Information processing systems, methods, and programs

Abstract

The objective is to achieve efficient delivery of goods. [Solution] An information processing system in one aspect of the present invention includes an information acquisition unit that acquires location information and time information corresponding to the location information from a delivery terminal of a package delivery person, and a determination unit that determines the delivery route of the delivery person based on the location information and the time information, and determines second road information of the delivery route that is not registered in the map information by referring to first road information registered in the map information of a map application and the delivery route. A method in one aspect of the present invention includes acquiring location information and time information corresponding to the location information from a delivery terminal of a package delivery person, determining the delivery route of the delivery person based on the location information and the time information, and determining second road information of the delivery route that is not registered in the map information by referring to first road information registered in the map information of a map application and the delivery route.

Description

Technical Field

[0001] The present invention mainly relates to an information processing system, method, and program for managing package delivery.

Background Art

[0002] In physical distribution, as part of efforts to improve operational efficiency, methods for managing package delivery are provided.

[0003] For example, Patent Document 1 discloses a method of detecting uninput delivery status by a delivery person based on the position information of the delivery person.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] Although the technique disclosed in Patent Document 1 is premised on that the delivery position of a package is known in advance, on the other hand, there arises a problem that information on appropriate delivery positions and routes corresponding to the delivery destinations of packages is scarce, especially in areas with little package delivery track record such as mountainous regions.

[0006] Accordingly, an object of the present invention is to provide highly accurate map information while achieving improved delivery efficiency, particularly in areas with little package delivery track record.

Means for Solving the Problems

[0007] An information processing system in one aspect of the present invention includes an information acquisition unit that acquires location information and time information corresponding to the location information from a delivery terminal of a package delivery person, and a determination unit that determines the delivery route of the delivery person based on the location information and the time information, and determines a second road information from the delivery route that is not registered in the map information of a map application, by referring to the first road information registered in the map information of a map application and the delivery route. [Effects of the Invention]

[0008] According to the present invention, it is possible to provide more accurate map information while improving delivery efficiency, especially in areas with little experience in package delivery. [Brief explanation of the drawing]

[0009] [Figure 1] This is a block diagram showing an information processing system according to the first embodiment of the present invention. [Figure 2] Figure 1 is a functional block diagram showing the server terminal 100. [Figure 3] Figure 1 is a functional block diagram showing the delivery terminal 200. [Figure 4] This figure shows an example of delivery data stored on the server terminal 100. [Figure 5] This figure shows an example of property data stored on the server terminal 100. [Figure 6] This figure shows an example of analysis data stored on the server terminal 100. [Figure 7] This flowchart shows an example of a process for generating a map according to the first embodiment of the present invention. [Figure 8] This flowchart shows an example of a detailed process for generating a map according to the first embodiment of the present invention. [Figure 9] This flowchart shows another example of the details of the process for generating a map according to the first embodiment of the present invention. [Figure 10] This is a schematic diagram showing an example of a map display on a delivery terminal according to the first embodiment of the present invention. [Modes for carrying out the invention]

[0010] Embodiments of the present invention will be described below with reference to the drawings. The embodiments described below are not intended to unduly limit the scope of the present invention as described in the claims. Furthermore, not all components shown in the embodiments are necessarily essential components of the present invention.

[0011] <Structure> Figure 1 is a block diagram showing an information processing system according to a first embodiment of the present invention. This system 1 is managed, for example, by a delivery company or a delivery information platform operator, and consists of a server terminal 100 that manages delivery data, package data, etc. of delivery personnel and analyzes information for improving delivery efficiency, multiple delivery personnel terminals 200A, 200B of delivery personnel, and multiple user terminals 300A, 300B of users at delivery destinations. For the sake of explanation, each terminal is described as a single or specific number, but there is no limit to the number of each. Hereafter, for the sake of explanation, unless otherwise explained, delivery personnel terminals will be collectively referred to as delivery personnel terminals 200, and user terminals will be collectively referred to as user terminals 300, respectively.

[0012] The server terminal 100, the delivery terminal 200, and the user terminal 300 are each connected via the network NW1. The network NW consists of the Internet, intranet, wireless LAN (Local Area Network), WAN (Wide Area Network), etc.

[0013] The server terminal 100 may be a general-purpose computer such as a workstation or personal computer, or it may be logically implemented through cloud computing.

[0014] The deliverer terminal 200 and the user terminal 300 are, for example, information processing apparatuses such as a personal computer and a tablet terminal, but may also be configured by a smartphone, a mobile phone, a PDA, or the like.

[0015] Figure 2 is a functional block configuration diagram of the server terminal 100 in Figure 1. The server terminal 100 includes a communication unit 110, a storage unit 120, and a control unit 130.

[0016] The communication unit 110 is a communication interface for communicating with the deliverer terminal 200 and the user terminal 300 via the network NW1, and communication is performed according to a communication protocol such as TCP / IP (Transmission Control Protocol / Internet Protocol).

[0017] The storage unit 120 stores programs for executing various control processes and each function in the control unit 130, input data, and the like, and is composed of a RAM (Random Access Memory), a ROM (Read Only Memory), and the like. Further, the storage unit 120 includes a delivery data storage unit 121 that stores various data related to package delivery, a property data storage unit 122 that stores various data related to properties as delivery destinations, and an analysis data storage unit 123 that stores analysis data generated by analyzing delivery data, property data, and the like. Note that a database (not shown) storing various types of data may be constructed outside the storage unit 120 or the server terminal 100.

[0018] The control unit 130 controls the overall operation of the server terminal 100 by executing programs stored in the storage unit 120, and is composed of a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), and the like. As functions of the control unit 130, it at least includes: an information acquisition unit 131 that receives instructions and information from each terminal, a determination unit 132 that determines the delivery position and the like of a package delivered by a deliverer based on the received information, a generation unit 133 that generates information such as the delivery position and screen information for displaying the information as a result of the above determination, and a display unit 134 that causes each terminal to display the generated information and the like. The information acquisition unit 131, the determination unit 132, the generation unit 133, and the display unit 134 are activated by the programs stored in the storage unit 120 and executed by the server terminal 100 which is a computer (electronic computer).

[0019] The information acquisition unit 131 receives predetermined information from the deliverer terminal 200 or the like via the communication unit 110. For example, it receives position information of the deliverer and time information corresponding to the position from the deliverer terminal 200.

[0020] The information acquisition unit 131 can acquire the position information and the time information from the deliverer terminal 200 in various ways. For example, there are a method of periodically polling the deliverer terminal 200, a method of receiving the information as a push notification from the deliverer terminal 200, or a method of acquiring the information based on a preset condition (for example, when the moving distance of the deliverer exceeds a predetermined threshold, or when the deliverer enters a specific area, etc.).

[0021] Further, the information acquisition unit 131 may include a function of evaluating the accuracy and reliability of the position information, and correcting or excluding position information with low accuracy. In this function, a quality score of the position data is calculated using indicators such as GPS accuracy information, the number of satellites, and signal strength, to detect data that falls below a predetermined threshold. For the detected low-accuracy data, interpolation processing using a statistical method can be applied, or the data can be replaced with a position estimated from the high-accuracy data before and after it.

[0022] In addition to location and time information, the information acquisition unit 131 can also acquire sensor information such as acceleration sensor data, gyroscope data, and barometric pressure sensor data from the delivery terminal 200. By utilizing this auxiliary data, the delivery person's location can be estimated even in environments where GPS is unstable (such as high-rise building districts, tunnels, and underground parking lots). Furthermore, by combining this data with weather information, traffic information, and event information obtained from external data sources, a comprehensive analysis of the delivery situation becomes possible.

[0023] Furthermore, the information acquisition unit 131 can also be equipped with a function to dynamically adjust the frequency and accuracy of data acquisition. For example, an adaptive sampling strategy can be implemented, such as increasing the frequency of location information acquisition when the delivery person is moving at high speed and decreasing the frequency when they are stationary. This makes it possible to optimize the battery consumption of the delivery person terminal 200 while maintaining the accuracy of location tracking.

[0024] The determination unit 132 performs a predetermined determination process based on the received information. For example, it performs a process to determine the delivery location and delivery route where the delivery person delivered the package, based on the location information and time information.

[0025] In determining delivery routes, high-precision determination is possible by combining multiple approaches. Specifically, the following determination methods can be implemented individually or in combination:

[0026] (1) Determination method based on the delivery person's stay time: A method for identifying a location as a delivery location candidate when a delivery person stays at a specific location for a certain period of time or longer. In this method, first, the determination unit 132 of the control unit 130 extracts multiple data points from the location information data in which the change in location coordinates is within a predetermined threshold (e.g., 5 meters). Next, it analyzes the time information corresponding to these data points and calculates the stay time by calculating the elapsed time from the time of the first data point to the time of the last data point. The stay time threshold is dynamically adjusted according to the characteristics of the delivery area (urban areas, suburbs, mountainous areas, etc.), time of day, and weather conditions. Specifically, it refers to the rule table stored in the analysis data storage unit 123 of the memory unit 120 and obtains the standard stay time set for each combination of region type, time of day, and weather type. For example, a relatively short standard value (3 minutes) is set for urban areas, a medium standard value (5 minutes) for suburban areas, and a longer standard value (7 minutes) for mountainous areas, and further adjustments are made such as multiplying by a coefficient of 1.2 in rainy weather and a coefficient of 1.3 at night. Finally, if the calculated actual dwell time exceeds the adjusted threshold, that location is marked as a delivery location candidate and recorded in the delivery data storage unit 121.

[0027] (2) Determination method based on the number of times a delivery person has stayed at the same location: A method of determining a location as a delivery location when the number of times a delivery person has stayed at a particular location exceeds a predetermined frequency threshold. Specifically, the determination unit 132 spatially clusters location data in order to detect locations from the location information of delivery persons. In this clustering process, location information is plotted in a two-dimensional space of latitude and longitude, and points that are close to each other (for example, within 20 meters) are grouped together as the same cluster. The center position of each cluster is calculated and used as the representative coordinate of the location of stay. Next, the number of times the same or nearby delivery persons have stayed at the same or nearby clusters is aggregated from past delivery data stored in the delivery data storage unit 121. In this aggregation process, the number of stays is calculated for each period of short term (past 7 days), medium term (past 30 days), and long term (past 90 days), and compared with a frequency threshold according to regional characteristics and the type of delivery route. The frequency threshold varies according to the importance of the route, and is set to a low value (about twice a week) for main routes and a high value (four or more times a week) for branch routes. Locations where the number of stays exceeds the above threshold are determined to be delivery locations, and this information is stored in association with the property data storage unit 122. Furthermore, when integrating stay information from multiple delivery personnel, the data is weighted based on each delivery person's performance and confidence score before aggregation, thereby achieving a more accurate determination.

[0028] (3) Determination method based on pattern analysis of the delivery person's movement path: A method for estimating the delivery location by detecting characteristic patterns (stopping points, turning points, dead ends, etc.) from the delivery person's movement trajectory data. In this method, the determination unit 132 first calculates the movement speed and direction from the time-series location information. Specifically, it calculates the speed by dividing the distance between two consecutive location information points by the difference in corresponding time information, and calculates the direction of movement from the vector between the two points. Next, in order to detect characteristic patterns from the calculated speed data, it identifies characteristic points of the speed profile (minimum values, points of sudden deceleration, values ​​near zero, etc.). In particular, it extracts points where the movement speed between consecutive location information points is below a predetermined low-speed threshold (e.g., 2 km / h or less), and also analyzes the direction of movement before and after these points. Points where the direction of movement changes significantly (e.g., 90 degrees or more) or where the direction of movement reverses (e.g., a change of 140 degrees or more) are likely to show characteristics of delivery behavior and are therefore identified as candidate delivery locations. Furthermore, in detecting dead-end patterns, a sequence is identified in which the vehicle moves slowly and then moves in the opposite direction along the same path, and the turning point is designated as a delivery candidate. Each detected pattern location is assigned a confidence score, and locations where multiple patterns overlap (for example, where the stopping point and the turning point coincide) are given particularly high scores and are given priority as delivery locations. This information is stored in the analysis data storage unit 123 and is also used as reference data for improving the accuracy of future determinations.

[0029] (4) Determination method based on association between delivery completion notification and location information: A method for identifying the delivery location by analyzing the location information at the time of the delivery completion operation of the delivery terminal 200. In this implementation, the determination unit 132 first associates the delivery completion operation event input through the user interface of the delivery terminal 200 with the location information before and after that operation. Specifically, it extracts location information for 2 minutes before and after the time the delivery completion operation was performed (a total of 4 minutes). The extracted set of location information is weighted based on the GPS accuracy information (number of satellites, signal strength, estimated error radius, etc.) of each data point. High-precision location information is assigned a high weight (e.g., 1.0), and low-precision location information is assigned a low weight (e.g., 0.3). Next, the centroid position (weighted average position) of the weighted location information is calculated and this is set as the first candidate for the delivery location. In addition, the most frequently occurring location within the extracted set of location information (after applying appropriate spatial clustering) is identified and this is set as the second candidate for the delivery location. If the distance between the first and second candidates is short (for example, within 10 meters), their average position is determined as the final delivery location. If they are far apart, the candidate with higher GPS accuracy is determined as the final delivery location. The determined delivery location is stored in the delivery data storage unit 121 along with the delivery event ID and the delivery item ID. Furthermore, if there is a corresponding property ID, it is also recorded in the property data storage unit 122 in association with it. This process makes it possible to identify the delivery location that accurately corresponds to the actual delivery actions of the delivery person.

[0030] (5) Determination method based on comparison with past delivery data: A method for estimating the delivery location in the current delivery by referring to past delivery location data for the same or similar delivery destinations. In this method, the determination unit 132 first extracts past delivery event data from the delivery data storage unit 121 that has the same property ID or the same address information as the current delivery event. Delivery location information is collected from the extracted data and its spatial distribution is analyzed. Specifically, density-based clustering (e.g., DBSCAN algorithm) is applied to the collected location information to identify high-density areas (areas where many delivery locations are concentrated). Each cluster is assigned a weighted score based on the number of delivery location data it contains and the recency of the data. For example, data from the last 30 days is assigned a high score (1.0), data from 30 to 90 days is assigned a medium score (0.7), and data from more than 90 days ago is assigned a low score (0.4). The overall score for each cluster is calculated, and the center of the cluster with the highest score is designated as the most likely delivery location candidate. Furthermore, a probability density map is generated based on the extracted past delivery location data, and by overlaying it with the current delivery person's travel route, a confidence score is calculated to evaluate how well each point on the route matches a past delivery location. By combining this confidence score with the results of other determination methods, it becomes possible to determine the delivery location more accurately. These analysis results are stored in the analysis data storage unit 123 and are continuously updated, contributing to improved determination accuracy.

[0031] The determination unit 132 calculates a confidence score for each determination result obtained from the methods (1) to (5) above, and determines the final delivery location by integrating multiple determination results. The calculation of the confidence score takes into account the reliability of each determination method, the weighting according to the situation, and the past determination success rate. For example, in urban areas, a high weight is given to the delivery completion notification method (4), and in mountainous areas, a high weight is given to the movement pattern analysis method (3) and the comparison of past data method (5). Adaptive weighting is performed according to the environment. If similar locations are indicated by multiple determination methods (for example, within 20 meters of each other), a highly confident integrated evaluation is performed. If significantly different locations are indicated, the determination result with the highest confidence is adopted, or a flag is set indicating that further verification is needed. The determination unit 132 records and analyzes the history of these determination results in the analysis data storage unit 123, and uses this information for automatic adjustment of determination parameters and continuous improvement of the determination logic. Furthermore, by also recording the reliability index of the judgment result, it can be used to evaluate the reliability of the second road information in the subsequent map generation process.

[0032] Furthermore, the determination unit 132 can also implement advanced predictions using machine learning models. In this machine learning-based delivery location prediction system, past delivery data stored in the delivery data storage unit 121 and analysis data storage unit 123 of the memory unit 120 is used as a training dataset. During the training dataset construction phase, various features such as location information (latitude and longitude), time information (time, day of the week, month, etc.), delivery person information (ID, years of experience, etc.), property information (type, number of past deliveries, etc.), environmental information (weather, season, etc.), and movement pattern features (stopping point patterns, frequency of direction changes, etc.) are extracted and processed. These features are combined with actual delivery location data as ground truth labels to create a dataset for supervised learning. The determination unit 132 uses this dataset to train multiple machine learning algorithms. For example, the decision tree algorithm automatically learns branching conditions based on feature thresholds and constructs rules for predicting delivery locations from various feature combinations. In areas with particularly complex terrain or diverse delivery patterns, a more advanced random forest model is applied, and accuracy and stability are improved by integrating the prediction results of multiple decision trees. In parallel, a recurrent neural network (RNN) model, which excels in analyzing time-series data, is trained to predict delivery locations from the temporal patterns of the delivery person's movement trajectory. Each model is trained and evaluated independently, and its prediction accuracy is periodically verified. Furthermore, an ensemble learning framework is implemented to integrate the prediction results of these different algorithms. In this ensemble method, the prediction results of each model are dynamically weighted according to the model's past accuracy performance and the current situation, and the most reliable delivery location is output as the final prediction. This machine learning system is periodically retrained, and is designed to continuously improve prediction accuracy as new delivery data is accumulated.

[0033] Furthermore, the determination unit 132 can also utilize 3D location information. By using location information including elevation data and altitude information estimated from barometric pressure sensor data, it is possible to pinpoint delivery locations within high-rise buildings and precise locations in mountainous areas. In this 3D location analysis, the information acquisition unit 131 first acquires altitude information (elevation or altitude above sea level) in addition to normal location information (latitude and longitude) from the delivery terminal 200. Two methods are implemented for acquiring altitude information: directly from GPS and by comparing the barometric pressure sensor data of the delivery terminal 200 with a reference pressure to estimate altitude changes. Measurement using the barometric pressure sensor is often more accurate (approximately ±1m) than the vertical accuracy of GPS (usually around ±10m), and functions effectively even in environments with weak satellite signals, such as inside high-rise buildings. The acquired 3D location data is processed in conjunction with a Geographic Information System (GIS) and compared with elevation databases and building information databases. For example, in mountainous areas, comparing detailed digital elevation model (DEM) data provided by the Geospatial Information Authority of Japan with location information enables accurate location identification based not only on planar coordinates but also on actual terrain. Within high-rise buildings, detailed analysis of vertical movement patterns within the building allows for the detection of elevator and stair use and the understanding of the floor structure. Specifically, rapid altitude changes (5-30m) over short periods (30 seconds to 1 minute) are identified as elevator use, and the elevator's location can be determined by combining this with the horizontal movement before and after. On the other hand, more gradual altitude changes (continuous changes of 2-5m per minute) are identified as stair use. By combining these vertical and horizontal movement patterns, delivery locations on each floor of the building can be identified, and a three-dimensional delivery map can be constructed. Furthermore, by integrating and analyzing three-dimensional location data collected from multiple delivery personnel, information such as the building's floor structure and delivery frequency on each floor is accumulated and used for more efficient delivery route planning. These 3D location data and analysis results are stored in the property data storage unit 122 and the analysis data storage unit 123, and are used in subsequent delivery operations.

[0034] The generation unit 133 performs a process to generate predetermined information based on the determination result. For example, based on the delivery location obtained as a determination result, it performs a process to generate information indicating the delivery location and / or screen information. Here, when generating screen information, the generation unit 133 uses the image and text data stored in the storage unit 120 as material and generates screen information for the user interface by arranging various images and text in predetermined areas of the user interface according to predetermined layout rules. Furthermore, the processing related to the generation of screen information can also be performed by a GPU (Graphics Processing Unit).

[0035] The generation unit 133 can perform advanced map generation processing that integrates multiple map information sources and delivery data when supplementing the first road information with the second road information. Specifically, it can have the following functions:

[0036] (1) Multiple map service support function: The generation unit 133 utilizes APIs from multiple map services, such as Google Maps, OpenStreetMap, and geographical survey maps of various countries, to acquire the first road information. In this function, the information acquisition unit 131 first acquires map data (road network, topographic information, facility information, etc.) for the target area through the API of each map service. The acquired map data is unified into a common coordinate system (e.g., WGS84) and managed as map layers. The generation unit 133 selectively combines these map layers according to regional characteristics and applications to construct the optimal first road information. Specifically, a regional classification module is implemented to determine the type of region (urban, suburban, mountainous, etc.), and the regional type of the current location is automatically determined based on GPS coordinates. If it is determined to be an urban area, data from Google Maps or private map services with comprehensive commercial facilities and address information is given priority, and POI (Point of Interest) information and detailed road attributes (one-way streets, number of lanes, etc.) are adopted as the first road information. On the other hand, in mountainous and rural areas, precise topographic information and contour data provided by the Geospatial Information Authority of Japan and other sources are highly valued, and road information obtained from these services is prioritized as primary road information. The primary road information obtained from each map service is assigned a reliability score, and the most reliable data is prioritized based on the freshness of the information, the frequency of updates, and the accuracy evaluation. The primary road information thus optimized is used as a base map to effectively complement the secondary road information.

[0037] (2) Road information generation and supplementation function: The generation unit 133 implements a function to extract roads and paths that do not exist in the first road information from the delivery person's movement trajectory data as second road information and supplement the map information. In this process, first, time-series location information data of multiple delivery events is extracted from the delivery data storage unit 121 and trace data that can be used as candidate movement routes is generated. In addition to GPS coordinates and acquisition time, metadata such as movement speed, direction and GPS accuracy information is added to this trace data. Next, spatial clustering is applied to this trace data to identify patterns in which multiple delivery persons pass through similar routes. In the clustering process, density-based algorithms such as DBSCAN (Density-Based Spatial Clustering of Applications with Noise) are used to group spatially close trace points. The clustering distance parameter is variable according to the road type, with a wider threshold (e.g., 20m) applied to major roads and a narrower threshold (e.g., 5m) applied to paths. From the clustering results, route patterns with sufficient reliability (for example, routes confirmed to have been passed through three or more times by different delivery drivers) are extracted and identified as candidates for second road information. This candidate road data is compared with existing first road information by the road verification module in the generation unit 133, and only new road sections not present on the official map are extracted as second road information. Attribute information based on travel history is assigned to the detected second road information. Specifically, passability estimated from the travel speed of the trace data (e.g., always passable, pedestrian-only, seasonally passable), road surface condition estimated from the vibration pattern of the trace data (e.g., paved, unpaved, gravel road), and road width estimated from the distance between multiple traces (e.g., one lane, two lanes, sidewalk width), are recorded. This information, along with a confidence score, is registered in the second road information database of the analysis data storage unit 123 and stored as high-quality data to complement the first road information. Ultimately, the generation unit 133 generates a more complete and accurate composite map by superimposing or combining this second road information with the first road information, thereby contributing to the efficiency of delivery operations.

[0038] (3) Route Optimization Function: The generation unit 133 implements a function to calculate the optimal delivery order and route based on a composite map that integrates the first road information and the second road information, taking into account conditions such as the location information of the delivery destination, estimated travel time, road conditions, and delivery time slots. In this route optimization process, first, a list of the day's scheduled deliveries (delivery destination location, delivery time slot, cargo characteristics, etc.) is obtained from the delivery data storage unit 121, and a travel cost matrix is ​​calculated between each delivery destination. In the travel cost calculation, the actual travel distance is calculated using the road network of the composite map that combines the first road information and the second road information, rather than simple straight-line distance. In particular, unpaved roads and narrow roads detected as second road information are reflected in the cost calculation according to the vehicle type and weather conditions. For example, for large vehicles, a high cost is set for narrow second road information routes, and for small vehicles and deliveries on foot, these shortcut routes are effectively utilized in the calculation model. Furthermore, attribute data such as average driving speed by time of day, past traffic congestion information, and gradient information are assigned to each road section, and travel time predictions are made taking these factors into account. In addition, differences in fuel consumption due to uphill and downhill slopes, and traffic restrictions (one-way streets, no right or left turns, time-based traffic restrictions, etc.) are also reflected in the travel cost. Based on the multidimensional travel cost information calculated in this way, the generation unit 133 applies a delivery order optimization algorithm. In this optimization process, a solution is sought that minimizes the total travel time and total travel distance while satisfying constraints such as arrival time constraints at the delivery destination (delivery must be made within a specified time frame), vehicle constraints (loading capacity, driving time limits, etc.), and priority (prioritizing urgent deliveries, etc.). The optimization results are visually displayed in an easy-to-understand manner on a composite map containing the first and second road information on the delivery terminal 200, and information such as recommended routes to each delivery destination, estimated arrival times, and estimated stay times are provided in a timeline format.

[0039] (4) Three-dimensional map generation function: The generation unit 133 implements a function to generate a 3D map by integrating elevation data and building floor information into a composite map that includes first road information and second road information. In this 3D map generation process, elevation information is first integrated into a composite map that is created by supplementing the second road information with the basic first road information. Elevation information is obtained from multiple sources, such as digital elevation model (DEM) data provided by the Geospatial Information Authority of Japan, terrain models generated from satellite images, and point cloud data obtained from aerial laser surveying (LiDAR). This elevation data is converted into a common coordinate system, and a continuous elevation map is generated by interpolation processing. In mountainous areas, a 3D mesh model is constructed based on this elevation map, in which the topography is represented three-dimensionally. On this model, major roads based on the first road information are displayed as thick three-dimensional lines, while smaller roads and unofficial routes based on the second road information are displayed as thin three-dimensional lines or dotted lines. Each road is positioned according to the actual terrain, and by accurately reflecting road gradient information, it supports the visual understanding of steep uphill and downhill slopes and curves. Furthermore, color-coding based on road type (first road information / second road information) and gradient degree, as well as warning icons for particularly steep sections requiring special attention, are implemented. In urban areas, 3D building shape data is integrated to represent the three-dimensional structure of high-rise buildings and mixed-use facilities. This building data is accompanied by floor information, allowing for the identification of underground parking lots, ground-floor entrances, and individual floors. For commercial facilities and apartment buildings with high delivery frequencies, the location of elevators and stairs, and delivery points on each floor are visualized on the 3D model based on the vertical movement pattern analysis results within the building provided by the judgment unit 132. The second road information plays a crucial role in representing shortcuts and back entrance access routes between buildings, contributing to the accurate identification of building access points. In addition, a route guidance function optimized for 3D display is implemented, displaying delivery routes traversing three-dimensional space in a three-dimensional manner, distinguishing between the first and second road information. In this way, the generation unit 133 helps delivery personnel understand the terrain and building structure more intuitively by transforming planar map information into a three-dimensional object, enabling them to select efficient delivery routes.

[0040] (5) Augmented Reality (AR) Display Function: The generation unit 133 implements an AR function that superimposes composite map information, including first road information and second road information, onto the camera image. In this AR function, real-time video is first acquired from the camera of the delivery terminal 200, and at the same time, location information (GPS) and orientation information (direction, tilt, rotation, etc. obtained from the accelerometer, gyroscope, and magnetometer) of the terminal are also collected. An AR alignment module is implemented that integrates and processes this sensor information and video information, and accurate mapping is performed between real-space coordinates within the camera's field of view and in-video coordinates. Visual SLAM (Simultaneous Localization and Mapping) technology is also introduced, which extracts feature points from the camera image and tracks them to maintain a stable AR display even with GPS errors. The AR content generation module generates digital content that visually distinguishes between the first road information and the second road information. Specifically, major roads based on the first road information are represented by thick solid lines and bright colors, while unofficial routes and side streets based on the second road information are represented by thin dashed lines and different colors. This allows delivery personnel to visually distinguish between official map information and supplementary information obtained from actual delivery records. In the AR navigation display, 3D arrow objects indicating the direction to go, information panels showing the distance to the destination and the estimated arrival time, and special markers indicating the entrance to unofficial shortcut routes based on the second road information are placed in accordance with the real space. In particular, when entering a route based on the second road information that is not displayed in the map application, highlighting and warning displays are provided to help delivery personnel use unofficial routes with confidence. In addition, a visibility optimization function automatically adjusts the color, transparency, and size of the AR content according to the surrounding ambient light and background complexity, so that the first and second road information can be clearly distinguished under various environmental conditions. These AR features allow delivery personnel to intuitively understand the optimal route, including secondary road information not found on official maps, while viewing the actual scenery, thereby improving operational efficiency, especially in complex delivery environments.

[0041] (6) Personalized display function: The generation unit 133 implements a function to customize the display of a composite map including first road information and second road information according to the experience level of each delivery person, preferred navigation style, and characteristics of the device used. This personalization function is centered around the user profile management module, and each delivery person's attribute information (years of experience, assigned area history, delivery performance data, etc.), setting information (display language, map style settings, display level of second road information, etc.), and usage history (frequently used functions, operation patterns, frequency of use of second road information, etc.) are managed in an integrated manner. In the display customization process, the amount of information is first adjusted according to the delivery person's experience level. For new delivery people (less than one year of experience or few visits to their assigned area), reliable information, mainly first road information, is displayed preferentially, and only second road information with a particularly high reliability is displayed. In addition, when using second road information, a wealth of auxiliary information such as precautions and usage tips is provided. On the other hand, experienced delivery drivers (3 years or more of experience or frequent visits to their assigned area) are shown a complete composite map integrating primary and secondary road information, with the reliability of the information visualized through a gradient display based on reliability. Furthermore, a settings interface is provided that allows delivery drivers to adjust the display level of secondary road information, with options such as "Show primary road information only," "Add only highly reliable secondary road information," and "Show all secondary road information." The visual representation of secondary road information (color coding, line type, icons, etc.) can also be customized according to the delivery driver's preferences. In addition, an adaptive learning function is implemented, recording and analyzing the secondary road information routes actually used and avoided by delivery drivers, and dynamically adjusting the reliability score of the secondary road information based on each delivery driver's behavior patterns. As a result, secondary road information that delivery drivers frequently use is gradually highlighted, and secondary road information that is avoided is given a lower display priority, achieving personalized optimization based on usage history.This personalization feature allows delivery drivers to efficiently utilize a composite map of primary and secondary road information optimized to their experience and preferences, thereby reducing the burden of delivery work and improving operational efficiency.

[0042] The generation unit 133 stores the information obtained from these generation processes in the analysis data storage unit 123 of the storage unit 120 and can also transmit it to the delivery terminal 200. The transmitted information is optimized considering the hardware performance, network status, battery level, etc. of the delivery terminal 200, and a mechanism can be implemented to display it in the most optimal format for each terminal.

[0043] The display unit 134 performs the process of displaying the generated information via a user interface such as the delivery terminal 200. For example, the display unit 134 transmits screen information to the delivery terminal 200, and the delivery terminal 200 displays the screen via its user interface based on the received screen information.

[0044] In the implementation of the display unit 134, an adaptive rendering function can be provided that corresponds to the characteristics of the delivery terminal 200 (screen size, resolution, processing power, etc.). This function automatically adjusts the map rendering quality and animation effects according to the performance of the terminal, providing an optimal user experience for various terminals. For example, it is possible to optimize the display by providing high-resolution 3D maps and smooth animation effects on high-performance terminals, and automatically switching to a simplified 2D display on low-performance terminals.

[0045] Furthermore, the display unit 134 can provide multiple display modes. For example, standard mode (normal map display), detailed mode (displays all information), simple mode (displays only important information in a larger size), night mode (color scheme that is easy to see even in dark environments), and power saving mode (display that minimizes battery consumption). Delivery personnel can manually switch between these modes depending on the situation, and it is also possible to switch them automatically based on conditions such as ambient light and battery level.

[0046] Furthermore, a voice guidance function can also be implemented. This function provides voice-guided route directions and arrival notifications even when the delivery person cannot focus on the screen (such as while driving). The voice guidance uses speech synthesis technology to generate natural-sounding instructions and is provided in the language specified by the delivery person. In addition, by linking with external audio devices such as headsets or Bluetooth earphones, clear voice guidance is possible even in noisy environments.

[0047] The display unit 134 also features a visual feedback function. Depending on the delivery person's actions (approaching the destination, deviating from the planned route, etc.), it provides visual notifications (highlighting, alert icons, etc.) and vibration feedback to appropriately attract the delivery person's attention. These feedbacks can be customized according to the delivery person's preferences and circumstances, and a frequency adjustment function is also implemented to prevent distraction due to excessive notifications.

[0048] These advanced display features allow delivery personnel to perform delivery tasks intuitively and efficiently, resulting in improved delivery efficiency and reduced burden on delivery personnel.

[0049] Figure 3 is a functional block diagram showing the delivery terminal 200 of Figure 1. The delivery terminal 200 includes a communication unit 210, a display and operation unit 220, a storage unit 230, a control unit 240, and a GPS 250.

[0050] The communication unit 210 is a communication interface for communicating with the server terminal 100 via the network NW, and communication is performed using a communication protocol such as TCP / IP.

[0051] The display operation unit 220 is a user interface used by the delivery person to input instructions and display text, images, etc., in accordance with the input data from the control unit 240. If the delivery person terminal 200 is a smartphone or tablet, the display operation unit 220 consists of a touch panel or the like. This display operation unit 220 is activated by a control program stored in the memory unit 230 and executed by the delivery person terminal 200, which is a computer (electronic calculator).

[0052] The memory unit 230 stores programs for executing various control processes and functions within the control unit 240, input data, etc., and is composed of RAM, ROM, etc. The memory unit 230 can also temporarily store the contents of communication with the server terminal 100.

[0053] The control unit 240 controls the overall operation of the delivery terminal 200 by executing a program stored in the memory unit 230, and is composed of a CPU, GPU, and the like.

[0054] GPS250 refers to the Global Positioning System, which uses artificial satellites to determine the current location and time (time information) of the delivery terminal 200.

[0055] Figure 4 shows an example of delivery data stored in the server terminal 100.

[0056] As shown in Figure 4, delivery data 1000 stores various data related to package delivery managed by the delivery company. As various data related to package delivery, first, a delivery ID (for example, a delivery event identified as "10001") is assigned as information to identify a single delivery event (delivery request, etc.), and associated with the delivery ID, package information (for example, information to identify the package (package ID, etc.), information about the sender (requester, shipper), information about the destination (recipient, recipient) (property ID, contact information, delivery location, etc.), other information such as date (reception date and time, specified delivery date and time, delivery completion date and time, etc.), package contents, package weight, fee, responsible delivery office, delivery vehicle, etc.), and delivery person information (information to identify the delivery person (delivery person ID, etc.), delivery person location information, time information corresponding to the location (time, etc.), delivery method (truck, motorcycle, UAV (drone), etc.), time spent at a specific location, delivery location where the package was delivered, parking location where the vehicle was parked, etc.) may be included, but is not limited to these.

[0057] Figure 5 shows an example of property data stored in the server terminal 100.

[0058] As shown in Figure 5, property data 2000 stores various data related to the property (house, facility, business premises, etc.) corresponding to the delivery destination of the package. As various data related to the property, a property ID (for example, a property identified as "20001") is assigned as information to identify the property, and associated with the property ID, property information (for example, location information to identify the location of the property, parking locations available for delivering packages to the property, information about the recipient (for example, name, gender, age, etc.), other information such as the type of property (detached house, apartment building, office, etc.)), and delivery information associated with that property (for example, delivery person's stay time, delivery location, drop-off location, number of deliveries, etc.) may be included, but is not limited to these.

[0059] Figure 6 shows an example of analysis data stored in the server terminal 100.

[0060] As shown in Figure 6, the analysis data 3000 stores information related to the analysis results obtained based on information received from the delivery person, and information referenced during the analysis. The analysis data 3000 may include, but is not limited to, map information obtained from external applications, etc., that the delivery person uses to identify the route and destination for delivery of the package, heat map information generated by superimposing on the map information based on the delivery person's location and time information, road information generated by superimposing on the map information to identify roads for the land transport of the package, route information that identifies the route to the delivery destination of the package, property information associated with the delivery destination of the package, and dwell time information related to the average dwell time of multiple delivery persons, etc.

[0061] <Processing flow> Referring to Figures 7 and onward, the processing method for streamlining package delivery performed by System 1 of this embodiment will be described. First, Figure 7 is a flowchart showing an example of the process for generating a map according to the first embodiment of the present invention.

[0062] First, as a premise, each delivery person (deliverer) of the delivery company carries a delivery terminal 200 and delivers packages requested by the consignor to a designated destination via delivery means such as a truck. In this case, especially in mountainous areas where roads are not well-developed or accessible, it is assumed that the delivery person will park the vehicle in a parking space near the destination, carry the packages on foot to the destination, and finally deliver the packages. Alternatively, it is also assumed that the packages will be delivered by leaving them at a designated delivery location specified by the user at the destination.

[0063] Next, as part of step S101, the information acquisition unit 131 of the control unit 130 of the server terminal 100 receives the delivery person's location information from the delivery person's terminal 200. Here, the information acquisition unit 131 can receive the delivery person's current location information at predetermined time intervals or through batch processing, receiving location information corresponding to each of multiple time points. The information acquisition unit 121 stores the received location information as delivery data 1000 in the delivery data storage unit 121 of the storage unit 120.

[0064] Next, as part of step S102, the information acquisition unit 131 of the control unit 130 of the server terminal 100 receives the location information along with time information (for example, date and time) corresponding to the location information. The information acquisition unit 131 associates the received time information with the location information and stores it as delivery data 1000 in the delivery data storage unit 121 of the storage unit 120.

[0065] Next, as part of step S103, the determination unit 132 of the control unit 130 of the server terminal 100 determines the delivery route to the delivery location of the package by the delivery person based on the received location and time information of the delivery person. Specifically, the determination unit 132 determines the delivery route by tracing multiple points to the delivery location based on multiple location information at different times obtained from the delivery person terminal 200. Here, the server terminal 100 can also similarly obtain location and time information from the same or different delivery persons, associated with multiple delivery events (delivery IDs) for the same or different delivery destinations, and determine the delivery route to each delivery location. The determination unit 132 stores the information regarding the delivery route determined in this way in the delivery data storage unit 121, the property data storage unit 122, and / or the analysis data storage unit 123 of the storage unit 120, associated with the delivery location and property information.

[0066] Here, the determination unit 132 can calculate the time spent at each location based on the location information and time information, and determine the delivery location and parking location of the package based on the time spent there.

[0067] Next, as part of step S104, the generation unit 133 generates a map based on the determined delivery route. Here, the map based on the delivery route is a map that traces the delivery route for the delivery of goods by the delivery vehicle or delivery person, and can also be a map that distinguishes between the roadway traveled by the delivery person and the sidewalk traveled by the delivery person based on the determined delivery location and parking location. Furthermore, map information can be generated so that the delivery location and parking location can be displayed on the generated map. The generation unit 133 stores the generated map information as analysis data 3000 in the analysis data storage unit 123 of the storage unit 120.

[0068] Figure 8 is a flowchart showing a detailed example of the map generation process according to the first embodiment of the present invention.

[0069] First, as part of step S201, the information acquisition unit 131 of the control unit 130 of the server terminal 100 acquires a map application (for example, Google Maps®) from a third-party map application server via an API or the like. The server terminal 100 can also store the data related to the map application acquired in this way in the analysis data storage unit 123 of the storage unit 120 beforehand.

[0070] Next, as part of step S202, the determination unit 123 of the control unit 130 of the server terminal 100 compares the map information of the acquired map application with the stored map information based on the delivery route and determines whether there is a difference. For example, a difference could be the difference between road information registered in the map application (first road information) and road information that is not registered in the map application but is registered in the map information based on the delivery route (second road information).

[0071] If the above determination result is that there is no difference, as the process of step S203, the generation unit 123 of the control unit 130 of the server terminal 100 can generate the map information of the acquired map application for display on the delivery terminal 200. If there is a difference, that is, if there is road information that is not registered in the map application but is registered in the map information based on the delivery route, as the process of step S204, the road information registered in the map information based on the delivery route, i.e., the difference in road information, is supplemented to the map information of the acquired map application, and as the process of step S205, it is generated for display on the delivery terminal 200 as the map information of the map application. In particular, in areas such as mountainous areas and remote islands, it is conceivable that there is road information that is not covered by existing map applications, and roads that can be detected by delivery routes based on the location information of delivery persons who deliver packages on a daily or regular basis can be used as map information in a way that supplements the map application.

[0072] Figure 10 is a schematic diagram showing an example of map information displayed on a delivery terminal according to the first embodiment of the present invention.

[0073] As shown in Figure 10, for example, on a map displayed via a map application, map information including road information is displayed by supplementing the first road information, which consists of major roads registered in the map application and shown with thick lines, with second road information, which consists of non-major roads (branch routes) shown with dashed lines. The first and second road information can be displayed separately and identifiable, or they can be displayed as a single unit without identification. In addition, delivery locations and parking locations can be indicated on this map to correspond to the delivery destination property.

[0074] Figure 9 is a flowchart showing a detailed example of a process for generating map information (a process for supplementing map information) according to the first embodiment of the present invention.

[0075] First, as part of the process in step S301, the information acquisition unit 131 of the control unit 130 of the server terminal 100 acquires a map application (for example, Google Maps®) from a third-party map application server via an API or the like, as described above. The server terminal 100 can also store the data related to the map application acquired in this way in the analysis data storage unit 123 of the storage unit 120 beforehand.

[0076] Next, as part of step S302, the generation unit 123 of the control unit 130 of the server terminal 100 refers to the map application and extracts the portion that can be identified as a route by major roads within the range registered in the map application as first road information.

[0077] Next, as part of the processing in step S203, the generation unit 123 of the control unit 130 of the server terminal 100 synthesizes the second road information (branch routes (roads not registered in the map application, unpaved roads, etc.) which has been determined to be the difference) with the first road information by superimposing the second road information onto the first road information, thereby supplementing the first road information with the second road information and generating the final road information. By synthesizing this road onto the map application's map, a composite map can be generated.

[0078] In the system of the present invention, as delivery data accumulates, advanced prediction and optimization functions can be implemented using machine learning models that utilize this data. These machine learning models can be used in various aspects, such as determining delivery routes, determining second road information, identifying delivery locations, and optimizing delivery efficiency.

[0079] The determination unit 132 can implement a second road information determination function that utilizes a machine learning model. In this implementation, past delivery data (location information, time information, delivery route, etc.) stored in the delivery data storage unit 121 and analysis data storage unit 123 of the memory unit 120 is processed as training data, and a supervised learning algorithm is applied. Specifically, the determination unit 132 can implement three core modules: a preprocessing module for converting time-series data of location information into input features, a prediction module for predicting the second road information from the features, and an evaluation module for evaluating the reliability of the prediction results.

[0080] The preprocessing module applies noise reduction filters, feature extraction (velocity calculation, acceleration calculation, direction change detection, etc.), and normalization to raw GPS data, converting it into a format suitable for input to machine learning models. It also extracts temporal features such as time of day, day of the week, and season from time information, enabling the learning of the time dependence of delivery patterns.

[0081] The prediction module can implement a deep learning architecture to accurately identify second road information not registered in the map application from the delivery route. These models can effectively learn spatiotemporal patterns of location data. The determination unit 132 optimizes model parameters using gradient descent during the training phase and can evaluate the generalization performance of the model through cross-validation using a validation dataset.

[0082] The generation unit 133 can utilize the output of a machine learning model to perform a process that supplements the first road information with more accurate second road information. In particular, by analyzing location data collected from multiple delivery events and adding only reliable second road information to the map information, useful information can be added while maintaining the accuracy of the map.

[0083] Furthermore, as a variation of the information processing system of this embodiment, the system may include an offline operation mode to ensure operation in environments where network connectivity is unstable or impossible. This mode is particularly designed to support delivery operations in mountainous areas and regions where communication infrastructure is inadequately developed.

[0084] In the implementation of the offline operation mode, the information acquisition unit 131 is responsible for acquiring the data necessary for the system's operation during the preparation phase before delivery. Specifically, it instructs the delivery terminal 200 to download map data (including first road information) of the planned delivery area in advance. This download process automatically determines the necessary map range by referring to the delivery plan information stored in the delivery data storage unit 121 of the memory unit 120. Furthermore, by downloading the second road information, which has been determined in the past, in advance, comprehensive map information can be used even in an offline state.

[0085] The determination unit 132 can provide logic that enables the determination of delivery routes and second road information through local processing within the delivery terminal 200, even when offline. This local determination logic is a lightweight version of the full algorithm implemented on the server side, optimized to match the processing power of the terminal, and retains basic functions such as time-series analysis of GPS information, calculation of dwell time, and recognition of movement patterns.

[0086] The generation unit 133 can implement a map synthesis and display function that works even in an offline environment. Specifically, it can combine the first and second road information downloaded in advance with locally recorded location information to generate a composite map and visualize delivery routes within the terminal. This process employs an efficient rendering algorithm that takes into account the processing load on the terminal, minimizing battery consumption while providing sufficient visual information.

[0087] As another variation, this system can implement advanced visualization functions that go beyond 2D map representation, including 3D map display and AR (augmented reality) technology. These functions are particularly useful for spatial awareness in delivery operations in complex terrain and within high-rise buildings.

[0088] The information acquisition unit 131 is responsible for acquiring the extended dataset necessary for implementing 3D maps and AR functions. Specifically, in addition to standard location information (latitude and longitude), it can acquire altitude data (GPS altitude data, barometric pressure sensor data, etc.), terminal attitude information (accelerometer, gyroscope, magnetometer data), and camera image data. By utilizing this data, it becomes possible to accurately represent the first and second road information in three-dimensional space.

[0089] The determination unit 132 can analyze the movement of the delivery person in three-dimensional space and perform processing to determine the three-dimensional delivery route and second road information. By using three-dimensional position data including elevation, it is possible to calculate the actual travel distance and difficulty, taking into account not only the planar distance but also the elevation difference. In particular, in mountainous areas and regions with large elevation differences, this three-dimensional analysis makes it possible to evaluate the actual difficulty of delivery more accurately.

[0090] The generation unit 133 can be equipped with the function of combining three-dimensional first road information and second road information to generate a three-dimensional composite map. In this 3D composite map, attribute information such as road elevation differences, gradients, and widths is also visually represented, helping delivery personnel to select more efficient routes. Furthermore, by utilizing augmented reality (AR) technology, it is also possible to overlay delivery route information, including the second road information, onto the camera image of the delivery terminal 200, providing intuitive navigation.

[0091] As described above, by applying various implementation methods and extended functions, the present invention can provide more accurate map information while improving delivery efficiency, especially in areas with limited delivery experience.

[0092] Although embodiments of the invention have been described above, these can be implemented in various other forms, and can be carried out by various omissions, substitutions, and modifications. These embodiments and variations, as well as those with omissions, substitutions, and modifications, are included within the technical scope of the claims and their equivalents. [Explanation of Symbols]

[0093] 1 System: 100 Server terminal, 110 Communication unit, 120 Storage unit, 130 Control unit, 200 Delivery terminal, 300 User terminal, NW1 Network

Claims

1. An information acquisition unit that acquires location information and time information corresponding to the location information from the delivery terminal of the package delivery person, Based on the location information and time information, the delivery route of the delivery person is determined. An information processing system comprising: first road information registered in the map information of a map application; and a determination unit that refers to the delivery route and determines second road information of the delivery route that is not registered in the map information.

2. Furthermore, the information processing system according to claim 1, further comprising a generation unit that complements the first road information with the second road information.

3. The information processing system according to claim 1, further comprising a generation unit that generates the second road information so that it can be displayed as map information in a map application.

4. The information processing system according to claim 1, wherein the information acquisition unit acquires the map application from a third-party application server.

5. The information processing system according to claim 3, wherein the determination unit determines the delivery location of the package based on the location information and the time information, and generates the delivery location so that it can be displayed on the map information.

6. Location information and time information corresponding to the location information are obtained from the delivery terminal of the package delivery person. Based on the location information and time information, the delivery route of the delivery person is determined. A method comprising referring to first road information registered in the map information of a map application and the delivery route, and determining second road information of the delivery route that is not registered in the map information.

7. The computer is instructed to obtain location information and time information corresponding to the location information from the delivery terminal of the package delivery person. Based on the location information and time information, the delivery route of the delivery person is determined. A program that refers to first road information registered in the map information of a map application and the delivery route, and determines second road information of the delivery route that is not registered in the map information.

Images