new light field control command is regenerated.

[0011] Preferably, the habitual behavior pattern of the target user is determined based on the target user's historical behavior data, specifically including: The historical behavior data is obtained, wherein the historical behavior data includes a time-stamped user location sequence and a task status record that is temporally associated with the user location sequence, and the task status record includes at least tasks being executed and tasks not being executed; From the historical behavior data, the user location sequence with timestamps during task execution is selected as valid job data from the task status records. Based on the effective operational data, calculate the statistical values ​​of the dwell time of the target user in each type of indoor area, and determine the statistical values ​​as the typical dwell time of the corresponding type of indoor area; Based on the effective job data, the target user's access sequences to various indoor areas and the frequency of occurrence of each indoor area access sequence are statistically analyzed when the target user completes a job task. The inertial order is determined by using one or more indoor area visit sequences that rank highest in frequency.

[0012] Preferably, the inertial order is determined using one or more indoor area visit sequences that rank highest in frequency, specifically including: Based on the most frequent one or more indoor area access sequences, calculate the conditional probability of entering other indoor areas after leaving each indoor area. The probability transition model, composed of all the aforementioned conditional probabilities, is determined as the inertial order.

[0013] Preferably, acquiring the real-time trajectory data of the target user specifically includes: when the target user completes authorized login verification, acquiring the real-time trajectory data of the target user through an indoor positioning system deployed in the multiple indoor areas, wherein the indoor positioning system includes at least one of the following: an ultra-wideband positioning system, a Bluetooth beacon positioning system, a Wi-Fi fingerprint positioning system, and a camera-based visual recognition positioning system.

[0014] Preferably, the real-time trajectory data of the target user is obtained by an indoor positioning system deployed in the multiple indoor areas. Specifically, this includes: when the target user is detected to enter a pre-defined work monitoring area, or when the target user is predicted to leave the current indoor area, the sampling frequency of the indoor positioning system is activated or increased to obtain the real-time trajectory data.

[0015] Secondly, the present invention provides an indoor light field adaptive tracking control system, wherein multiple indoor areas are respectively equipped with lighting units, each lighting unit having a base brightness level and a working brightness level, wherein the lighting brightness of the working brightness level is greater than that of the base brightness level, and the system includes: The acquisition unit is used to acquire real-time trajectory data of the target user. The determining unit is used to determine the target user's habitual behavior pattern based on the target user's historical behavior data. The habitual behavior pattern includes at least a habitual sequence and the typical duration of stay in various types of rooms. The prediction unit is used to predict the target indoor area where the target user will be located in the next time period based on the real-time trajectory data, the inertial behavior pattern and the spatial topology distribution map of each indoor area, and to determine the target indoor area as a spotlight area and to determine each indoor area outside the target indoor area as a non-spotlight area. The instruction generation unit is used to generate light field control instructions to control the lighting units corresponding to the focusing area to increase to the working brightness level, and to control the lighting units corresponding to the non-focusing area to decrease to the basic brightness level.

[0016] The indoor light field adaptive tracking control method provided in this application embodiment has multiple indoor areas equipped with lighting units. Each lighting unit has a base brightness level and a working brightness level, where the working brightness level is greater than the base brightness level. The method includes acquiring real-time trajectory data of a target user, then determining the target user's inertial behavior pattern based on the target user's historical behavior data. The inertial behavior pattern includes at least an inertial sequence and typical dwell time in various types of rooms. Then, based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the method predicts the target indoor area where the target user will be located in the next time period, and identifies the target indoor area as a focused area, and identifies all indoor areas outside the target indoor area as non-focused areas. Finally, a light field control command is generated to control the lighting unit corresponding to the focused area to increase to the working brightness level, and to control the lighting unit corresponding to the non-focused area to decrease to the base brightness level. This method combines the target user's real-time trajectory data, inertial behavior patterns, and spatial topology information to proactively predict the target indoor area where the target user will be in the next time period. As a result, the corresponding lighting units are turned up to working brightness in advance before the target user actually enters. This method fundamentally changes the passive response mode of traditional sound and light sensors that "turn on the light after detecting a person", solves the lighting delay problem between entering a dark area and the light fixture turning on, provides users with a seamless and continuous lighting-first experience, and significantly improves the ease of operation. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of an interior scene with multiple rooms in the prior art; Figure 2 A flowchart illustrating an indoor light field adaptive tracking control method provided by the present invention; Figure 3 A schematic diagram of the process for predicting the target indoor area in the indoor light field adaptive tracking control method provided by the present invention; Figure 4 A structural block diagram of an indoor light field adaptive tracking control system provided by the present invention; Figure 5 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0020] As mentioned earlier, the indoor smart lighting fixtures currently used in hotels are mainly controlled by sound and light sensors. However, this sound and light sensor control method is a passive response trigger, and the lighting of the fixtures has a significant lag. For example, it needs to wait for the sound intensity in the environment to reach a certain intensity threshold before it can be triggered. This causes staff to experience a brief period of no lighting or insufficient lighting when they first enter a room with insufficient light, which brings inconvenience to operation.

[0021] In view of this, embodiments of this application provide an indoor light field adaptive tracking control method and system, which can be used to solve the problems in the prior art. For ease of understanding, the embodiments of this application can be described in general here, such as... Figure 1 The diagram shown illustrates an indoor scenario according to this application. This scenario includes multiple indoor areas, such as different rooms in a hotel. Each room is equipped with a lighting unit, which can be a smart indoor light fixture. It is important to note that each lighting unit has a base brightness level and a working brightness level, with the working brightness level being greater than the base brightness level. The working brightness level is sufficient to provide adequate illumination for the daily work of management personnel, while the base brightness level is relatively lower, thus achieving energy savings (generally, hotel rooms need to maintain a certain level of lighting to provide better comfort for visitors and guests).

[0022] In practical applications, there are usually several ways to implement this method of setting two different brightness levels for a lighting unit: a base brightness level and a working brightness level. One approach is to directly set two different brightness levels for the lighting unit (indoor smart luminaire), where the higher brightness level can be used as the working brightness level, and the lower brightness level as the base brightness level. Another approach is to set multiple individual luminaires (each of which can be an indoor smart luminaire) within the lighting unit. This allows control over the number of luminaires lit, thus setting two different brightness levels. For example, the working brightness level could illuminate most or all luminaires, while the base brightness level could illuminate only a small portion or even just one luminaire.

[0023] like Figure 2 The diagram shown is a schematic flowchart of the indoor light field adaptive tracking control method provided in this application embodiment. The method includes the following steps: Step S21: Obtain the real-time trajectory data of the target user.

[0024] The real-time trajectory data reflects the current location and movement trend of the target user, serving as the direct input for subsequent area prediction. In practical applications, this real-time trajectory data can refer to a sequence of spatial coordinates of the target user within multiple indoor areas, recorded sequentially by timestamps. These spatial coordinates can be represented as two-dimensional planar coordinates (x, y) or three-dimensional spatial coordinates (x, y, z), with an accuracy sufficient to distinguish the boundaries of different indoor areas.

[0025] In this application, when the target user completes authorized login verification, the real-time trajectory data of the target user can be obtained by an indoor positioning system deployed in multiple indoor areas. For example, the indoor positioning system deployed in each indoor area can collect the location information of the terminal device (such as employee badge, smart bracelet) carried by the target user in real time, and generate a time-stamped location point sequence according to a fixed sampling period (0.5 seconds). The indoor positioning system includes at least one of the following: ultra-wideband positioning system, Bluetooth beacon positioning system, Wi-Fi fingerprint positioning system, and camera-based visual recognition positioning system.

[0026] Of course, this application can also, for example, fuse the outputs of multiple positioning subsystems (such as ultra-wideband base station ranging data and camera visual recognition results) and perform Kalman filtering to output smooth, continuous, and low-jitter trajectory data, remove abnormal jump points and interpolate short-term lost data, and output a trajectory stream with spatiotemporal consistency. This application obtains real-time trajectory data for characterizing the user's dynamic spatial behavior based on any of the above methods.

[0027] For example, this application could involve a hotel cleaning staff member wearing a work badge that supports UWB positioning. Once the target user completes authorized login verification, the indoor positioning system can acquire the (x, y) coordinates of the user on the floor plan every 0.5 seconds, forming a sequence like t0:(2.3, 5.1); t1:(2.4, 5.2); t2:(2.6, 5.3)... This sequence, as real-time trajectory data, can completely reflect the real-time movement process of the target user.

[0028] In addition, as a condition for triggering step S21 to obtain the real-time trajectory data of the target user, the sampling frequency of the indoor positioning system can be activated or increased when the target user is detected to have entered a pre-defined work monitoring area or when it is predicted that the target user is about to leave the current indoor area, thereby obtaining the real-time trajectory data. If the target user has not entered the pre-defined work monitoring area (that is, the target user is outside the work monitoring area), the indoor positioning system can be turned off to collect the real-time trajectory data of the target user. Alternatively, when the target user is working normally in the current indoor area, the real-time trajectory data can be obtained at a relatively low sampling frequency.

[0029] Step S22: Based on the target user's historical behavior data, determine the target user's habitual behavior pattern, which includes at least the habitual sequence and the typical duration of stay in various types of rooms.

[0030] In practical applications, step S22 can be implemented in the following way: Specifically, the historical behavior data of the target user can be obtained first. The historical behavior data includes a time-stamped user location sequence and a task status record that is associated with the user location sequence in time. For example, the historical behavior data can be a time-stamped user location sequence generated by the target user in the past period (30 days) when performing similar tasks, and a task status record that is associated with the location sequence in time. Of course, the task status record can be a job status marker aligned with the above location record in the time dimension. The task status record can be used to characterize the business stage that the target user is in at the corresponding time. For example, it can be executing a task or not executing a task. Executing a task means that the target user is currently executing a task, and not executing a task means that the target user is not executing a task at this time.

[0031] The timestamped user location sequence can be a set of location records arranged chronologically, each containing an indoor area identifier and a corresponding timestamp. For example, the timestamped user location sequence could be (Room A, 2024-01-01T08:02:15), (Room B, 2024-01-01T08:05:33), (Room C, 2024-01-01T08:09:47), etc. Therefore, in this embodiment, there is a temporal correlation between the task status record and the timestamped user location sequence. This correlation enables the system to identify which location movement behaviors belong to the actual work process, thereby providing a basis for subsequent data cleaning.

[0032] After obtaining the historical behavior data, the user location sequence with timestamps whose task status records are in the process of executing a task can be further filtered out as valid work data. This valid work data refers to a subset of movement trajectory data that reflects only the target user's actual work status after task status filtering. In this embodiment, the filtering action uses task status records as the criterion; only when the task status corresponding to a location record is "in execution" is it included in the valid work data. Therefore, this method eliminates location disturbances caused by users being in non-work states such as standby, rest, or off-duty, allowing subsequent modeling to focus on the actual work movement.

[0033] After obtaining the valid operational data, the statistical values ​​of the duration of stay of the target user in each type of indoor area can be calculated based on the valid operational data, and the statistical values ​​can be determined as the typical duration of stay in the corresponding type of indoor area.

[0034] The various types of indoor areas can refer to room categories classified according to functional attributes, such as guest rooms, restrooms, storage rooms, corridors, and office areas. The dwell time can refer to the time span during which a target user continuously resides in the same indoor area, with the start time being the timestamp of the first valid location record upon entering the area and the end time being the timestamp of the first valid location record upon leaving the area. This statistical value can be a measure of central tendency obtained by summarizing multiple dwell times for the same type of indoor area, such as the mean, median, or mode. In this embodiment, this statistical value is directly used as the typical dwell time for this type of indoor area to support the setting of the subsequent second preset threshold and as the basis for estimating the theoretical travel time in path cost.

[0035] Furthermore, based on this valid task data, we can statistically analyze the various indoor area access sequences and their frequencies when the target user completes a task. The various indoor area access sequences can refer to the ordered combinations of indoor areas experienced by the target user within a complete task cycle, such as (guest room → bathroom → storage room), (office area → corridor → guest room), (guest room A → corridor → guest room B), etc. The frequency of occurrence can refer to the proportion of a certain type of access sequence appearing in all valid task data out of the total number of tasks.

[0036] In this embodiment, the statistical action is performed on a unit of a single task, that is, each segment is identified as a continuous position sequence that starts and ends in the task and has no interruption in between, and the access order of the indoor area covered by it is extracted accordingly. This processing method ensures that the statistical access sequence truly reflects the work process logic, rather than random moving segments.

[0037] After obtaining the frequency of occurrence of various indoor area access sequences, the inertial order can be determined by utilizing one or more indoor area access sequences with the highest frequency ranking. This inertial order can refer to a stable preference ranking of indoor area access formed by the target user during long-term operation. In one aspect of this application, the inertial order can be directly derived from the most frequent indoor area access sequence. For example, when (guest room → bathroom → storage room) is the highest frequency sequence, the inertial order reflects that after departing from the guest room, the most likely destination is the bathroom, and then the storage room. Therefore, this inertial order will serve as the core basis for subsequently determining candidate indoor areas, enabling the narrowing of the prediction search space and improving prediction efficiency and rationality.

[0038] For example, this application could involve hotel cleaning staff completing cleaning tasks for a number of guest rooms daily. The system pre-collects their historical behavioral data for the past 30 days, including timestamped location records (e.g., entering room 301 at 08:02:15, leaving room 301 at 08:07:42) and synchronously recorded task statuses (e.g., performing tasks from 08:00 to 09:30). The system first filters out all location records with the task status "performing tasks," removing data from non-work periods such as dining in the restaurant or staying in the staff break room. Then, for the filtered data, it calculates the average time spent in various guest rooms (12 minutes), in the bathroom (3 minutes), and in the storage room (5 minutes). Next, it analyzes the access sequence when completing a single task, finding that (guest room → bathroom → storage room) appears most frequently (68%), followed by (guest room → storage room → bathroom) (22%). Finally, (guest room → bathroom → storage room) is determined as the inertial sequence, and the corresponding dwell time is used as a typical value for subsequent predictions.

[0039] Therefore, this application filters the historical behavior data by recording the task status to obtain effective operation data that only reflects the actual operation process; it uses this data to calculate the typical stay time of each type of indoor area and counts the high-frequency indoor area access sequence; then it maps the high-frequency sequence to an inertial sequence, so that the constructed inertial behavior pattern has clear business semantics and empirical support, thereby improving the accuracy and robustness of the prediction of the target indoor area in the next period.

[0040] After obtaining the occurrence frequencies of various indoor area access sequences, the above-mentioned method uses one or more indoor area access sequences with the highest occurrence frequency to determine the inertial order. That is, the inertial order is directly derived from the indoor area access sequence with the highest occurrence frequency. In practical applications, other methods can also be used to determine the inertial order. For example, based on one or more indoor area access sequences with the highest occurrence frequency, the conditional probability of entering other indoor areas after leaving each indoor area can be calculated, and then the probability transition model composed of all conditional probabilities can be determined as the inertial order.

[0041] In this embodiment, the conditional probability can obviously be used to quantify the behavioral preference intensity of a target user moving from one indoor area to other indoor areas during the completion of a task, thereby supporting the subsequent differentiated modeling of the possibility of multi-path transfers. Specifically, based on the 3-5 indoor area access sequences with the highest frequency, the conditional probability of entering other indoor areas after leaving each indoor area can be calculated. For example, assuming the 3 most frequent indoor area access sequences are A→B→C, A→B→D, and A→E→F, all transfer pairs between adjacent areas can be extracted, including A→B, B→C, B→D, A→E, and E→F. The number of times each transfer pair appears in all valid task data can be counted and divided by the total number of departures from the corresponding departure areas (A, B, E) to obtain P(B|A), P(C|B), P(D|B), P(E|A), and P(F|E). This constitutes a conditional probability set covering all indoor area nodes, which includes the conditional probability of entering other indoor areas after leaving each indoor area.

[0042] The probabilistic transition model is a two-dimensional structure organized in matrix form. Its row index represents the departure indoor area, and its column index represents the arrival indoor area. The matrix element values ​​are the corresponding conditional probabilities. The model is a sparse matrix, and only the regions where transitions have actually occurred are assigned non-zero values. In practical applications, the probabilistic transition model can be a first-order state transition matrix of a Markov chain, where the sum of the elements in each row is equal to 1. It is used to characterize the normalized transition tendency distribution of the target user under any given departure area.

[0043] In this embodiment, after obtaining the conditional probabilities of entering other indoor areas after leaving each indoor area, a Markov chain-related scheme from the prior art can be used to generate a probability transition model from these conditional probabilities. This probability transition model, as a mathematical expression of inertial order, replaces the simple sequence-based empirical rules, allowing the prediction stage to directly call the corresponding row vector based on the current area to obtain all possible next areas and their probability weights, thus supporting a multi-select or multi-target area parallel prediction mechanism. For example, when the target user is currently in indoor area B, looking up the data in row B of the probability transition model, we find P(C|B)=0.35, P(D|B)=0.62, and P(A|B)=0.03. Therefore, we can determine that the target user is most likely to go to D, followed by C, while the probability of returning to A is low.

[0044] Step S23: Based on real-time trajectory data, inertial behavior patterns, and spatial topology distribution maps of various indoor areas, predict the target indoor area where the target user will be located in the next time period, and determine the target indoor area as a spotlight area, and determine all indoor areas outside the target indoor area as non-spotlight areas.

[0045] The spatial topology distribution map can refer to map data expressed in graph structure, showing the connectivity and geometric constraints between multiple indoor areas. Its nodes represent each indoor area, and edges represent the passable paths between two areas. The edges can be labeled with attributes such as physical distance, direction of travel, and access control status. The next time period can refer to a time interval with the current time as the starting point and a preset time window length (e.g., 1 minute, 5 minutes, 10 minutes, etc.). The spotlight area can refer to the indoor area where the system predicts that the target user will actually enter and carry out work in the next time period, and the lighting brightness needs to be increased in advance to ensure working conditions. The non-spotlight area can refer to all other indoor areas except the spotlight area, and its lighting only needs to maintain the basic brightness level to save energy.

[0046] It should be further explained that, for step S23, before execution, it can be determined whether the preset conditions are met. The preset conditions can refer to a set of judgment criteria used to trigger the location prediction process. Its function is to limit the prediction logic to only when the target user shows the possibility of leaving the area (i.e. leaving the current indoor area), thereby avoiding the waste of system resources and miscontrol of the light field due to frequent and meaningless predictions. In practical applications, the preset conditions specifically include any one or more of the following two independent judgment conditions that can be used to quantitatively identify the user's behavior state. The first condition is to determine, based on the real-time trajectory data, that the target user's actual movement direction points to the exit of the current indoor area and the actual movement speed is greater than the first preset threshold. The second condition is to calculate, based on the real-time trajectory data, the actual stay time of the target user in the current indoor area and the actual stay time is greater than the second preset threshold, which is determined based on the typical stay time. Conditions one and two, from two orthogonal dimensions of spatial movement trend and time dwell pattern, jointly characterize the target user's intention to leave, improving the robustness of the prediction timing judgment. When either condition one or condition two is satisfied, it means that the preset conditions are met, thus constituting the conditions for executing step S23. Thus, step S23 can be: under the condition of satisfying the preset conditions, based on real-time trajectory data, inertial behavior patterns, and spatial topology distribution maps of various indoor areas, predict the target indoor area where the target user will be located in the next time period.

[0047] The first condition can be further explained. The actual movement direction in the first condition can refer to the direction of movement trend represented by the displacement vector formed by two or more consecutive time-stamped position coordinate points. Its function is to reflect the user's instantaneous movement orientation in space and provide a geometric basis for determining whether it is heading towards the exit. The exit of the current indoor area can be a physical access interface connecting the current indoor area with other areas, which is marked in advance on the spatial topology distribution map. Its function is to serve as a benchmark target for direction comparison and ensure that the movement trend judgment has a clear spatial reference. The actual movement speed can be the Euclidean distance of the user's position change per unit time. Its function is to quantify the urgency of the user's departure behavior and eliminate the misjudgment interference caused by slow movement during the execution of tasks (such as cleaning). The first preset threshold can be an empirical threshold value set according to the normal walking speed range of indoor workers. Its function is to establish the operable boundary of the speed criterion and ensure that only rapid movement behavior with a clear intention to leave is responded to.

[0048] Therefore, based on the real-time trajectory data of the target user, the position coordinates P1(x1, y1) and P2(x2, y2) corresponding to adjacent timestamps t1 and t2 can be calculated to obtain the displacement vector (x2-x1, y2-y1). Then, the displacement vector (x2-x1, y2-y1) and the exit direction unit vector (calculated by the target user's current actual position coordinates and the exit position coordinates) are multiplied by a dot product. If the dot product value is greater than cos(30°) and the target user's actual moving speed (calculated by the magnitude of the displacement vector / (t2-t1)) is greater than a first preset threshold (e.g., 1 meter per second), then condition one is satisfied. Alternatively, this application can also use a sliding window method to linearly fit the coordinates of multiple nearest position points, and use the fitted straight line direction instead of the displacement vector and the exit direction unit vector to perform the dot product operation to improve the accuracy of direction judgment.

[0049] The second condition can be further explained. The actual dwell time in condition two can refer to the elapsed time from the moment the target user last entered the current indoor area to the present moment. Its function is to characterize the user's progress in completing tasks within the area, providing a time dimension for predictive initiation. The second preset threshold can be a dynamic threshold derived from the typical dwell time defined in this application by setting an offset (e.g., +1 minute), percentile (e.g., the 90th percentile), or upper limit of the confidence interval. It can also be 70% to 80% of the typical dwell time. The purpose of this second preset threshold is to adapt the judgment standard to individual behavioral differences, balancing universality and personalization. The typical dwell time, as a previously defined technical feature, is only referenced in this step as the basis for generating the second preset threshold. Thus, in condition two, the actual dwell time and the second preset threshold can be directly compared. If the actual dwell time is greater than the second preset threshold, condition two is satisfied; otherwise, it is not.

[0050] It should be further explained that, in the embodiments of this application, the specific implementation of step S23 can be further described in conjunction with the target indoor area prediction method shown in the figure. The method includes the following steps: Step S231: Determine the current indoor area of ​​the target user based on the real-time trajectory data.

[0051] The current indoor area refers to the indoor area where the target user is located at the current moment. This current indoor area can be directly obtained through the real-time trajectory data, and will not be further explained here.

[0052] Step S232: Taking the exit of the current indoor area as the starting point of the path, and combining the inertial order in the inertial behavior pattern, determine M candidate indoor areas, where M is a positive integer greater than or equal to 2.

[0053] The exit of the current indoor area can refer to the location of the boundary passage between other indoor areas that are adjacent to the current indoor area and have accessibility in the spatial topology distribution map. It can be directly obtained from the spatial topology distribution map. The inertial sequence and inertial behavior pattern have been explained in step S22 above, and will not be repeated here.

[0054] The M candidate indoor areas are extracted from various indoor areas and are several indoor areas that may be accessed after the current user's current indoor area. The selection is based on the frequency of the area appearing next to the current area in the inertial order, without considering physical accessibility. The purpose of this step is to transform the user's long-term established work path preferences into a structured candidate set, reduce the prediction search space, improve computational efficiency, and retain the statistical representativeness of behavioral patterns.

[0055] For example, this application can retrieve all successor regions with the current indoor area as the predecessor node from the inertial sequence, sort them in descending order according to their joint occurrence frequency in historical valid operation data, and select the top M as candidates. For example, as mentioned in the aforementioned step S22, the inertial sequence can be a probability transition model, and the probability transition model can be composed of all conditional probabilities. Therefore, by querying the probability transition model, the conditional probability of the current user leaving the current indoor area and arriving at other indoor areas can be obtained. After further arranging the conditional probabilities in descending or ascending order, the M indoor areas with the highest conditional probabilities are selected as candidate indoor areas.

[0056] Step S233: Based on the spatial topology distribution map, calculate the path cost from the path starting point to each of the M candidate indoor areas.

[0057] The starting point of the path is the node position of the current indoor area's exit in the spatial topology distribution map. The path cost includes at least one of the following: the physical distance from the starting point of the path to the candidate indoor area and the number of turns in the path. The physical distance can refer to the shortest Euclidean distance from the starting point of the path along the connected edges in the topology map to the entrance node of the target candidate indoor area. The number of turns in the path can refer to the number of nodes whose direction changes by more than a preset angle threshold (e.g., 45°) in the path from the starting point of the path to the entrance of the target candidate indoor area. Therefore, the role of step S233 is to introduce hard constraints of the physical environment, perform feasibility filtering on the purely behavior-driven candidate set, so that the prediction results reflect user habits and meet spatial accessibility requirements, avoiding recommending invalid areas that require excessive detours.

[0058] In this embodiment, the spatial topology map can be a digital map that represents the connectivity and spatial adjacency relationships between multiple indoor areas in a graph structure. Nodes represent indoor areas, and edges represent travel paths between areas. Edge attributes can include physical distance, number of turns, etc. Alternatively, the spatial topology map can be modeled as a weighted undirected graph, where nodes represent indoor area exits / entrances, edges represent connecting channels, and edge weights comprehensively represent physical distance, turning penalties, and travel resistance. In this case, Dijkstra's algorithm can be used to calculate the shortest weighted path length from the path origin to the entrance node of each candidate indoor area, and this length can be used as the path cost. Furthermore, this application can also involve pre-constructing a path cost lookup table that stores the standardized path costs between all pairs of areas. During runtime, the table is directly consulted to obtain the path costs from the path origin to each of the M candidate indoor areas. This application obtains the quantified path costs of each candidate indoor area based on any of the above methods, providing a numerical basis for subsequent screening.

[0059] Step S234: Based on the path cost of M candidate indoor areas, select the N candidate indoor areas with the lowest path cost from the M candidate indoor areas as the predicted target indoor areas, where N is greater than or equal to 1 and less than M.

[0060] Among them, the lowest path cost can refer to the smallest path cost value. When there are ties, the original frequency in the inertial order can be sorted from high to low as a secondary criterion. N candidate indoor areas constitute the final prediction output set, and the number N can be dynamically configured according to the system resource scheduling strategy. For example, it can be set to 1 in a computing power-limited scenario and 2 to 3 in a multi-area collaborative lighting scenario. The role of this step is to achieve a weighted balance between behavioral preferences and physical feasibility, and output a finite number of high-confidence prediction results to support the accurate delineation of the spotlight area and the generation of instructions.

[0061] For example, in practical applications, the path costs of M candidate indoor areas can be arranged in descending or ascending order, and then the N candidate indoor areas with the lowest path costs can be selected as the predicted target indoor areas. Alternatively, a path cost threshold can be set, retaining only candidate areas with costs below the threshold, and then the top N from these can be selected in ascending order of cost. This application can obtain N candidate indoor areas with optimal path costs based on either of the above methods, which can then be used as the final set of predicted target indoor areas.

[0062] It should be further explained that after selecting the N candidate indoor areas with the lowest path costs as the predicted target indoor areas, the business occupancy status of each indoor area can be considered to further eliminate target indoor areas that are currently occupied. This business occupancy status can refer to a logical identifier reflecting whether an indoor area is currently in use; typically, this status includes both unoccupied and occupied. This business occupancy status can be synchronously obtained from a pre-set business management system (the hotel's business management system).

[0063] In this embodiment, the service occupancy status is used to perform secondary screening on the predicted N candidate indoor areas. Its function is to exclude areas that are currently unavailable or unsuitable for turning on the working brightness lighting, thereby ensuring that the light field control command only applies to indoor areas that meet the user's movement expectations and are actually available.

[0064] Step S24: Generate light field control commands to control the lighting units corresponding to the spotlight area to increase to the working brightness level, and control the lighting units corresponding to the non-spotlight area to decrease to the basic brightness level.

[0065] The light field control command can refer to a structured control message containing a target area identifier, target brightness level, execution timestamp, and priority fields. In this application, after generating the light field control command, a PWM duty cycle increase command can be sent to the corresponding lighting unit in the focusing area according to the light field control command, so that its output luminous flux increases to the level corresponding to the working brightness. At the same time, a dimming command is broadcast to the lighting units in the non-focusing area through the DALI bus, uniformly setting their brightness to the base level. Alternatively, this application can also use a Zigbee network to distribute the light field control command to each area gateway, and the gateway drives the local LED driver circuit to complete the brightness switching.

[0066] For example, this application could send a brightness = 300 lx command to the lighting unit of guest room A (as a spotlighting area) to control the lighting unit of guest room A to adjust to the working brightness level, and at the same time send a brightness = 50 lx command to the lighting units of non-spotlighting areas such as guest room B and guest room C, thereby controlling the lighting units corresponding to these non-spotlighting areas to reduce to the basic brightness level.

[0067] It should be further explained that after generating the light field control command to control the lighting unit corresponding to the spotlight area to increase to the working brightness level and control the lighting unit corresponding to the non-spotlight area to decrease to the basic brightness level, in order to further improve the accuracy of control, the embodiments of this application may also include a correction step. Therefore, after generating the light field control command, the method may further include using the spatial topology distribution map to generate predicted trajectory data from the starting point of the path to each target indoor area.

[0068] This application can, for example, use a shortest path algorithm between nodes in a spatial topology map to generate a path sequence containing the intermediate regions and the coordinates of key turning points, starting from the path origin and sequentially traversing to the terminal nodes corresponding to each target indoor area. This serves as the predicted trajectory data from the path origin to each target indoor area. Alternatively, this application can use multiple pre-stored typical travel path templates in the spatial topology map to match the spatial relative relationship between the path origin and each target indoor area, retrieve the corresponding template, and perform endpoint adaptation to generate predicted trajectory data. This application obtains predicted trajectory data with spatial continuity and logical reachability for subsequent comparison based on any of the above methods.

[0069] After generating predicted trajectory data from the starting point of the path to each target indoor area, subsequent real-time trajectory data of the target user at the starting point of the path can be continuously acquired. This subsequent real-time trajectory data is a time-stamped sequence of two-dimensional or three-dimensional spatial coordinates continuously collected after the target user passes through the starting point of the path, covering the entire process from the starting point of the path until entering any indoor area. Here, the starting point of the path serves as a dual reference, serving as both a time zero point and a spatial origin, to align the predicted trajectory with the time and spatial axes of the actual movement. Of course, the method of acquiring this subsequent real-time trajectory data can be the same as the method of acquiring real-time trajectory data described above. For example, if the target user completes authorized login verification, the subsequent real-time trajectory data can also be acquired through the deployed indoor positioning system. The method of acquiring this subsequent real-time trajectory data will not be elaborated here.

[0070] After obtaining the predicted trajectory data from the starting point of the path to each target indoor area, as well as the subsequent real-time trajectory data, it is possible to further determine whether the subsequent real-time trajectory data deviates from each of the predicted trajectory data. Here, deviation can mean that the Euclidean distance between the spatial coordinates of multiple consecutive sampling points in the subsequent real-time trajectory data and the expected coordinates of the corresponding time or path segment in any predicted trajectory data exceeds a third preset threshold, and the duration of the deviation is greater than a fourth preset threshold. Deviation can mean that after performing deviation judgment on the predicted trajectory data corresponding to each of the N target indoor areas one by one, the result is always yes, that is, no predicted trajectory can effectively fit the actual motion path.

[0071] Therefore, this application may, for example, calculate the vertical distance from each sampling point in the subsequent real-time trajectory data to the predicted trajectory data for each predicted trajectory data. If the vertical distance from K consecutive sampling points to the predicted trajectory data exceeds a preset distance, it indicates that the subsequent real-time trajectory data deviates from the predicted trajectory data. In this way, it can be determined whether the subsequent real-time trajectory data deviates from each predicted trajectory data. Of course, this application may also, for example, project the subsequent real-time trajectory data into each predicted trajectory path in segments, and then determine whether its cumulative path offset exceeds a preset tolerance bandwidth, and then determine whether the subsequent real-time trajectory data deviates from each predicted trajectory data.

[0072] If the subsequent real-time trajectory data does not deviate from at least one of the predicted trajectory data, no processing is required. Otherwise, if the subsequent real-time trajectory data deviates from all the predicted trajectory data, it indicates that correction is needed. In this case, a new target indoor area can be predicted based on the subsequent real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map, and a new light field control command can be generated.

[0073] Here, re-prediction can refer to replacing the initial judgment in step S23, which determines the target user's current indoor area based on real-time trajectory data, with the latest position and movement trend represented by subsequent real-time trajectory data as input, and re-execute all prediction sub-steps defined in step S23. This will not be elaborated further here.

[0074] The indoor light field adaptive tracking control method provided in this application embodiment has multiple indoor areas equipped with lighting units. Each lighting unit has a base brightness level and a working brightness level, where the working brightness level is greater than the base brightness level. The method includes acquiring real-time trajectory data of a target user, then determining the target user's inertial behavior pattern based on the target user's historical behavior data. The inertial behavior pattern includes at least an inertial sequence and typical dwell time in various types of rooms. Then, based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the method predicts the target indoor area where the target user will be located in the next time period, and identifies the target indoor area as a focused area, and identifies all indoor areas outside the target indoor area as non-focused areas. Finally, a light field control command is generated to control the lighting unit corresponding to the focused area to increase to the working brightness level, and to control the lighting unit corresponding to the non-focused area to decrease to the base brightness level. This method combines the target user's real-time trajectory data, inertial behavior patterns, and spatial topology information to proactively predict the target indoor area where the target user will be in the next time period. As a result, the corresponding lighting units are turned up to working brightness in advance before the target user actually enters. This method fundamentally changes the passive response mode of traditional sound and light sensors that "turn on the light after detecting a person", solves the lighting delay problem between entering a dark area and the light fixture turning on, provides users with a seamless and continuous lighting-first experience, and significantly improves the ease of operation.

[0075] Based on the same inventive concept as the indoor light field adaptive tracking control method provided in the embodiments of this application, the embodiments of this application can also provide an indoor light field adaptive tracking control device. For any unclear points regarding the content of this device embodiment, please refer to the relevant content in the above method embodiments. In this device embodiment, multiple indoor areas are respectively provided with lighting units. Each lighting unit has a base brightness level and a working brightness level, with the working brightness level having a higher brightness than the base brightness level. Figure 4 The diagram shown is a schematic representation of the indoor light field adaptive tracking control device 30 (hereinafter referred to as device 30). Device 30 includes: an acquisition unit 301, a determination unit 302, a prediction unit 303, and a command generation unit 304, wherein: Acquisition unit 301 is used to acquire real-time trajectory data of the target user; The determining unit 302 is used to determine the target user's habitual behavior pattern based on the target user's historical behavior data. The habitual behavior pattern includes at least a habitual sequence and the typical duration of stay in various types of rooms. The prediction unit 303 is used to predict the target indoor area where the target user will be located in the next time period based on the real-time trajectory data, the inertial behavior pattern and the spatial topology distribution map of each indoor area, and to determine the target indoor area as a spotlight area and to determine each indoor area outside the target indoor area as a non-spotlight area. The instruction generation unit 304 is used to generate light field control instructions to control the lighting unit corresponding to the focusing area to increase to the working brightness level and control the lighting unit corresponding to the non-focusing area to decrease to the basic brightness level.

[0076] The device 30 provided in the embodiments of this application adopts the same inventive concept as the method provided in the embodiments of this application. Since the method can solve the problems in the prior art, the device 30 can also solve the problems in the prior art. This will not be elaborated here.

[0077] Specifically, predicting the target indoor area where the target user will be located in the next time period based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area can include: Under preset conditions, based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the target indoor area where the target user will be located in the next time period is predicted; wherein, the preset conditions specifically include any one or more of the following: based on the real-time trajectory data, it is determined that the actual movement direction of the target user points to the exit of the current indoor area, and the actual movement speed is greater than a first preset threshold; based on the real-time trajectory data, the actual stay time of the target user in the current indoor area is calculated, and the actual stay time is greater than a second preset threshold, wherein the second preset threshold is determined based on the typical stay time.

[0078] Specifically, predicting the target indoor area where the target user will be located in the next time period based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area can include: The current indoor area of ​​the target user is determined based on the real-time trajectory data; Taking the exit of the current indoor area as the starting point of the path, and combining the inertial order in the inertial behavior pattern, determine M candidate indoor areas, where M is a positive integer greater than or equal to 2; Based on the spatial topology distribution map, calculate the path cost from the starting point of the path to each of the M candidate indoor areas, wherein the path cost includes at least one of the following: the physical distance from the starting point of the path to the candidate indoor area, and the number of turns in the path; Based on the path cost of M candidate indoor areas, N candidate indoor areas with the lowest path cost are selected from the M candidate indoor areas as the predicted target indoor areas, where N is greater than or equal to 1 and less than M.

[0079] The device 30 may further include a rejection unit for combining the service occupancy status of each indoor area and rejecting target indoor areas with a service occupancy status of "occupied" from the predicted target indoor areas, wherein the service occupancy status includes "unoccupied" and "occupied".

[0080] After generating the light field control command, the device 30 may further include a deviation adjustment unit, which is used to generate predicted trajectory data from the starting point of the path to each target indoor area using the spatial topology distribution map; continuously acquire subsequent real-time trajectory data of the target user at the starting point of the path; determine whether the subsequent real-time trajectory data deviates from each predicted trajectory data; if a deviation occurs, then based on the subsequent real-time trajectory data, the inertial behavior pattern and the spatial topology distribution map, re-predict a new target indoor area and regenerate a new light field control command.

[0081] Specifically, determining the target user's habitual behavior pattern based on the target user's historical behavior data may include: The historical behavior data is obtained, wherein the historical behavior data includes a time-stamped user location sequence and a task status record that is temporally associated with the user location sequence, and the task status record includes at least tasks being executed and tasks not being executed; From the historical behavior data, the user location sequence with timestamps during task execution is selected as valid job data from the task status records. Based on the effective operational data, calculate the statistical values ​​of the dwell time of the target user in each type of indoor area, and determine the statistical values ​​as the typical dwell time of the corresponding type of indoor area; Based on the effective job data, the target user's access sequences to various indoor areas and the frequency of occurrence of each indoor area access sequence are statistically analyzed when the target user completes a job task. The inertial order is determined by using one or more indoor area visit sequences that rank highest in frequency.

[0082] Specifically, determining the inertial order using one or more indoor area visit sequences that rank highest in frequency can include: Based on the most frequent one or more indoor area access sequences, calculate the conditional probability of entering other indoor areas after leaving each indoor area. The probability transition model, composed of all the aforementioned conditional probabilities, is determined as the inertial order.

[0083] Specifically, obtaining the real-time trajectory data of the target user may include: after the target user has completed authorized login verification, obtaining the real-time trajectory data of the target user through an indoor positioning system deployed in the multiple indoor areas, wherein the indoor positioning system includes at least one of the following: an ultra-wideband positioning system, a Bluetooth beacon positioning system, a Wi-Fi fingerprint positioning system, and a camera-based visual recognition positioning system.

[0084] Specifically, acquiring the real-time trajectory data of the target user through an indoor positioning system deployed in the multiple indoor areas may include: when the target user is detected to enter a pre-defined work monitoring area, or when the target user is predicted to leave the current indoor area, activating or increasing the sampling frequency of the indoor positioning system to acquire the real-time trajectory data.

[0085] Figure 5 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other through the communication bus 440. The processor 410 can call logic instructions in the memory 430 to execute the indoor light field adaptive tracking control method provided in this application embodiment. Multiple indoor areas are each equipped with a lighting unit, which has a base brightness level and a working brightness level. The working brightness level is greater than the base brightness level. The method includes acquiring real-time trajectory data of a target user, then determining the target user's inertial behavior pattern based on the target user's historical behavior data. The inertial behavior pattern includes at least inertial sequence and typical dwell time in various types of rooms. Then, based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the method predicts the target indoor area where the target user will be located in the next time period, and identifies the target indoor area as a focused area, and identifies all indoor areas outside the target indoor area as non-focused areas. Finally, a light field control instruction is generated to control the lighting unit corresponding to the focused area to increase to the working brightness level, and to control the lighting unit corresponding to the non-focused area to decrease to the base brightness level. This method combines the target user's real-time trajectory data, inertial behavior patterns, and spatial topology information to proactively predict the target indoor area where the target user will be in the next time period. As a result, the corresponding lighting units are turned up to working brightness in advance before the target user actually enters. This method fundamentally changes the passive response mode of traditional sound and light sensors that "turn on the light after detecting a person", solves the lighting delay problem between entering a dark area and the light fixture turning on, provides users with a seamless and continuous lighting-first experience, and significantly improves the ease of operation.

[0086] Obviously, since the processor 410 can call the logical instructions in the memory 430 to execute the method provided in the embodiments of this application, it can also solve the problems in the prior art.

[0087] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0088] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the indoor light field adaptive tracking control method provided in the embodiments of this application. Multiple indoor areas are respectively equipped with lighting units, each lighting unit having a base brightness level and a working brightness level. The working brightness level has a higher illumination brightness than the base brightness level. The method includes acquiring real-time trajectory data of a target user, then determining the target user's inertial behavior pattern based on the target user's historical behavior data. The inertial behavior pattern includes at least an inertial sequence and typical dwell time in various types of rooms. Then, based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, predicting the target indoor area where the target user will be located in the next time period, and identifying the target indoor area as a focused area, and identifying all indoor areas outside the target indoor area as non-focused areas. Then, generating light field control commands to control the lighting units corresponding to the focused area to increase to the working brightness level, and controlling the lighting units corresponding to the non-focused areas to decrease to the base brightness level. This method combines the target user's real-time trajectory data, inertial behavior patterns, and spatial topology information to proactively predict the target indoor area where the target user will be in the next time period. As a result, the corresponding lighting units are turned up to working brightness in advance before the target user actually enters. This method fundamentally changes the passive response mode of traditional sound and light sensors that "turn on the light after detecting a person", solves the lighting delay problem between entering a dark area and the light fixture turning on, provides users with a seamless and continuous lighting-first experience, and significantly improves the ease of operation.

[0089] Obviously, since the computer can execute the method provided in the embodiments of this application when the computer program is executed by the processor, it can also solve the problems in the prior art.

[0090] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, provides the method provided in the embodiments of this application.

[0091] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0092] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An indoor light field adaptive tracking control method, characterized in that, Multiple indoor areas are each equipped with a lighting unit, which has a base brightness level and a working brightness level. The working brightness level has a higher illumination level than the base brightness level. The method includes: Obtain real-time trajectory data of the target user; Based on the target user's historical behavior data, determine the target user's habitual behavior pattern, which includes at least a habitual sequence and the typical duration of stay in various types of rooms; Based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the target indoor area where the target user will be located in the next time period is predicted, and the target indoor area is identified as a spotlight area, while each indoor area outside the target indoor area is identified as a non-spotlight area. Generate light field control commands to control the lighting units corresponding to the focused area to increase to the working brightness level, and control the lighting units corresponding to the non-focused area to decrease to the basic brightness level.

2. The method according to claim 1, characterized in that, Based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the target indoor area where the target user will be located in the next time period is predicted, specifically including: Under preset conditions, based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the target indoor area where the target user will be located in the next time period is predicted; wherein, the preset conditions specifically include any one or more of the following: based on the real-time trajectory data, it is determined that the actual movement direction of the target user points to the exit of the current indoor area, and the actual movement speed is greater than a first preset threshold; based on the real-time trajectory data, the actual stay time of the target user in the current indoor area is calculated, and the actual stay time is greater than a second preset threshold, wherein the second preset threshold is determined based on the typical stay time.

3. The method according to claim 1, characterized in that, Based on the real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map of each indoor area, the target indoor area where the target user will be located in the next time period is predicted, specifically including: The current indoor area of ​​the target user is determined based on the real-time trajectory data; Taking the exit of the current indoor area as the starting point of the path, and combining the inertial order in the inertial behavior pattern, determine M candidate indoor areas, where M is a positive integer greater than or equal to 2; Based on the spatial topology distribution map, calculate the path cost from the starting point of the path to each of the M candidate indoor areas, wherein the path cost includes at least one of the following: the physical distance from the starting point of the path to the candidate indoor area, and the number of turns in the path; Based on the path cost of M candidate indoor areas, N candidate indoor areas with the lowest path cost are selected from the M candidate indoor areas as the predicted target indoor areas, where N is greater than or equal to 1 and less than M.

4. The method according to claim 3, characterized in that, The method further includes: combining the service occupancy status of each indoor area, removing the target indoor areas with the service occupancy status as occupied from the predicted target indoor areas, wherein the service occupancy status includes unoccupied and occupied.

5. The method according to claim 3, characterized in that, After generating the light field control command, the method further includes: Using the spatial topology map, predictive trajectory data from the starting point of the path to each target indoor area is generated; Continuously acquire the target user's subsequent real-time trajectory data at the starting point of the path; Determine whether the subsequent real-time trajectory data deviates from all predicted trajectory data; If deviations occur, a new target indoor area is re-predicted based on the subsequent real-time trajectory data, the inertial behavior pattern, and the spatial topology distribution map, and a new light field control command is regenerated.

6. The method according to claim 1, characterized in that, Based on the target user's historical behavior data, determine the target user's habitual behavior pattern, specifically including: The historical behavior data is obtained, wherein the historical behavior data includes a time-stamped user location sequence and a task status record that is temporally associated with the user location sequence, and the task status record includes at least tasks being executed and tasks not being executed; From the historical behavior data, the user location sequence with timestamps during task execution is selected as valid job data from the task status records. Based on the effective operational data, calculate the statistical values ​​of the dwell time of the target user in each type of indoor area, and determine the statistical values ​​as the typical dwell time of the corresponding type of indoor area; Based on the effective job data, the target user's access sequences to various indoor areas and the frequency of occurrence of each indoor area access sequence are statistically analyzed when the target user completes a job task. The inertial order is determined by using one or more indoor area visit sequences that rank highest in frequency.

7. The method according to claim 1, characterized in that, The inertial order is determined using one or more indoor area visit sequences that rank highest in frequency, specifically including: Based on the most frequent one or more indoor area access sequences, calculate the conditional probability of entering other indoor areas after leaving each indoor area. The probability transition model, composed of all the aforementioned conditional probabilities, is determined as the inertial order.

8. The method according to claim 1, characterized in that, Obtaining real-time trajectory data of the target user specifically includes: when the target user completes authorized login verification, obtaining the target user's real-time trajectory data through an indoor positioning system deployed in the multiple indoor areas, wherein the indoor positioning system includes at least one of the following: an ultra-wideband positioning system, a Bluetooth beacon positioning system, a Wi-Fi fingerprint positioning system, and a camera-based visual recognition positioning system.

9. The method according to claim 8, characterized in that, By deploying an indoor positioning system in the multiple indoor areas, the real-time trajectory data of the target user is obtained. Specifically, when the target user is detected to enter a pre-defined work monitoring area, or when the target user is predicted to leave the current indoor area, the sampling frequency of the indoor positioning system is activated or increased to obtain the real-time trajectory data.

10. An indoor light field adaptive tracking control system, characterized in that, Multiple indoor areas are each equipped with a lighting unit. Each lighting unit has a base brightness level and a working brightness level, where the working brightness level is greater than the base brightness level. The system includes: The acquisition unit is used to acquire real-time trajectory data of the target user. The determining unit is used to determine the target user's habitual behavior pattern based on the target user's historical behavior data. The habitual behavior pattern includes at least a habitual sequence and the typical duration of stay in various types of rooms. The prediction unit is used to predict the target indoor area where the target user will be located in the next time period based on the real-time trajectory data, the inertial behavior pattern and the spatial topology distribution map of each indoor area, and to determine the target indoor area as a spotlight area and to determine each indoor area outside the target indoor area as a non-spotlight area. The instruction generation unit is used to generate light field control instructions to control the lighting units corresponding to the focusing area to increase to the working brightness level, and to control the lighting units corresponding to the non-focusing area to decrease to the basic brightness level.

Images