System and method for personalized ground transportation handling and user intent prediction

Abstract

The present disclosure provides methods and systems for predicting a trip intent or destination while a user travels along a route. The method can include: (a) receiving a starting geographic location of a route and data about an identity of a user; (b) retrieving a trained classifier based at least in part on the data about the identity of the user; (c) using the trained classifier to predict a trip intent or destination based on the starting geographic location; and (d) presenting one or more transaction options to the user on an electronic device while the user travels in a land vehicle along at least a portion of the route, wherein the one or more transaction options are identified based on the trip intent or destination predicted in (c).

Description

[0001] CROSS-REFERENCE

[0002] This application claims priority to U.S. Provisional Application No. 62 / 969,472, filed February 3, 2020, the entirety of which is incorporated by reference herein. BACKGROUND

[0003] The rapid expansion of computing power of mobile computing devices, such as smartphones, tablet computers, and other portable devices, and the number and growth of software program applications (or “apps”) for mobile devices have greatly increased individual dependence on the devices, apps, and related platforms in the area of personal productivity. For example, apps are widely used to schedule meetings, determine travel routes, select modes of transportation, and other functions.

[0004] The advent and acceptance of mobility services, such as the adoption of consumer-to-consumer car sharing and ride-hailing services (e.g., Uber and Lyft ) have encouraged the convergence of transportation and mobile applications. Next-generation mobility is about autonomous and self-driving vehicles, electric vehicles, and on-demand shared mobility and the use cases they support. Self-driving vehicles that operate without human intervention are rapidly improving. As such vehicles become autonomous or automated, commercial and transactional opportunities can be presented to passengers or users during the course of transportation. SUMMARY

[0005] It is recognized herein that there is a need for methods and systems for providing products or services for use with vehicles such as fully autonomous, driverless vehicles. Beneficially, such products or services can help users of vehicles save time and resources. Further, they can allow companies providing these products or services to more directly engage with end consumers in a personalized manner within the vehicle transportation environment.

[0006] The present disclosure provides systems and methods for generating personalized transportation-centric experiences with customized monetizable driver- and / or passenger-centric services. For example, the systems and methods of the present disclosure can be used or configured to predict an intent or destination of a trip based on limited location data. The personalized transportation-centric experiences can be provided with any mode of transportation, such as, for example, autonomous vehicles, ride-hailing services, fleet-based services, micro-mobility (e.g., fleet-based demand-responsive mobility), rail transportation, and / or land-based public transportation vehicles. Machine learning systems can be used to generate, predict, estimate, or determine an intent or destination of a trip with minimal human intervention. The provided systems and methods can allow for unmanned / driverless vehicles in a range of new use cases in industries such as hotels and hospitality, restaurants and dining, travel and entertainment, healthcare, service provision, and the like.

[0007] In an aspect, a method for predicting a trip intent or destination while a user travels along a route is provided. The user of the provided system / method can be a driver, a service driver (such as a driver for a ride-hailing service), a passenger, or any user transported by a vehicle. The method can include: (a) receiving a starting geographic location of a route and data about a user profile; (b) retrieving a trained classifier based at least in part on the data about the user profile; (c) using the trained classifier to predict a trip intent or destination based on the starting geographic location; and (d) presenting one or more transaction options to the user on an electronic device while the user is traveling in the vehicle along at least a portion of the route, wherein the one or more transaction options are identified based on the trip intent or destination predicted in (c).

[0008] In some embodiments, the starting geographic location is received in the form of global positioning system (GPS) data. In some embodiments, the starting geographic location is determined in part using a geographic location of the electronic device, which is determined by a global positioning system or signal triangulation. In some embodiments, the starting geographic location is input by the user via a graphical user interface (GUI) on the electronic device.

[0009] In some embodiments, the unclassified geographic space data includes irrelevant GPS data. In some cases, the one or more training data sets include labeled data obtained using cluster analysis of a plurality of trip data records. In some cases, the method further includes generating the plurality of trip data records by associating the unclassified or irrelevant GPS data with one or more personal identities. In some cases, the plurality of trip data records are augmented by social data, traffic data, or purchase data of the corresponding personal identities.

[0010] In some embodiments, training the classifier includes creating labels for segments of a trip based on one or more labeling rules. In some embodiments, the method further includes predicting a mode of transportation for one or more portions of a travel route. In some cases, the mode of transportation includes an autonomous vehicle, a ride-hailing service, rail transit, and / or land public transit vehicles. In some embodiments, the method further includes updating the trip intent or destination when new location data is received during the trip.

[0011] Another aspect of the present disclosure provides a non-transitory computer- readable medium comprising machine executable code that, when executed by one or more computer processors, implements any of the methods above or elsewhere herein.

[0012] Yet another aspect of the present disclosure provides a system for predicting a trip intent or destination of a user. The system includes one or more computer processors and computer memory coupled with the same. The computer memory contains machine executable code that, when executed by the one or more computer processors, implements any of the methods above or elsewhere herein. In some embodiments, the one or more processors are configured to execute a set of instructions to: (a) receive a starting geolocation of a travel route and data regarding an identity of the user; (b) train a classifier based at least in part on (i) the data regarding the identity of the user and (ii) one or more training data sets containing unclassified geospatial data; (c) use the classifier trained in (b) to predict a trip intent or destination based at least in part on the starting geolocation; and (d) present one or more transaction options to the user on an electronic device when the user is traveling in a vehicle along at least a portion of the travel route, wherein the one or more transaction options are identified based at least in part on the trip intent or destination predicted in (c).

[0013] In some embodiments, the starting geolocation is received in the form of global positioning system (GPS) data. In some embodiments, the starting geolocation is determined in part using a geolocation of the electronic device, which is determined by a global positioning system or signal triangulation. In some embodiments, the starting geolocation is input by the user via a graphical user interface (GUI) on the electronic device.

[0014] In some embodiments, the unclassified geospatial data includes irrelevant GPS data. In some cases, the one or more training data sets include labeled data obtained using cluster analysis of a plurality of trip data records. In some cases, the one or more processors are further configured to generate the plurality of trip data records by associating unclassified or irrelevant GPS data with one or more personal identities. In some cases, the plurality of trip data records are augmented by social data, traffic data, or purchase data of the corresponding personal identities. In some embodiments, the one or more processors are configured to train the classifier by creating labels for segments of a trip based on one or more labeling rules. In some embodiments, the one or more processors are further configured to predict a mode of transportation for one or more portions of the travel route. In some cases, the mode of transportation includes an autonomous vehicle, a ride-hailing service, rail transit, and / or land public transit vehicles. In some embodiments, the one or more processors are further configured to update the trip intent or destination when new location data is received during the trip.

[0015] Other aspects and advantages of the present disclosure will become apparent to those skilled in the art from the following detailed description, in which only illustrative embodiments of the present disclosure are shown and described. As will be realized, the present disclosure is capable of other and different embodiments, and its several details are capable of modifications in various apparent respects, all without departing from the disclosure. Accordingly, the drawings and description are to be regarded as illustrative in nature, and not as restrictive.

[0016] INCORPORATION BY REFERENCE

[0017] All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference. To the extent there is a contradiction between the disclosure herein and that of any such incorporated publication, patent or patent application, the present specification shall control. BRIEF DESCRIPTION OF DRAWINGS

[0018] The novel features of the application are set forth with particularity in the appended claims. A better understanding of the features and advantages of the present application will be obtained by reference to the following detailed description that sets forth illustrative embodiments, in which the principles of the application are utilized, and the accompanying drawings (also “figure” and “figures” herein), of which:

[0019] FIG. 1 An example of a network environment in which a personal traffic management and intent prediction system can operate is schematically illustrated in accordance with some embodiments.

[0020] FIG. 2 A trip identification engine configured to process location data is schematically illustrated in accordance with some embodiments.

[0021] FIG. 3 A point-of-interest (POI) assignment engine of a system is shown in accordance with some embodiments.

[0022] FIG. 4 A device identifier (ID) to personal identity (ID) converter is schematically illustrated in accordance with some embodiments.

[0023] FIG. 5 A personal data augmentation engine configured to augment personal data records in a personal database is schematically illustrated in accordance with some embodiments.

[0024] FIG. 6 An automatic trip labeling engine configured to further augment data stored in an augmented personal database with intent or destination related labels is schematically illustrated in accordance with some embodiments.

[0025] FIG. 7 An itinerary selector configured to group multiple itineraries by trip intent is schematically illustrated in accordance with some embodiments.

[0026] FIG. 8 An example of an itinerary cluster is shown in accordance with some embodiments.

[0027] FIG. 9 A tag rule knowledge base updated with tags created for segments of an itinerary is illustrated in accordance with some embodiments.

[0028] FIG. 10 An automatic itinerary tagging engine configured to tag itineraries with tags obtained using cluster analysis is schematically illustrated.

[0029] FIG. 11 An example of an augmented personal data record with different additional tags is shown.

[0030] FIG. 12 An itinerary abstractor of a system is schematically illustrated in accordance with some embodiments.

[0031] FIG. 13 Multiple itinerary classifiers stored in an itinerary classifier knowledge base are schematically illustrated.

[0032] FIG. 14 A personal segmentation engine configured to segment users is schematically illustrated.

[0033] FIG. 15 An example of a user segmentation tag is shown.

[0034] FIG. 16 An example of an augmented personal data record with all tags and organized according to user segmentation is shown.

[0035] FIG. 17 A new trip intent predictor of a system is schematically illustrated in accordance with some embodiments.

[0036] FIG. 18 An example process of continuously generating and updating predictions of itinerary destinations or intents as new data is collected during an itinerary is shown.

[0037] FIG. 19 An example process of predicting a user's itinerary intent is shown.

[0038] FIG. 20 A computer system programmed or otherwise configured to implement a ground travel analysis and destination / intent prediction system is shown.

[0039] FIG. 21An example of an augmented itinerary dataset processed by the systems and methods described herein is shown.

[0040] FIG. 22 An example of insight data extracted from an itinerary associated with a user is shown.

[0041] FIG. 23 An example of a system including an insight generator and a recommendation engine is shown.

[0042] FIG. 24 An insight generator is shown schematically.

[0043] FIG. 25 An example of a transportation ontology is shown.

[0044] FIG. 26 and FIG. 27 An example of insight data about an individual is shown.

[0045] FIG. 28 An example of insight data about frequent business travelers from a particular geographic location is shown.

[0046] FIG. 29 An example of inferences or insights about a group or user is shown. DETAILED DESCRIPTION

[0047] While various embodiments have been shown and described herein, it will be apparent to those skilled in the art that many changes, modifications, and substitutions can be made thereto without departing from the disclosure. It is to be understood that various alternatives to the embodiments described herein can be employed.

[0048] As used herein, the terms “autonomous control,” “autonomous driving,” “autonomous,” and “driverless” when used to describe a vehicle generally refer to a vehicle that is capable of performing all driving tasks and monitoring the driving environment along at least a portion of a route on its own. An autonomous vehicle can travel from one point to another without requiring any intervention by a human on the autonomous vehicle. In some cases, an autonomous vehicle can refer to a vehicle having capabilities specified in the National Highway Traffic Safety Administration (NHTSA) definition of vehicle automation, and in particular Level 4 of the NHTSA definition, “In certain situations, an automated driving system (ADS) on a vehicle can itself perform all driving tasks and monitor the driving environment, essentially completing all driving. In these situations, a human need not be attentive,” or Level 5 of the NHTSA definition, “In all situations, an automated driving system (ADS) on a vehicle can complete all driving. Human occupants are simply passengers, and no longer need to be involved in driving.” In some cases, an autonomous vehicle can refer to a vehicle having capabilities specified in Level 2 (“In some situations, an advanced driver assistance system (ADAS) on a vehicle can itself actually control both steering and braking / acceleration at the same time. A human driver must remain fully attentive (“monitor the driving environment”) and perform the rest of the driving tasks”) or Level 3 (“An automated driving system (ADS) on a vehicle can itself perform all aspects of the driving task in some situations. In these situations, a human driver must be ready to take back control at the request of the ADS. In all other situations, the human driver performs the driving tasks”) of the NHTSA definition. An autonomous vehicle can also include a vehicle having a Level 2+ autonomous driving capability, where AI is used to improve a Level 2 ADAS while still requiring continuous driver control.

[0049] As used herein, the term “passenger vehicle” generally refers to a vehicle for passengers, such as a car or truck, but does not include public transit vehicles.

[0050] As used herein, the term “public transit vehicle” generally refers to a multi-passenger vehicle capable of transporting one or more groups of passengers, such as a train or bus.

[0051] As used herein, the term "trip" generally refers to the total time and / or route taken from a first location to a second location. A trip can include one or more routes. The term "route" generally refers to a set of one or more directions that allow a user to travel from a first location to a second location. A route can have one or more legs. A leg can refer to a portion of a route between a pick-up point and a drop-off point.

[0052] As used herein, the term "contextual information" generally refers to any information associated with a geographic location and / or an event. Contextual information can be derived from information that indicates or is related to such geographic locations and / or events.

[0053] As used herein, the terms "labeled data" or "labeled data set" generally refer to a paired data set used to train a model using supervised learning. The methods provided herein can utilize the intent or destination extracted by the clustered analysis ground trip analysis and destination / intent prediction system as part of the labeled data set. Alternatively, the methods provided herein can utilize a non-paired training means that allows a machine learning method to train and apply to an existing data set available to an existing system.

[0054] Whenever the term "at least," "greater than," or "greater than or equal to" precedes the first numerical value in a series of two or more numerical values, the term "at least," "greater than," or "greater than or equal to" applies to each of the numerical values in that series of numerical values. For example, greater than or equal to 1, 2, or 3 is equivalent to greater than or equal to 1, greater than or equal to 2, or greater than or equal to 3.

[0055] Whenever the term "no greater than," "less than," or "less than or equal to" precedes the first numerical value in a series of two or more numerical values, the term "no greater than," "less than," or "less than or equal to" applies to each of the numerical values in that series of numerical values. For example, less than or equal to 3, 2, or 1 is equivalent to less than or equal to 3, less than or equal to 2, or less than or equal to 1.

[0056] As used herein, the terms "a," "an," and "the" generally refer to both the singular number of and the plural number of a noun, unless the context clearly indicates otherwise.

[0057] The present disclosure provides systems and methods that can be used or configured to perform ground travel analytics and predict a trip destination / intention of one or more users. The systems and methods of the present disclosure can be used or configured to predict an intention and / or destination of a user based at least in part on location data, such as global positioning system (GPS) data. In some cases, the predicted destination can be used to further generate a personalized transportation plan for the user, e.g., to instruct the user (e.g., a driver) to take a particular route to avoid an accident, to process, recommend to, and / or present to the user personalized travel data, routing data, scheduling data, traffic data, and other forms or types of data. In some cases, machine learning techniques can be utilized to predict the intention or destination of a trip. In some cases, machine learning techniques can also be utilized to create a personalized transportation plan that includes the predicted intention / destination, a travel schedule (e.g., a start time, an end time), transaction-based purchase options for goods, services, and content during the transportation, a vehicle type (e.g., an autonomous vehicle type such as a sedan or a minivan, a brand), a transportation mode type (e.g., an autonomous vehicle, public transportation such as a train, a light rail, or a city bus, a shuttle, a ride share, a ride hail, a shared or private trip, walking, a bicycle, an e-scooter, a taxi, etc.), and others.

[0058] The intention / destination prediction capability can be utilized or implemented in a driver and / or passenger monetization platform. For example, the driver and / or passenger monetization can include activities and services related to: a) transaction-based purchases of goods such as gasoline, food, coffee, services, parking, and content (e.g., a podcast about an artist’s work on exhibit at a museum at the destination of the driver or passenger) during the transportation; b) subscriptions to access content, e.g., an annual subscription to a music streaming service, a news service, a concierge service, etc.; c) transaction-based purchases of goods, services, and content while in transit and while the vehicle is intermittently parked, such as at a gas station, a restaurant, a coffee shop, etc. (e.g., a refueling station operator such as an energy company can partner with a coffee chain to offer a discount to passengers who purchase a coffee drink while refueling a vehicle); and d) redemption of loyalty points, e.g., car manufacturers and ride service fleet operators can use a system similar to that used by airlines or chain hotels to reward the loyalty of their customers, where the redemption of loyalty points can be in substantially the same manner as these and other industries use such programs.

[0059] Artificial intelligence, including machine learning algorithms, can be used to train a prediction model for predicting trip intent or destination. For example, the machine learning algorithm can be a neural network. Examples of neural networks that can be used with the implementations herein can include deep neural networks, convolutional neural networks (CNNs), and recurrent neural networks (RNNs). In some cases, the model trained by the machine learning algorithm can be pre-trained and implemented on a user device or on-ground trip analysis and destination / intent prediction system, and the pre-trained model can undergo continuous retraining - which can involve continuously tuning the prediction model or components of the prediction model (e.g., classifiers) to adapt to changes over time in the implementation environment (e.g., changes in customer / user data, sensor data, model performance, third-party data, etc.).

[0060] METHOD AND SYSTEM FOR GROUND TRAVEL ANALYSIS AND DESTINATION INTENT PREDICTION

[0061] The present disclosure provides systems and methods that can be used or configured to perform on-ground trip analysis and predict trip destinations / intents for one or more users. The systems and methods of the present disclosure can be used or configured to predict a user’s intent and / or destination based at least in part on location data. In some implementations, the location data can include global positioning system (GPS) data. The GPS data can be unclassified data or can be unassociated with a user (e.g., de-identified data). The proposed systems and methods can process unclassified location data and identify a trip intent or destination associated with a user. The trips, and corresponding users and intents, can be used to generate a training dataset for training a classifier in order to determine / infer trip intents or destinations at deployment. The trip intent and / or destination can be predicted during a trip based on real-time location data or limited location data (e.g., location of trip start location). Alternatively or additionally, the intent and / or destination can be predicted for a new / next trip before the trip begins.

[0062] ​A user can pre-register or subscribe to one or more travel services provided by the system herein in the system herein. The user can be a potential requester of a travel service. The user can utilize a user mobile application to receive transaction options provided by the system during a trip. The application can provide one or more transaction services or monetizable driver and / or passenger centric service options to the user based on a predicted intent or destination. Service or transaction offers related to the predicted intent or destination can be presented to the user via the application. The user can acquire a service or conduct a transaction via the application during a trip. The user can be transported from a first location to a second location using and / or while acquiring one or more services including travel services and user experience services provided by the system during a trip. A user of the provided system / method can be a driver, a service driver such as a driver of a ride-hailing service, a passenger, or any user transported by a vehicle as described above.

[0063] FIG. 1 An example of a network environment 100 in which the ground travel analytics and destination / intent prediction system 101 can operate is schematically illustrated. The ground travel analytics and destination / intent prediction system 101 can interact with a plurality of user devices 103 over one or more networks 110. The ground travel analytics and destination / intent prediction system 101 can be coupled to or be part of a personal transportation platform for providing personalized transportation experiences including providing personalized services / products during a trip. In some implementations, one of the plurality of user devices 103 can be a device associated with a user. In some implementations, one user device can be used by multiple users. For example, a user device can be a built-in device or system inside or coupled to a vehicle. In some implementations, two or more user devices can be associated with a single user.

[0064] In some embodiments, the ground travel analytics and destination / intent prediction system 101 can be configured to provide a user interface for a user to view a travel route or interact with one or more related transaction options for a predicted trip intent / destination during a trip via the user device 103. In some cases, the user interface can include a GUI presented on a display on the user device or in the vehicle. The ground travel analytics and destination / intent prediction system can be configured to predict a trip intent or destination based on limited location data such as a trip start location. The trip intent or destination can be generated using a machine learning based model based on limited location data (e.g., GPS data of trip start location) and / or personal data (e.g., personal ID). The ground travel analytics and destination / intent prediction system can be configured to predict a trip intent / destination with improved prediction accuracy based on limited real-time data. The prediction of the trip intent or destination can be dynamically updated and / or improved as the trip progresses. Details regarding intent and destination prediction are described below.

[0065] In some cases, the ground travel analytics and destination / intent prediction system can be coupled to a personal transportation planning system configured to generate a personalized transportation plan including a travel route, a schedule of departure times and arrival times for one or more segments or one or more stops during the trip, a mode of transportation for segments of the travel route (e.g., type of transportation, type / brand of vehicle, vehicle configuration, etc.), and one or more services or monetizable driver and / or passenger centric services (e.g., digital services, transaction events, or business activities related to the destination) during the trip. In some cases, the personalized transportation plan can also include transporting the user using an autonomous vehicle for at least one segment. In some cases, the personalized transportation plan or at least a portion of the personalized transportation plan (e.g., mode of transportation) can be dynamically updated based on updated trip intent / destination prediction.

[0066] The personalized transportation plan can be generated based on data related to the user and / or data related to transaction services. The data related to the user can include a personal identity (ID), historical data such as user preferences, transportation history, or purchase history. Such data can be collected from a variety of data sources such as a mobile application (e.g., a map application, a navigation application, an email, a text message, a social network application, a personal health application, etc.), a social network software, a third party service provider such as a transportation service provider (e.g., Uber and Lyft ), a vendor, a business entity (e.g., a fast food restaurant, a restaurant, a coffee shop, a hotel, a convenience store, a gas station, a theater, etc.), a content provider (e.g., Apple Music digital virtual assistants, smart home devices such as Alexa , interactive voice response (IVR) systems, social media channels, and messaging APIs such as Facebook channels, Twilio SMS channels, Skype channels, and various other sources. Data related to transaction services can include user rejections or acceptances of previous transaction services or data from third party service providers. Personalized transportation plans can be generated using machine learning based models based at least in part on predicted intent or destination. Input data can be data derived from various data as described above. For example, input data can include social graphs, purchase graphs, transportation graphs, demographic information, weather data, provider or service provider catalogs, and various other data. Output of the model can be a travel route, a schedule for one or more legs of the travel route (e.g., departure time, arrival time, etc.), a mode of transportation for each leg (e.g., vehicle, car type), and one or more transaction options or services during the trip. In some cases, the system can provide transaction offers to the user in real time. For example, upon receiving user input indicating rejection of a service offer, a new transaction offer can be selected and provided to the user in real time.

[0067] As used herein, real-time generally refers to a response time of, for example, less than 1 second, one tenth of a second, one hundredth of a second, a millisecond, or less for a computer processor. Real-time can also refer to an occurrence of a first event being simultaneous or substantially simultaneous with respect to an occurrence of a second event.

[0068] The ground travel analysis and destination / intent prediction system 101 can include one or more servers 105 and one or more database systems 107, 109, which can be configured to store or retrieve relevant data. Relevant data can include processed GPS data, trip data, augmented trip data, augmented personal data records (labeled with additional data related to trip intent, trip type, user segmentation, etc.), user profile data (e.g., user preferences, personal data such as identity, age, gender, contact information, demographic data, ratings, etc.), historical data (e.g., social graphs, transportation history, transportation subscription plan data, purchase or transaction history, loyalty programs, and various other data as described elsewhere herein. In some instances, the ground travel analysis and destination / intent prediction system 101 can obtain data (e.g., via one or more networks 110) or otherwise communicate with one or more external systems or data sources, such as one or more location data services, ontological knowledge bases, maps, weather or traffic application program interfaces (APIs), or map databases. In some instances, the ground travel analysis and destination / intent prediction system 101 can retrieve data from the database systems 107, 109, which communicate with one or more external systems (e.g., location data sources, transportation service providers, autonomous vehicle dispatch systems, third-party monetizable driver- and / or passenger-centric service entities, such as fast food restaurants, restaurants, coffee shops, hotels, convenience stores, gas stations, theaters, digital service providers, etc.). In some instances, the databases can be synchronous databases that maintain tables or records of information such as weather, traffic, public transportation, global positioning system (GPS) inputs or logs, schedule data, personal data, and other data obtained from external data sources.

[0069] Each component (e.g., servers, database systems, user devices, external systems, etc.) can be operatively interconnected via one or more networks 110 or any type of communication links allowing transfer of data from one component to another. For example, respective hardware components can include network adapters that allow one-way and / or two-way communication with one or more networks. For example, servers and database systems can communicate with user devices 103 and / or data sources via one or more networks 110 to send and / or receive relevant data.

[0070] A server (e.g., server 105) can include a web server, a mobile application server, an enterprise server, or any other type of computer server, and can be programmed with computers to accept requests (e.g., HTTP or other protocols that can initiate data transmissions) from computing devices (e.g., user devices, other servers), and provide the requested data to the computing devices. The server can be a single server or a distributed server across multiple computers or multiple data centers. The server can be various types, such as, but not limited to, a web server, a news server, a mail server, a messaging server, an advertising server, a file server, an application server, a switch server, a database server, a proxy server, another server suitable for performing the functions or processes described herein, or any combination thereof. In addition, the server can be a broadcast facility for distributing data, such as free-to-air, cable, satellite, and other broadcast facilities. The server can also be a server in a data network (e.g., a cloud computing network).

[0071] The server can include various computing components, such as one or more processors, one or more memory devices that store software instructions and data for execution by the processors. The server can have one or more processors and at least one memory for storing program instructions. The processor can be a single or multiple microprocessors, field-programmable gate arrays (FPGAs), or digital signal processors (DSPs) capable of executing a particular instruction set. The computer-readable instructions can be stored on a tangible, non-transitory computer-readable medium, such as a floppy disk, a hard disk, a CD-ROM (compact disk read-only memory), and MO (magnetic optical), DVD-ROM (digital versatile disk read-only memory), DVD RAM (digital versatile disk random access memory), or semiconductor memory. Alternatively, the method can be implemented in hardware components or a combination of hardware and software (e.g., an ASIC, a special purpose computer, or a general purpose computer).

[0072] The one or more databases 107, 109 can utilize any suitable database technology. For example, a structured query language (SQL) or "NoSQL" database can be utilized to store processed / raw GPS data, user profile data, historical data, prediction models or algorithms for predicting trip intent / destination, maps or other data. Some databases can be implemented using various standard data structures, such as arrays, hashes, (linked) lists, structs, structured text files (e.g., XML), tables, JavaScript Object Notation (JSON), NOSQL, etc. Such data structures can be stored in memory and / or (structured) files. In another alternative, an object-oriented database can be used. Object databases can include multiple collections of objects that are grouped and / or linked together by common attributes; they can be related to other collections of objects by some common attributes. Object-oriented databases are similar in execution to relational databases, except that objects are not just pieces of data, but can have other types of functionality encapsulated within a given object. In some embodiments, the databases can include a graph database that uses a graph structure with nodes, edges, and properties for semantic queries to represent and store data. If the databases of the present application are implemented as data structures, the use of the databases of the present application can be integrated into another component, such as a component of the present application. Further, the databases can be implemented as a mix of data structures, objects, and relational structures. The databases can be merged and / or distributed in varying forms by standard data processing techniques. Portions of the databases can be exported and / or imported and thereby decentralized and / or integrated, e.g., tables.

[0073] In some embodiments, the ground travel analysis and destination / intent prediction system 101 can construct databases in order to efficiently deliver data to users. For example, the ground travel analysis and destination / intent prediction system 101 can provide customized algorithms to extract, transform, and load (ETL) data. In some embodiments, the ground travel analysis and destination / intent prediction system 101 can use a proprietary database architecture or data structure to construct databases to provide an efficient database model that is adapted to large-scale databases, easily scalable, efficient querying and data retrieval, or reduced memory requirements compared to using other data structures.

[0074] One or more databases (e.g., augmented personal databases) can be accessed by a variety of applications or entities that can be related to a transaction, although in some cases such a variety of applications or entities can be unrelated to a transaction. In some cases, data stored in an augmented personal database can be utilized or accessed by other applications through an application programming interface (API). Data accessed by a variety of applications can include predicted intent / destinations, predicted traffic patterns, and / or data managed by the system, such as personal data records. Access to databases can be authorized at each API level, each data level (e.g., data type), each application level, or according to other authorization policies.

[0075] The ground travel analysis and destination / intent prediction system 101 can be implemented anywhere in the network. The ground travel analysis and destination / intent prediction system 101 can be implemented on one or more servers in the network, in one or more databases in the network, on one or more electronic devices built-in or coupled to a vehicle, or on one or more user devices. For example, the ground travel analysis and destination / intent prediction system 101 can be implemented in a distributed architecture (e.g., multiple devices collectively performing to implement or otherwise execute the ground travel analysis and destination / intent prediction system 101 or operations thereof) or in a replicated manner (e.g., multiple devices each implementing or otherwise executing the ground travel analysis and destination / intent prediction system 101 or operations thereof as independent systems). The ground travel analysis and destination / intent prediction system 101 can be implemented using software, hardware, or a combination of software and hardware in one or more of the above-described components within the network environment 100.

[0076] A user device of the plurality of user devices 103 can be an electronic device. The user device can be a computing device configured to perform one or more operations consistent with the disclosed implementations. Examples of a user device can include, but are not limited to, a mobile device, a smart phone / cell phone, a tablet computer, a personal digital assistant (PDA), a smart wearable device, a smart watch, a laptop or notebook computer, a desktop computer, a media content player, a television, a video game station / system, a virtual reality system, an augmented reality system, a microphone, or any electronic device configured to enable a user to view a travel route, and interact with transaction or service related information, and display other travel related information, for example. The user device can be a handheld object. The user device can be portable. The user device can be carried by a human user. In some cases, the user device can be located away from the human user, and the user can control the user device using wireless and / or wired communication. The user device can be a computing device in communication with a wearable device worn by the user. In some cases, the wearable device can be configured to monitor user activity, vital signs (e.g., blood pressure and heart rate), or health conditions of the user. In some cases, the user device can be an electronic device coupled to or located on a vehicle.

[0077] In some implementations, the user device can be capable of detecting a location of the device / user. The user device can have one or more sensors on the device to provide instantaneous positioning or location information of the user device. In some implementations, the instantaneous location information can be provided by sensors such as location sensors (e.g., global positioning system (GPS)), inertial sensors (e.g., accelerometer, gyroscope, inertial measurement unit (IMU)), altitude sensors, attitude sensors (e.g., compass), pressure sensors (e.g., barometer), field sensors (e.g., magnetometer, electromagnetic sensor), and / or other sensor information (e.g., WiFi data). The location of the user device can be used to locate a starting point of a travel route. As an additional or alternative, the location of a place of interest (e.g., a trip starting point, a stop during a trip) can be provided by a user via the user device 103, such as by manually entering the location via a user interface.

[0078] The user device can include a communication unit that can allow for communication with one or more other components in a network. In some cases, the communication unit can include a single communication module or multiple communication modules. In some cases, the user device can be capable of interacting with one or more components in a network environment using a single communication link or multiple different types of communication links. The user device 103 can interact with the ground travel analytics and destination / intent prediction system 101 by requesting and obtaining the above-mentioned data via the network 110.

[0079] A user device can include one or more processors capable of executing instructions of a non-transitory computer-readable medium that can provide one or more operations consistent with the disclosed implementations. A user device can include one or more memory storage devices comprising a non-transitory computer-readable medium comprising code, logic or instructions for performing one or more operations.

[0080] In some implementations, a user can utilize a user device 103 to interact with the ground travel analysis and destination / intent prediction system 101 through one or more software applications (i.e., client software) running on and / or accessed by the user device, where the user device 103 and the ground travel analysis and destination / intent prediction system 101 can form a client-server relationship. For example, the user device 103 can run a dedicated mobile application provided by the ground travel analysis and destination / intent prediction system 101.

[0081] In some embodiments, the client software (i.e., the software application installed on the user device 103) can be used as a downloadable mobile application for various types of mobile devices. Alternatively, the client software can be implemented in a combination of one or more programming languages and markup languages for execution by various web browsers. For example, the client software can be executed in web browsers that support JavaScript and HTML rendering, such as in Chrome, Mozilla Firefox, Internet Explorer, Safari, and any other compatible web browsers. Various embodiments of the client software application can be compiled for various devices, across multiple platforms, and can be optimized for their respective native platforms. In some cases, third-party user interfaces or APIs can be integrated into the mobile application and in the front-end user interface (e.g., within the graphical user interface). The third-party user interfaces can be hosted by third-party servers. The third-party servers can be provided by a range of third-party entities, such as by original equipment manufacturers (OEMs), hotels and hotels, restaurants and dining, travel and entertainment, service delivery, and various other entities described elsewhere herein. In some cases, APIs or third-party resources (e.g., map service providers, travel service providers, digital service providers, Starbucks, McDonalds, Ticketmaster, etc.) can be used to provide transactions and transact with users. In some cases, one or more third-party services can be invoked by the ground travel analytics and destination / intent prediction system 101 and integrated into the user application, such that users can access such services in a familiar front-end user experience. In some cases, one or more of the above-described services can be built-in components of the ground travel analytics and destination / intent prediction system 101 and can be provided to users without outsourcing to third-party entities. In some cases, data retrieved from third-party service providers can be organized and stored by the ground travel analytics and destination / intent prediction system 101 to form a vendor / service catalog that can be used to determine transaction offers related to predicted intents to provide to users during transit. In some cases, the ground travel analytics and destination / intent prediction system 101 can provide a graphical user interface (GUI). The GUI can allow users to access, accept, reject, select one or more transaction offers / options, information, services related to predicted destinations by interacting with graphical elements and viewing information such as travel routes and travel schedules during transit.

[0082] The user device can include a display. The display can be a screen. The display can be a touchscreen. Alternatively, the display can not be a touchscreen. The display can be a light emitting diode (LED) screen, an OLED screen, a liquid crystal display (LCD) screen, a plasma screen, or any other type of screen. The display can be configured to show a user interface (UI) or graphical user interface (GUI) presented by an application (e.g., via an application programming interface (API) executed on the user device). For example, the GUI can show a graphical element that allows the user to accept or reject a transaction offer, as well as view information related to a predicted intent / destination, travel route, and transaction options.

[0083] The network 110 can be a communication pathway between the personal transportation management system 101, the user device 103, and other components of the network. The network can include any combination of local and / or wide area networks using wireless and / or wired communication systems. For example, the network 110 can include the Internet and mobile telephone networks. In one embodiment, the network 110 uses standard communications technologies and / or protocols. As such, the network 110 can include links using technologies such as Ethernet, 802.11, worldwide interoperability for microwave access (WiMAX), 2G / 3G / 4G, or long term evolution (LTE) mobile communications protocols, infrared (IR) communication, Bluetooth®, and / or any of the wireless technologies used to enable remote units, such as mobile phones and laptops, to communicate together, and can be implemented using a combination of technologies. Other networking protocols, such as MPLS, transmission control protocol / Internet protocol (TCP / IP), user datagram protocol (UDP), hypertext transfer protocol (HTTP), simple mail transfer protocol (SMTP), file transfer protocol (FTP), and / or the like, can be used on the network 110. Data exchanged over the network can use technologies and / or formats including binary-formatted image data (e.g., portable network graphics (PNG)), hypertext markup language (HTML), extensible markup language (XML), and / or the like. Moreover, all or some of links can be encrypted using conventional encryption technologies, such as secure sockets layer (SSL), transport layer security (TLS), Internet Protocol security (IPsec), and / or the like, in an embodiment, entities on the network can use custom and / or dedicated data communications technologies instead of or in addition to the foregoing. The network can be wireless, wired, or a combination thereof.

[0084] A system for predicting a trip destination or intent can include a plurality of components as described and shown herein. FIGS. 2-18Various components of the system are illustratively shown that are configured to perform functions and operations to collectively predict destinations or intents based on limited location data. In some implementations, the various components of the system can include a trip identification engine, a point of interest (POI) assignment engine, a device identifier (ID) to personal ID converter, a personal data augmentation engine, an automatic trip labeling engine, a machine learning (ML) clustering system, a personal partitioning engine, a database, and various other components.

[0085] In some implementations, one or more components of the system (e.g., a trip classifier) can be developed using unclassified or uncorrelated location data. In some implementations, personal data records generated and managed by the ground travel analysis and destination / intent prediction system can include data extracted from unclassified or uncorrelated location data. In some cases, unclassified or uncorrelated location data can be processed by a trip identification engine of the system to output a plurality of individual trips. FIG. 2 A trip identification engine 210 for processing location data according to some implementations is illustratively shown.

[0086] In some implementations, the location data 201 can include GPS data. In some cases, the GPS data can include a plurality of records or datasets containing corresponding GPS information. For example, a table can be maintained having an entry for each record containing corresponding coordinate data (e.g., latitude, longitude), a timestamp, resolution, and other information such as a device ID. The data entries of a location record can include any suitable data structure. For example, the data structure can include a plurality of data fields as described above. The data structure can depend on the original data format or the data source from which the data is retrieved.

[0087] The location data 201 can be retrieved from one or more data sources. The location data can be obtained with suitable location-based technologies such as Global Navigation Satellite System (GNSS), cellular triangulation, Assisted GPS (A-GPS), Differential Global Positioning System (DGPS), and the like.

[0088] In some implementations, the location data can include at least geographic location information and an identifier such as a device ID. Depending on the data source, the identifier can be, for example, an advertising identifier (IDFA) or an Android advertising ID (AAID).

[0089] The plurality of records of location data can be obtained from one or more data sources in a streaming or batched manner. The location data can be time series data, such as spatiotemporal point measurements. In some cases, the plurality of records or datasets of location data can be unclassified, which can not be streamed or organized in a time series. The plurality of records or datasets can be associated with a plurality of trips associated with one or more devices / persons. In some cases, the plurality of records or datasets of location data can not include explicit information about a person or a trip.

[0090] The location data can be processed by the trip identification engine 210 to identify one or more individual trips 221-1, 221-2, 221-3. In some cases, the one or more individual trips can be stored in the trip database 220. A trip can include an ordered sequence of location data records that includes at least a set of location data corresponding to a trip origin and a set of location data corresponding to a trip destination. A trip can be identified from the plurality of sets of location data using any suitable method such that trip information including a starting geographic location / starting time of the trip, a destination geographic location / ending time of the trip, and a device ID associated with the trip is identified. One or more unique trips can be identified and associated with a device ID. For example, different trips 221-1, 221-2 at different starting / ending time points can be identified and associated with the same device ID.

[0091] In some embodiments, location context data such as points of interest (POIs) can be assigned to the geographic locations of the origin and / or destination of a trip. FIG. 3 A POI assignment engine 303 of a system according to some embodiments is shown. The POI assignment engine 303 can be coupled to the trip database 301 for assigning location context data to at least the origin and destination of individual trips. In some cases, location context data can be assigned to one or more intermediate location points (e.g., intermediate stops) in a trip. Such location context data can be associated with the identity of a destination, origin, or stop along a trip. The location context data may, for example, include a coffee shop, airport, venue, restaurant, stadium, theater, dental office, etc. The trip database 301 can be the same as the trip database described in the FIG. 2

[0092] ​In some cases, the POI assignment engine can retrieve one or more itineraries from the itinerary database 301 and assign location context data collected using crowd-sourcing methods. For example, the POI assignment engine can be coupled to one or more third-party data sources 307-1, 307-2, 307-3 for collecting local business, landmark, and point of interest (POI) data. As an example, the POI data can have geographic tags associated therewith, and the POI assignment engine can search the POI data for a good match to the itinerary's start or destination geographic coordinates (e.g., start itinerary latitude / longitude, end itinerary latitude / longitude). The POI data with the best match (e.g., start itinerary POI, end itinerary POI) can be returned.

[0093] In some cases, one or more third-party data sources 307-1, 307-2, 307-3 can be selected over other data sources to provide POI data. For example, when multiple candidate POIs are associated with the same location or address, the system can select one POI from the multiple candidate POIs based on selection criteria. In some cases, the selection criteria can be user-specific. For example, a POI can be selected from multiple candidate POIs based on historical user data (e.g., the user's transportation, purchase, and / or transaction history, the user's social activities, etc.). Alternatively or additionally, the POI data can be a combination of multiple candidate POIs. For example, the POI data can be aggregated information about all POIs belonging to the same geographic location point (e.g., street name, landmark reference, area name, etc.). In other cases, a POI with an unstructured address that does not follow any format (such as "apartment number, street name, zone number") can be supplemented by information of other POIs belonging to the same geographic location.

[0094] In some cases, the location context data can be augmented by various other data sources such as an ontology knowledge base 305. For example, some POIs can be supplemented with additional attributes provided by the ontology knowledge base (e.g., an airport has a terminal number, airlines within the terminal). The ontology knowledge base 305 can be developed manually by one or more individuals, agencies, imported from external systems or resources, or can be learned in part using a machine learning system that collaborates with the user (e.g., extracting transportation terminology from natural language text). The itinerary data set 310 supplemented with POI data 311, 313 can in turn be ingested back into the itinerary database 301.

[0095] As described above, native / raw location data can not be associated with an individual. In some implementations, a device ID to individual ID converter of the system can be used to associate location data with an individual based on a device ID. FIG. 4 A device ID to individual ID converter 401 is schematically illustrated in accordance with some implementations.

[0096] Device ID to Personal ID converter 401 can be configured to associate one or more device IDs with personal IDs. Device ID to Personal ID converter 401 can process device ID data and associate device IDs with personal identifiers. Personal identifier 413 can correspond to an individual's identity that can be obtained using a suitable identity resolution method. Identity resolution methods can cover different identity attributes and matching algorithms. For example, identity-related data such as personal identity attributes, social behavior attributes, and social relationship attributes can be processed using matching algorithms such as pairwise comparison, transitive closure, and collective clustering to extract an individual's identity. Identity-related data and device IDs can be further processed and associated to form a personal data record 411. Entries 413 of the personal data record can include multiple data fields, such as personal ID and device ID. The personal data record 411 can then be stored in a personal database 410. In some cases, unidentified device IDs 403 can be stored in a separate database.

[0097] In some implementations, personal data records can be expanded with additional personal data. FIG. 5 A personal data augmentation engine 500 is schematically illustrated, configured to augment personal data records to form an augmented personal database 420. In some embodiments, the personal data augmentation engine can retrieve user data from multiple data sources (e.g., demographic data, purchase data, social graphs, etc.) and travel data from a travel database 301 to augment the personal data records.

[0098] User data can include personal data related to an individual, such as identity, age, gender, contact information, demographic data, and the like. Such data can be extracted from other data sources or third-party applications. In some cases, personal data can also include user preferences. User preferences can include travel preferences and transaction / service preferences. Travel preferences can be derived from one or more of various parameters obtained by the system and used to generate personalized travel routes or to predict traffic patterns. For example, a travel preference such as a "fastest route" preference indicates a preference for the fastest (in time) route between two points. A "shortest route" preference can indicate a preference for the shortest (in distance) route between two points. A "most fuel-efficient route" preference can indicate a preference for fuel savings. A travel preference can indicate a preference for "effort," which can be particularly relevant to cyclists, pedestrians, runners, hikers, and swimmers, who can, for example, want a large variation in grade (e.g., hills) or a small variation in grade (e.g., flat). A travel preference can indicate a preference for routes with various points of interest, more vegetation than urban landscapes, a preference for museums, theaters, playhouses, and the like, a preference for routes that include shopping opportunities, a preference for food, a user preference to avoid being stuck in traffic (even if a route with heavy traffic is the fastest path to a destination), and various other preferences. Travel preferences can include a user preference for a mode of transportation (e.g., autonomous vehicle, public transportation such as train, light rail, or city bus, shuttle, carpool, ride-hail, shared or private ride, walking, cycling, e-scooter, taxi, and the like) or in-vehicle user experience (e.g., access to music, games), and the like. Travel preferences can be used to determine a travel route, a segment of a route, a mode of transportation for a segment, and / or a stop (e.g., a point of interest, a restaurant, a coffee shop, and the like) during a travel route. Such user preferences can be input by a user and / or extracted from other data sources or historical data.

[0099] Purchase data can include any purchase or transaction history made by a user during a trip or at the end of a trip. The purchase or transaction can be made at any location that can not be in the vehicle (e.g., at a destination). The purchase or transaction can be an in-vehicle or in-cabin transaction.

[0100] In some cases, the social data can depict relationships between individual users or vehicles to facilitate carpooling, etc. In some cases, the social data can indicate relationships between users and other individuals and entities (e.g., families, businesses, friends, etc.), road networks, and potential meeting locations within a community. In some cases, the social data can be used to facilitate carpooling based on common interests and travel activities, provide recommended vehicles and locations, suggest carpooling partners. In some cases, the social data can be used to predict or recommend locations and / or schedules for trips. For example, if a user is scheduled to meet someone with whom the user has a business relationship, the arrival time can be scheduled based on business meeting preferences. The social data can be obtained from social networks (e.g., Facebook, Twitter, Linkedln, etc.), historical communications (e.g., emails, SMS, video chats, etc.), common memberships of clubs, common memberships of institutions, common memberships and affiliations, family relationships, common employers, common work places, etc.

[0101] In some cases, the additional user data can be used to predict a mode of transportation for at least one leg of a trip. For example, the user data can include a mode of transportation (e.g., autonomous vehicle, public transportation (such as train, light rail, or city bus), shuttle, carpool, ride-share, shared trip or private trip, walking, biking, e-scooter, taxi, etc.) collected from historical transportation data associated with the user.

[0102] The personal data augmentation engine 500 can be configured to augment personal data records in the personal database 410 with user data (e.g., demographic data, purchase data, social graph, etc.) and trip data retrieved from the trip database 301, thereby generating augmented personal data records. In some implementations, the trip data and the personal data can be merged based on device IDs. In some implementations, the personal data augmentation engine 500 can employ suitable techniques to merge different databases (i.e., trip database, personal database). For example, the trip database can be merged with multiple records in the personal database by incorporating the trip data into one of the multiple personal data records based on matching device IDs. The merged personal data records can then be saved in the augmented personal database 420 as augmented personal data records. For example, the augmented personal data records can include data fields such as a personal ID, a device ID, and corresponding trip data (e.g., a series of location data tagged with POIs).

[0103] In some implementations, the augmented personal data records can be further updated and augmented with additional intent or destination related information. FIG. 6An automated trip tagging engine 600 is illustratively shown configured to further augment data stored in the augmented personal database 420 with intent or destination related tags.

[0104] In some embodiments, the intent of a destination can be context information related to an activity, or an intent inferred from the destination. For example, the intent of a supermarket can be food shopping, the intent associated with a weekday morning trip can be a daily commute, the intent associated with a food store, a gas station, etc. can be shopping, and the intent associated with a sports field, a cinema can be entertainment. Such intents can be obtained from a tagging rules knowledge base 610 that stores a list of intent tags. The automated trip tagging engine 600 can tag the destination of each trip with the intent tags provided by the tagging rules knowledge base 610. The augmented personal data records can then be updated by incorporating the intent of the destination associated with each trip.

[0105] In some embodiments, the augmented personal data records can be processed by a trip selector of the system to group trips by intent. The trip groups can be further processed by a machine learning based clustering system to further cluster by destination. FIG. 7 A trip selector 700 is illustratively shown configured to group a plurality of trips retrieved from the augmented personal database 420 by intent. For example, trips with the same intent tag can be grouped together. Trips with the same intent (e.g., to communicate to work) can or can not have the same destination.

[0106] In some cases, the trips grouped by intent are further processed by a machine learning based clustering system. The machine learning based clustering system can perform a clustering analysis on the trips to determine a natural grouping of trips for the group intent. For example, trips of the same cluster can have the same destination, and can have different origins, intermediate stops, or start / end times. In some cases, a group of trips belonging to the same cluster can have one or more common characteristics (e.g., origin, mode of transportation, intermediate stops) in addition to the intent. In some cases, one or more common characteristics can be identified and used as a tag for the trip or trip segments.

[0107] In some cases, the tags associated with the clusters can be used as part of a training data set to train a classifier. The tags can be used as part of a training data set to pair with trip data. The tags can be created manually by one or more individuals, institutions, or imported from external systems or resources. The tags can be identified based on the natural clusters generated by the machine learning based clustering system as described above (e.g., private car, morning trip, direct commute to work, etc.).

[0108] FIG. 8An example of a trip cluster is shown. This set of trips can be contained in a "commute to work using private car" cluster. This set of trips can have the same destination and the same intent, while the intermediate stops can be different.

[0109] In some embodiments, trip segments can be tagged with augmenting data. FIG. 9 An example of updating the tagging rules knowledge base with tags created for trip segments is shown. Tags can be created manually by one or more individuals, institutions, or imported from external systems or resources. Tags can be identified based on natural clusters generated by the machine learning based clustering system as described above (e.g., private car, morning trip, direct commute to work, etc.).

[0110] In some embodiments, tags created for trips or trip segments can be used to augment the personal data records. FIG. 10 An automatic trip tagging engine 1000 is schematically illustrated, which is configured to tag trips with tags obtained using clustering analysis as described above. The automatic trip tagging engine 1000 can be the same as the automatic trip tagging engine described in FIG. 6 The automatic trip tagging engine 1000 can automatically tag trips or one or more segments of a trip. In some cases, the tagging can be performed based on tagging rules retrieved from the tagging rules knowledge base 610. The tagging rules can be developed manually by one or more individuals, institutions, imported from external systems or resources, or can be learned in part using machine learning systems in collaboration with users (e.g., extracting transportation terminology from natural language text). The following are example rules stored in the tagging rules knowledge base 610:

[0111] Rule 1 : If trip start time is after 5:00 AM and trip end time is before 12:00 PM, then tag = "morning trip" (here we compute local time zone AM, PM)

[0112] Rule 2: If trip type = "daily commute" and intermediate stop duration is > 5 minutes and < 20 minutes, and POI type in trip = "coffee shop", then tag = "commute to work coffee shop"

[0113] Trips augmented with additional tags can be stored in the augmented personal database. As FIG. 10 shown in FIG. 6 the augmented personal data records include additional tags (e.g., morning, direct commute to work, from home, to office, private car, etc.) compared to the augmented personal data records shown in

[0114] FIG. 11Examples of augmented personal data records with different additional labels are shown. The two augmented personal data records as shown in the examples have the same intent, multiple common features (e.g., morning, from home, to office, private car), but one different feature (e.g., direct commute to work, commute to work coffee).

[0115] In some implementations, the plurality of augmented personal data records can be used to train a machine learning based trip abstractor. FIG. 12 A trip abstractor 1200 of a system according to some implementations is schematically illustrated. The trip abstractor can be a classifier trained using augmented personal data records including additional labels. The trip abstractor can process trip data and predict an intent associated with a trip destination. The trip abstractor can be trained using artificial intelligence such as a machine learning algorithm. For example, the machine learning algorithm can be a neural network. Examples of neural networks include deep neural networks, convolutional neural networks (CNNs), and recurrent neural networks (RNNs). In some cases, a model trained by a machine learning algorithm can be pre-trained and implemented on a system, and the pre-trained model can undergo continuous retraining - which involves continuously tuning a prediction model or components of a prediction model (e.g., a classifier) to adapt to changes over time in the implementation environment (e.g., changes in customer / user data, available labeled data, model performance, third party data, etc.).

[0116] In some implementations, the trip classifier can correspond to an intent of a trip. FIG. 13 A plurality of trip classifiers stored in a trip classifier knowledge base 1300 is schematically illustrated. In some cases, training a model can involve selecting a model type (e.g., a CNN, an RNN, a gradient boosting classifier or regressor, etc.), selecting an architecture of the model (e.g., number of layers, nodes, ReLU layers, etc.), setting parameters, creating training data (e.g., pairing data, generating input data vectors), and processing the training data to create the model. The trained model can be tested and optimized using test data retrieved from the trip classifier knowledge base. Test results can be compared to performance characteristics to determine whether the prediction model meets performance requirements. If the performance is good, i.e., meets the performance requirements, the model can be inserted into the trip classifier knowledge base.

[0117] FIG. 21An example of an augmented trip dataset 2100 processed by the system and method described above is shown. In the example shown, the augmented trip dataset 2100 can be generated using at least the irrelevant GPS data. One or more augmented trips can be associated with a device, such as a device ID 2101. Each augmented trip can include, for example, a device ID 2101, predicted or extracted information about a trip origin, such as a POI / name of the origin 2103 (e.g., one stop foods, Galter Life Center), a category of the POI 2105 (e.g., grocery store, office building, gym, restaurant), an address of the origin 2107, predicted or extracted information about a destination of the trip, such as a POI of the destination 2109 (e.g., home, CTA-Harrison, etc.), a predicted category of the POI 2111 (e.g., home, subway station, etc.), an address of the destination 2113, a predicted mode of the trip 2115 (e.g., car, carpool, etc.), a start time of the trip, and various information. In the example shown, the entries / rows of the augmented trip dataset can be trips or segments of trips.

[0118] In some embodiments, individuals / users can be divided and organized into subgroups / clusters. FIG. 14 An individual division engine 1400 is schematically illustrated and configured to divide individuals in an individual database. Individual division can allow the system to provide offers that can be relevant (e.g., found to be most relevant) to a particular subgroup of users. Individuals or users can be divided (or organized into one or more subgroups) using any suitable division technique. For example, the division technique can be based on fixed division rules 1410. As described above, users can be grouped based on one or more attributes such as geography (or geographic location), social graph, purchase graph, traffic graph, demographic information, user preferences, installed mobile applications, and other user attributes or features extracted from user profile data. Users in the same group can share one or more user attributes or user features (e.g., age, gender, geographic location, social graph, frequent flyer, frequent food purchaser, etc.). Individuals can belong to one or more subgroups. In some cases, subgroups can be continuously expanded and updated automatically as new data is collected. In some embodiments, new subgroups can be created as new users (or categories of users) are added or subscribed to the system. In some cases, subgroups can be discrete. In other cases, two or more subgroups can overlap and can share a set of commonalities or features.

[0119] In some cases, the partitioning technique can be based on patterns extracted from historical data (e.g., user profile data). The patterns can be extracted using a machine learning algorithm. In some cases, an initial set of patterns can be generated, and an algorithm can be employed to identify an optimal allocation of the patterns to the subpopulations that is both feasible and maximizes a desired outcome. The desired outcome can be to provide a small number of service or transaction options related to a predicted destination to be sent to a properly selected customer (e.g., customer group) at an appropriate time and / or location such that the selected customer is most likely to accept the service. Any suitable method such as a decision tree or other pattern recognition algorithm can be used to generate the initial set of patterns. In some cases, the algorithm for identifying an optimal allocation of the patterns to the subpopulations can be a trained machine learning algorithm (e.g., a support vector machine or neural network).

[0120] In some implementations, labels can be created for the user subpopulations. FIG. 15 Examples of user subpopulation labels are shown (e.g., likes to drive, frequent flyer, likes Italian food, sports fan, etc.). The labels can be manually created by one or more individuals, agencies, or imported from external systems or resources. Alternatively or additionally, the labels for user partitioning (e.g., likes to drive, frequent flyer, likes Italian food, sports fan, etc.) can be identified using a machine learning method as described above. In some cases, the labels for user partitioning can be used to supplement the augmented personal data records in the personal database.

[0121] FIG. 16 An example augmented personal data record 1600 is shown having all labels and organized according to user partitioning. The augmented personal data records can be managed, maintained, and updated by the system on a regular basis (e.g., hourly, daily, weekly, etc.) or upon detecting that new data is added to the database. The augmented personal data records can be used to form a training data set for continuous training of the classifier to improve performance. As described above, the training data set can be labeled data including intents or destinations obtained using the clustering system.

[0122] The system can be able to predict intents for future trips without relying on user schedule / planner data (e.g., calendar, emails, etc.). The system can be able to predict likely destinations for current trips based on real-time or limited location data. For example, a trained classifier can be deployed for making predictions of intents and / or travel patterns based on real-time location data. In some cases, the predicted intents and / or travel patterns can be generated based on location data of a trip origin. In some cases, the predicted intents and / or travel patterns can be generated during a trip and updated as new location data flows in.

[0123] FIG. 17A new journey intent predictor 1700 of a system according to some embodiments is schematically illustrated. The new journey intent predictor 1700 can use one or more trained classifiers downloaded from the journey classifier knowledge base 1300 to predict the destination and / or intent of a journey based on real-time location data. The new journey intent predictor 1700 can be coupled to the augmented personal database 420 and the journey classifier knowledge base 1300 to download the appropriate classifiers and retrieve the corresponding augmented personal data records to form an input data set along with the real-time location data. In some cases, when receiving real-time data such as a personal ID, a journey start location (e.g., a starting GPS location), location data updated as the journey progresses, or a mode of transportation, the new journey intent predictor 1700 can retrieve the corresponding augmented personal data records from the augmented personal database 420, extract data such as customer segment, journey type, etc. to supplement the real-time data to form an input data set to be processed by the classifiers downloaded from the journey classifier knowledge base 1300.

[0124] FIG. 18 An example process of continuously generating and updating the prediction of journey destination or intent as new data is collected during the journey is shown. As shown in the example, upon receiving the origin location data (e.g., journey start location), the new journey intent predictor can initially generate three predicted intents, e.g., a journey to work in the morning, a journey to the airport, a journey to buy food in the morning. In some cases, as new location data is collected during the journey, the predicted intents can be refined and updated. For example, as the journey progresses and location data indicating a stop at a coffee shop is collected, the predicted intent is updated to buy food and the predicted destination is a supermarket.

[0125] In some embodiments, the provided system can employ an edge intelligence paradigm such that at least a portion of the data processing can be performed at the edge. For example, the data processing and inference can be performed by the new journey intent predictor deployed on the user device. In some cases, the machine learning models or classifiers can be built and trained on the cloud, stored and maintained in the journey classifier knowledge base 1300, and run on the edge device or edge system (e.g., hardware accelerator). The system and method of the present disclosure can provide an efficient and highly scalable intent prediction platform that enables real-time, on-the-spot journey destination and intent prediction.

[0126] Predicted intentions or trip destinations can be used in a variety of applications. For example, predicted intentions or destinations can be used to provide transaction offers, travel convenience services, provide offers related to predicted destinations, and various other services. For example, predicted intentions or destinations can be used to assist a user in getting to a scheduled destination by providing information on a quick route, parking, or available modes of transportation. In another example, predicted intentions or destinations can be used to provide offers related to a destination, such as a discount on a good offered by a predicted destination (e.g., a supermarket) or a service near the destination (e.g., a coffee shop of the supermarket). In another example, predicted intentions or destinations can be used to influence a user's original intention to divert the user away from a predicted destination (e.g., provide a discount, a service offered by a competing service provider).

[0127] FIG. 19 An example process 1900 of predicting a user's trip intention is shown. In the example shown, location data such as GPS data can be acquired and analyzed to create a plurality of individual trips (operation 1901). Real-time location data can be obtained from a user device (e.g., sensors on the user device) or a user application operating on the user device, such as a map and navigation application. POIs can be identified and assigned to trip end locations and / or trip start locations (operation 1903). Personal trip data can be created by associating one or more individual trips with a personal ID (operation 1905). The personal trip data can be augmented with additional user data (operation 1907). The augmented personal trip data is clustered to identify trip intentions (operation 1909). One or more trip classifiers are trained using the augmented personal trip data (operation 1911). The trip classifiers can be executed and trip intentions, destinations, or modes of transportation can be predicted before a new trip, at the start of a trip, or during a trip (operation 1913). In some cases, one or more offers / services related to the predicted intention / destination can be displayed to the user on a user interface.

[0128] Although FIG. 19 A method according to some embodiments is shown, but one of ordinary skill in the art will recognize that there are many modifications to the various embodiments. For example, the operations can be performed in any order. Some operations can be excluded, some operations can be performed simultaneously in one step, some operations repeated, and some operations can include sub-steps of other operations. In some cases, the timing of providing one or more business options can be based on the user's current geographic location and / or travel time. The method can also be modified according to other aspects of the disclosure as provided herein.

[0129] An individual's trip data can be analyzed to extract various insights. Such insights can be extracted based at least in part on the predicted intent of each trip associated with the individual. FIG. 22 Examples of insight data extracted from trips associated with a user are shown. For example, a plurality of predicted trip intents associated with an individual can be analyzed based on the city or location in which the trips occurred. For example, trips of an individual occurring in a home city can be analyzed and a chart 2210 showing the number of trips organized by trip intent can be generated. Similarly, a chart 2220 can be provided showing the number of trips organized by trip intent in a visited city. Such charts can beneficially provide insights into the individual's travel patterns in different cities. In some embodiments, graphs showing trip patterns associated with an individual can also be generated. For example, a graph 2230 can show the number of trips using different modes (e.g., private car, car rental, ride hailing, walking, multi-modal, etc.) in a home city, while another graph 2240 can show the number of trips using different modes in a visited city. Such insights can beneficially provide a measure of mobility as an effort by a town, city, state, country, or any level of service. For example, car manufacturers, vehicle insurance companies, city, county, state, and national governments, retailers (e.g., supermarkets, home improvement retailers, department stores, etc.), hospitality service companies (e.g., hotels, restaurants, coffee chains), and / or mobility service companies / agencies can use such measures to understand the extent of consumer acceptance of mobility and other types of transportation-related services.

[0130] In some embodiments, the personal database can be further augmented with predicted insights (e.g., travel preferences) about an individual and / or recommendations predicted based on the insights. FIG. 23 An example of a system including an insight generator 2305 and a recommendation engine 2307 is shown. The traveler database 2301 can be the same as the personal database as described above. For example, as described above, the traveler database can store augmented personal data records including user segment labels, trips, demographic characteristics, trip types, and various other data generated by ingesting third party data 2301 and performing trip creation and user segmentation 2303. The augmented personal data records can also include insight data generated by the insight generator 2305 and the recommendation engine 2307. The insight generator 2305 and the recommendation engine 2307 can include one or more machine learning algorithm trained models for predicting a set of preferences of an individual and recommendations related to transportation services.

[0131] FIG. 24The insight generator 2405 is shown schematically. In some cases, the insight generator 2405 can be trained to predict insights about one or more user preferences. The one or more user preferences can relate to an individual's travel preferences, such as preferred modes, preferred destination types, types of accommodations, or various other preferred services, such as preferred restaurants, price points, brand affinities, and the like. For example, the insight generator 2405 can include one or more machine learning-based models 2407 and a transportation ontology 2403 that processes personal data records from a traveler database 2401 to predict user preferences, augments the personal data records with the predicted user preferences, and stores the augmented personal data records in a traveler database 2409. FIG. 25 An example of a transportation ontology 2403 is shown. In some cases, the insight generator can utilize one or more machine learning-based models or machine learning techniques as described in the user segmentation to predict segmentation labels associated with an individual. In some cases, the insight generator can make inferences based on the segmentation labels, other personal data, and different ontologies to generate additional insights about a user.

[0132] In some cases, the recommendation engine can be trained to generate recommendations based at least in part on predicted preferences of an individual and predicted intent of an individual's itinerary. For example, the recommendation engine can be trained to generate recommendations during an itinerary (e.g., recommendations for restaurants, modes for next itinerary leg, and the like).

[0133] FIG. 26 and FIG. 27Examples of insight data about an individual are shown. Insight data can be extracted from augmented trips 2601, 2603 associated with an individual. Insight data can be generated by an insight generator and / or a recommendation engine as described above. For example, augmented trips associated with an individual in different cities (resident city 2601, visited city 2603) can be processed to extract further insights about the individual. For example, based on predicted intent, patterns, or other information in the augmented trip data, user preference insights can be inferred, such as preferred mode 2605 (e.g., personal vehicle, carpool, walk), preferred hotel price range 2607 (e.g., mid-range), preferred cuisine 2609, persona 2611 (e.g., active lifestyle traveler, foodie, health and fitness, etc.), preferred restaurant price range 2613, most preferred brands (e.g., Peets, Target, etc.), and various other insights. Inferred user preferences can also include POI preferences such as preferred dining out days 2601, preferred modes of travel to restaurants in a particular city 2703, 2705 (e.g., carpool in Denver, drive in Chicago), preferred business travel days 2707, preferred travel accommodations, preferred travel restaurants 2711, or preferred travel modes 2713. In some cases, such insight data can be utilized to make further predictions 2715, such as recommended restaurants, hotels, services to provide in a particular city, or predicted trips.

[0134] In some cases, insights about groups of travelers can be provided. FIG. 28 Examples of insight data about frequent business travelers from a particular geographic location are shown. Insight data can include users 2801 who frequently take business trips from a particular location, the number of trips from the particular location associated with each individual 2805, and device IDs 2803 associated with the individuals. Such insights can be generated using methods and systems as described above, such as associating unrelated trips with an individual and predicting intent of trips.

[0135] Inferences about trips associated with a group of users can also be made using the provided methods and systems. FIG. 29 Examples of inferences or insights about a group or user are shown. For example, a chart 2601 showing the frequency of trip categories (e.g., restaurants, national parks, gyms) associated with a group of users (e.g., travelers in a particular city) can be generated. In some cases, a distribution chart about trip patterns associated with a group of users, such as in a resident city 2920 or a visited city 2930, can be generated.

[0136] COMPUTER SYSTEM

[0137] The systems, various components of the systems, or processes described herein can be implemented by one or more processors. In some embodiments, the processor can be a processing unit of a computer system. FIG. 20 A computer system 2001 is shown, programmed or otherwise configured for implementing a ground travel analysis and intent prediction system. The computer system 2001 can regulate various aspects of the present disclosure. The computer system 2001 can be an electronic device of a user or a computer system located remotely with respect to the electronic device. The electronic device can be a mobile electronic device.

[0138] The computer system 2001 includes a central processing unit (CPU, also "processor" and "computer processor" herein) 2005, which can be a single core or multi core processor, or a plurality of processors for parallel processing. The computer system 2001 also includes memory or memory location 2010 (e.g., random access memory, read only memory, flash memory), electronic storage unit 2015 (e.g., hard disk), communication interface 2020 (e.g., network adapter) for communicating with one or more other systems, and peripheral devices 2025, such as cache, other memory, data storage, and / or electronic display adapters. The memory 2010, storage unit 2015, interface 2020, and peripheral devices 2025 are in communication with the CPU 2005 through a communication bus (solid lines), such as a motherboard. The storage unit 2015 can be a data storage unit (or data repository) for storing data. The computer system 2001 can be operatively coupled to a computer network (“network”) 2030 by means of the communication interface 2020. The network 2030 can be the Internet, an internet and / or an extranet, or an intranet and / or extranet that in turn can include a network, such as the Internet. In some cases, the network 2030 is a telecommunication and / or data network. The network 2030 can include one or more computer servers that can implement a distributed computing methodology, such as cloud computing. In some cases, the network 2030 can implement a peer-to-peer network by means of the computer system 2001, which can enable devices coupled to the computer system 2001 to behave as a client or a server.

[0139] The CPU 2005 can execute a sequence of machine-readable instructions, which can be embodied in a program or software. The instructions can be stored in a memory location, such as the memory 2010. The instructions can be directed to the CPU 2005, which can subsequently program or otherwise configure the CPU 2005 to implement methods of the present disclosure. Examples of operations performed by the CPU 2005 can include fetch, decode, execute, and writeback.

[0140] CPU 2005 may be part of a circuit such as an integrated circuit. One or more other components of system 2001 may be included in the circuit. In some cases, the circuit is an application-specific integrated circuit (ASIC).

[0141] Storage unit 2015 may store files, such as drivers, libraries, and saved programs. Storage unit 2015 may store user data, such as user preferences and user programs. In some cases, computer system 2001 may include one or more additional data storage units located outside computer system 2001 (e.g., on a remote server communicating with computer system 2001 via an intranet or the Internet).

[0142] Computer system 2001 can communicate with one or more remote computer systems via network 2030. For example, computer system 2001 can communicate with a user's remote computer system (e.g., a user device). Examples of remote computer systems include personal computers (e.g., portable PCs), tablets, or tablet PCs (e.g., Apple PCs). iPad, Samsung Galaxy Tab), telephone, smartphone (e.g., Apple). iPhone, Android-enabled devices, Blackberry (or personal digital assistant). Users can access computer systems via network 2030 2001.

[0143] The methods described herein can be implemented using machine-executable code (e.g., a computer processor) stored in an electronic storage location (e.g., such as memory 2010 or electronic storage unit 2015) of computer system 2001. The machine-executable or machine-readable code can be provided in software form. During use, the code can be executed by processor 2005. In some cases, the code can be retrieved from storage unit 2015 and stored in memory 2010 for access by processor 2005 at any time. In some cases, electronic storage unit 2015 can be excluded, and the machine-executable instructions are stored in memory 2010.

[0144] The code can be pre-compiled and configured for use with machines having processors suitable for executing the code, or it can be compiled at runtime. The code can be supplied in a programming language, which can be selected to enable the code to be executed in a pre-compiled or on-site compiled manner.

[0145] Various aspects of the systems and methods provided herein, such as computer system 2001, can be embodied in programming. Various aspects of the technology can be thought of as "products" or "articles of manufacture" typically in the form of machine (or processor) executable code and / or associated data that is carried or embodied on or in a type of machine readable medium. Machine-executable code can be stored on or transmitted from an electronic storage medium, such as memory (e.g., read-only memory, random-access memory, flash memory) or a hard disk. "Storage" type media can include any or all of the tangible memory of the computers, processors or the like, or associated modules thereof, such as various semiconductor memories, tape, magnetic materials, optical media, and the like, which can be used to provide non-transitory storage of software to the computers or processors as needed. The software elements can be embodied in many ways, including a performance program, routine, software component, proprietary software application, proprietary software packages, script, database, database records, objects, exes, spreadsheets, formatting information, and the like. A typically, software embodiments are implemented as tangible, machine executable code and / or data which can be stored on a non-transitory, tangible medium, or in transit on a carrier wave or propagated signal. A typical process program, when executed, allows the computing system to perform one or more activities. The software and / or code elements can be stored on any type of non-transitory, tangible medium or memory, including non-volatile and volatile memory, such as hard drives, floppy disks, CD ROMs, DVDs, RAM, ROM, PROM, EPROM, EEPROM, and the like. The software elements can also be transmitted or downloaded from a network using a transmission medium, such as the Internet or Intranet. Transmissions media can include a wired or wireless links as well as physical media such as optical, electromagnetic, and the like. The embodiments are not limited in this context.

[0146] Accordingly, a machine readable medium, such as a computer-readable medium, can take many forms, including but not limited to, a tangible storage medium, a carrier wave medium or physical transmission medium. Non-volatile storage media include, for example, optical or magnetic disks, such as any of the storage devices in any computer(s) or the like, such as can be used to implement the databases, etc. shown in the drawings. Volatile storage media include dynamic memory, such as the main memory of such a computer platform. Tangible transmission media include coaxial cables; copper wire and fiber optic cables, including the wires that comprise a bus within a computer system. Carrier-wave transmission media can take the form of electric or electromagnetic signals, or the like, propagating through a physical transmission medium or over the air, such as during radio frequency (RF) and infrared (IR) data communications. Thus, the computer-readable medium can also take the form of a signal propagating in a network such as the Internet or over the air, such as during radio frequency (RF) and infrared (IR) data communications, or the like. Accordingly, computer-readable media are medium that can be accessed by a computer to provide program code to the computer for execution during alternate phases of operation, such as the instructions providing operating logic for performing various aspects. The various aspects will be described in the general context of computer-readable media of a computer system, such as those systems described above and in the figures. Any suitable computer readable medium can be utilized including a tangible or computational memory of the computer(s) 2001, or the like, including magnetic bases, optical storage devices, and / or any combination thereof. The computer-readable media includes permanent memory, i.e., memory that does not change when the computer reads from or writes to it. On the other hand, the computer-readable media includes volatile memory, i.e., where the computer stores temporary variables or other intermediate information used during the execution of instructions by the computer. Use of the word "or" in reference to terms such as "computer-readable medium" is intended to cover both

[0147] The computer system 2001 can include or be in communication with an electronic display 2035, which can comprise a user interface (UI) 2040 for providing a graphical user interface, e.g., as otherwise described herein. Examples of UIs include, without limitation, graphical user interfaces (GUIs) and web-based user interfaces.

[0148] The methods and systems of this disclosure can be implemented by way of one or more algorithms. An algorithm can be implemented by way of software upon execution by the central processing unit 2005. For example, an algorithm can train a model, such as a transportation planning engine.

[0149] While preferred embodiments of the application have been shown and described herein, it will be apparent to those skilled in the art that such embodiments are provided by way of example. The application is not intended to be limited by the specific examples provided in the specification. While the application has been described with reference to the aforementioned specification, the descriptions and illustrations of the embodiments in this document are not meant to be construed in a limiting sense. Numerous variations, changes, and substitutions will now occur to those skilled in the art without departing from the application. Furthermore, it shall be understood that all aspects of the application are not limited to the specific depictions, configurations or relative proportions set forth herein which depend on a variety of conditions and variables. It should be understood that various alternatives to the embodiments of the application described herein can be employed in practicing the application. It is therefore contemplated to cover by the present application any and all alternatives, modifications, variations, or equivalents. The following claims are intended to define and cover the application and all alternatives and equivalents thereof.

Claims

1. A method for predicting a trip intent of a user trip taken by a user to determine information provided to the user during the user trip, comprising: (a) receiving a starting geographic location of a travel route and data regarding an identity of the user; (b) training a trip abstractor by: (1) obtaining a plurality of training prior trip records, wherein a prior trip record of the plurality of training prior trip records includes trip features and a trip intent of a prior trip for which the intent has been determined; (2) generating a plurality of models, wherein for at least two models of the plurality of models, each model of the at least two models has a related trip intent that is different from a trip intent associated with the other model of the at least two models, and each model of the at least two models includes a trip classifier trained on a subset of the plurality of training prior trip records, the subset being prior trip records having trip intents that match the related trip intent; and (3) storing the plurality of models into a trip classifier database; (c) identifying a trip data set from location data collected by the user trip using a trip recognition engine, wherein the location data provides geospatial points visited by the user during the user trip, wherein at least some of the visited points are provided to the trip recognition engine in real time; (d) applying the trip data set and the trip classifier database to an intent predictor to determine a determined trip intent of the user trip, wherein the intent predictor uses at least two models of the plurality of models to determine the determined trip intent of the user trip; and (e) presenting to the user on an electronic device one or more transaction options identified based at least in part on the determined trip intent of the user trip when the user is traveling along at least a portion of the travel route.

2. The method of claim 1, wherein the starting geographic location is received in the form of global positioning system (GPS) data.

3. The method of claim 1, wherein the starting geographic location is input by the user through a graphical user interface (GUI) on the electronic device, or is determined in part using a geographic location of the electronic device determined through global positioning system or signal triangulation.

4. The method of claim 1, wherein the plurality of training prior trip records includes unassociated GPS data.

5. The method of claim 4, wherein the plurality of training prior trip records includes tagged data obtained using a cluster analysis of a plurality of trip data records, wherein the cluster analysis automatically identifies subsets of trips having common features.

6. The method of claim 5, further comprising generating the plurality of trip data records by associating the unassociated GPS data with one or more personal identities.

7. The method of claim 5, wherein the plurality of trip data records are augmented by social graphs, traffic data, or purchase data of corresponding personal identities. ​ ​ 8. The method of claim 1, wherein training the trip abstractor includes creating a label for a leg of a trip based on one or more labeling rules.

9. The method of claim 1, further comprising predicting a mode of transportation for one or more portions of the travel route.

10. The method of claim 9, wherein the mode of transportation includes an autonomous vehicle, a ride-hailing service, rail transit, and / or a land-based public transit vehicle.

11. The method of claim 1, further comprising updating the trip intent or destination when new location data is received during the user trip.

12. A system for predicting a trip intent of a user trip taken by a user to determine information provided to the user during the user trip, comprising: a trip abstractor trained by: (1) obtaining a plurality of training prior trip records, wherein a trip record of the plurality of training prior trip records includes trip characteristics and a trip intent of a prior trip for which the intent has been determined; (2) generating a plurality of models, wherein for at least two models of the plurality of models, each model of the at least two models has a related trip intent that is different from a trip intent associated with the other model of the at least two models, and each model of the at least two models includes a trip classifier trained on a subset of the plurality of training prior trip records, the subset being prior trip records having a trip intent that matches the related trip intent; a trip classifier database for storing the plurality of models generated by the trip abstractor; a trip recognition engine for recognizing a trip data set from location data collected from a user trip, wherein the location data includes a travel route represented by geospatial points visited by the user during the user trip, wherein at least some of the visited points are provided to the trip recognition engine in real-time; an intent predictor coupled to the trip recognition engine and the trip classifier database for determining a determined trip intent of the user trip based on at least the following factors while the user is traveling in a vehicle along at least a portion of the travel route: (1) a matching model of the plurality of models that matches in that its trip intent matches the determined trip intent of the user trip, (2) a starting geolocation of the travel route, and (3) data about the user’s identity; and a user interface of an electronic device adapted to present one or more transaction options identified based at least in part on the determined trip intent of the user trip for presentation to the user while the user is traveling along the travel route.

13. The system of claim 12, wherein the starting geolocation is received in the form of global positioning system (GPS) data.

14. The system of claim 12, wherein the starting geolocation is input by the user through a graphical user interface (GUI) on the electronic device, or is determined in part using a geolocation of the electronic device determined by a global positioning system or signal triangulation.

15. The system of claim 12, wherein the plurality of training prior trip records comprise irrelevant GPS data.

16. The system of claim 15, wherein the plurality of training prior trip records comprise labeled data obtained using a cluster analysis of a plurality of trip data records.

17. The system of claim 16, wherein the plurality of trip data records associate the irrelevant GPS data with one or more personal identities.

18. The system of claim 16, wherein the plurality of trip data records are augmented by social graphs, traffic data, or purchase data of the respective personal identities.

19. The system of claim 12, wherein the intent predictor is further configured to predict a mode of transportation for one or more portions of the travel route.

20. The system of claim 19, wherein the mode of transportation comprises an autonomous vehicle, a ride-hailing service, rail transit, and / or land public transit vehicles.

Images