Artificial intelligence-based commuting network evaluation method and system

Abstract

The invention provides a commuting network assessment method and system based on artificial intelligence, and the method comprises the steps: generating a smooth manifold of a commuting network in a Lie group space through S1, constructing a dynamic traffic manifold according to S2, constructing a multi-agent forest model through S3, and generating a commuting score vector through the agent forest model. And S4, generating a score relation equation to integrate commuting score vectors, marking three commuting paths with the highest scores and pushing the three commuting paths to the user, considering various commuting factors to make the final commuting score more accurate, adding historical data of the user at the same time, better meeting the requirements of the user, and improving the user experience. And S5, obtaining a final commuting path of the user, performing tracing, generating a tracing report, optimizing the dynamic traffic manifold according to S6, returning to S2, searching and analyzing factors influencing scores through a tracing method, and optimizing the dynamic traffic manifold according to an analysis result, thereby achieving the purpose of optimization evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of traffic commuting control, in particular to a commuting network evaluation method and system based on artificial intelligence. BACKGROUND

[0002] With the acceleration of urbanization and the popularization of intelligent transportation systems, efficient commuting path planning has become a key link to improve the efficiency of urban operation. Traditional path evaluation methods mainly rely on static road network topology and fixed weight scoring models, such as shortest path search based on xx algorithm or time optimal planning based on xx algorithm, which have single evaluation dimension and are difficult to adapt to dynamic traffic environment. To cope with complex road conditions, existing technologies gradually introduce artificial intelligence learning models to respond to some dynamic factors.

[0003] Most of the existing technologies focus on the evaluation of a single indicator, lack of systematic analysis and evaluation of multi-source data, and cannot integrate user opinions and feedback. Therefore, there is an urgent need for a commuting evaluation method that combines continuous spatiotemporal modeling and user feedback to improve the accuracy of evaluation. SUMMARY

[0004] In view of the above-mentioned problems, in combination with the first aspect of the present application, the embodiments of the present application provide a commuting network evaluation method based on artificial intelligence, which comprises: S1: obtaining a commuting network and mapping the commuting network to a Lie group space to generate a smooth manifold of the commuting network in the Lie group space; S2: obtaining real-time traffic flow data, traffic light phase timing, and road topology metadata, and constructing a dynamic traffic manifold based on the real-time traffic flow data, the traffic light phase timing, the road topology metadata, and the smooth manifold; S3: constructing a multi-agent forest model, obtaining a user input commuting starting point and a commuting ending point, and generating a commuting score vector based on the commuting starting point, the commuting ending point, and the dynamic traffic manifold, wherein the commuting score vector includes a travel time score, an energy consumption score, and an abnormal event influence probability score; S4: generating a score relationship equation based on user historical preferences, integrating the commuting score vector based on the score relationship equation, obtaining a commuting path score set, and marking and pushing the three commuting paths with the highest scores in the path score set to the user; S5: obtaining a user's final commuting path and tracing the final commuting path, and generating a trace report; S6: optimizing the dynamic traffic manifold based on user feedback and the trace report, and returning to step S2.

[0005] As a further solution of the present invention, the step of obtaining a commuting network and mapping the commuting network to a Lie group space to generate a smooth manifold of the commuting network on the Lie group space includes: The commuting network includes commuting nodes and commuting routes. The commuting nodes and commuting routes are mapped to the Lie group space. The commuting nodes correspond to elements in the Lie group space, and the commuting routes correspond to group action operators in the Lie group space. A smooth manifold is generated based on the commuting nodes and commuting routes.

[0006] As a further embodiment of the present invention, the step of acquiring real-time traffic flow data, traffic light phase sequence, and road topology metadata, and constructing a dynamic traffic manifold based on the real-time traffic flow data, traffic light phase sequence, road topology metadata, and the smooth manifold includes: Acquire real-time traffic flow data with a sampling frequency of once per second, including vehicle location and motion status; Obtaining a traffic light phase sequence with a time resolution of 0.1 seconds, wherein the traffic light phase sequence includes the change time and state of the traffic signal; Obtaining road topology metadata, wherein the road topology metadata includes the number of lanes, slope, and curvature radius of the road; A commuting differential equation is constructed based on real-time traffic flow data, traffic light phase sequence, and road topology metadata, and a dynamic traffic manifold is constructed based on the commuting differential equation and the smooth manifold.

[0007] As a further solution of the present invention, the method of constructing a commuting differential equation based on real-time traffic flow data, traffic light phase sequence, and road topology metadata, and constructing a dynamic traffic manifold based on the commuting differential equation and the smooth manifold includes: Obtaining a commuting position of a vehicle on a smooth manifold, obtaining a traffic flow tensor based on real-time traffic flow data and road topology metadata, wherein the traffic flow tensor represents a flow state of traffic flow, and composing the position and the traffic flow tensor into a commuting vector; Obtain traffic signal changes based on traffic light phase timing; The commuting vector is differentiated over the changing time to obtain a commuting differential equation, and a dynamic traffic manifold is constructed with the smooth manifold. The changing time is represented by the time corresponding to before and after the traffic signal changes.

[0008] As a further solution of the present invention, the multi-agent forest model is constructed to obtain the commuting start point and commuting destination input by the user. The agent forest model generates a commuting score vector based on the commuting start point, commuting destination, and dynamic traffic manifold. The commuting score vector includes a travel time score, an energy consumption score, and an abnormal event impact probability score, including: The user enters the commuting start and end points into the mobile user terminal. The mobile user terminal analyzes the commuting start and end points and the dynamic traffic manifold based on the multi-agent forest model to generate a commuting score vector. The multi-agent forest model includes a time agent, an energy consumption agent and a risk agent.

[0009] As a further embodiment of the present invention, the method further comprises: The time agent predicts the travel time based on the LSTM to obtain the predicted travel time, and scores the predicted travel time to obtain a travel time score; The energy consumption agent solves energy based on a dynamic equation to obtain commuting energy consumption, and scores the energy consumption to obtain an energy consumption score; The risk agent analyzes the impact of abnormal events based on the Bayesian network to obtain the impact probability of the abnormal events, and scores the impact probability of the abnormal events to obtain the impact probability score of the abnormal events.

[0010] As a further solution of the present invention, generating a rating relationship equation based on user historical preferences, integrating commuting rating vectors based on the rating relationship equation, obtaining a commuting path rating set, and marking the three highest-rated commuting paths in the commuting path rating set and pushing them to the user include: Obtaining user historical preferences based on historical data, evaluating and proportioning travel time scores, energy consumption scores, and abnormal event impact probability scores based on user historical preferences, and generating a scoring relationship equation; The travel time score, energy consumption score, and abnormal event impact probability score are input into the scoring relationship equation to generate a commuting path score set. The three commuting paths with the highest scores in the commuting path score set are extracted and marked, and the three commuting paths with the highest scores are pushed to the user.

[0011] As a further solution of the present invention, extracting the three commuting routes with the highest scores from the commuting route score set and marking them, and pushing the three commuting routes with the highest scores to the user, includes: The marks include time marks, energy consumption marks and comprehensive marks; If the travel time score corresponding to the extracted commuting route is the highest, the extracted commuting route is marked as having the shortest travel time; If the energy consumption score corresponding to the extracted commuting route is the highest, the extracted commuting route is marked as having the lowest energy consumption; If the extracted commuting route has the highest score, the extracted commuting route is marked as having the highest comprehensive score; The three commuting routes and their corresponding scores after being marked are pushed to the user.

[0012] As a further solution of the present invention, obtaining the user's final commuting route, tracing the final commuting route, and generating a tracing report include: Obtaining a final commuting route and a corresponding score based on the user's selection, wherein the final commuting route is represented as the commuting route ultimately selected by the user, obtaining a commuting score vector corresponding to the score, and performing a partial derivative of the corresponding commuting score vector to obtain a sensitivity of the commuting score vector, thereby obtaining a commuting impact matrix; Screen key features based on the commuting impact matrix to obtain a commuting impact list; Build a dynamic decision tree based on the commuting impact list, and generate a traceability report based on the dynamic decision tree.

[0013] In another aspect, an embodiment of the present invention further provides an artificial intelligence-based commuting network evaluation system, comprising: An acquisition module, which is used to acquire the commuting network, obtain real-time traffic flow data, traffic light phase timing, and road topology metadata, and obtain the user's final commuting path; A generation module is configured to map the commuting network to a Lie group space, generating a smooth manifold of the commuting network in the Lie group space. The agent forest model generates a commuting score vector based on the starting point, the end point, and the dynamic traffic manifold, and generates a score relationship equation based on the user's historical preferences. A construction module, wherein the construction module constructs a dynamic traffic manifold based on real-time traffic flow data, traffic light phase time series, road topology metadata, and the smooth manifold, and constructs a multi-agent forest model; An integration module, which integrates the commuting score vectors based on the score relationship equation to obtain a commuting path score set, and marks the three commuting paths with the highest scores in the path score set and pushes them to the user; A tracing module, which is used to trace the final commuting path and generate a tracing report; An optimization module optimizes the dynamic traffic flow shape based on user feedback and traceability reports.

[0014] Based on the above aspects, the embodiment of the present application realizes obtaining the commuting network through step S1, and mapping the commuting network to the Lie group space, generating a smooth manifold of the commuting network on the Lie group space, obtaining real-time traffic flow data, traffic light phase time series and road topology metadata according to step S2, constructing a dynamic traffic manifold through real-time traffic flow data, traffic light phase time series, road topology metadata and the smooth manifold, constructing a multi-agent forest model through step S3, obtaining the commuting start point and commuting end point input by the user, and the agent forest model generates a commuting score vector based on the commuting start point, commuting end point and dynamic traffic manifold, and the commuting score vector includes a travel time score, an energy consumption score and an abnormal event impact probability score, generating a score relationship equation through step S4, integrating the commuting score vector according to the score relationship equation, and obtaining a commuting path score. The three commuting paths with the highest scores in the path scoring set are marked and pushed to the user. By mapping the commuting network to the Lie group space, the travel time and energy consumption are converted into commuting differential equations and a dynamic traffic manifold is constructed. The paths generated by the dynamic traffic manifold are scored through the multi-agent forest model, and a variety of commuting factors are taken into account to make the final commuting score more accurate. At the same time, the user's historical data is added to better meet the user's needs. The user's final commuting path is obtained through step S5, and the final commuting path is traced, and a traceability report is generated. The dynamic traffic manifold is optimized according to step S6, and the process returns to step S2. The factors affecting the score are found through the traceability method, and the influencing factors are analyzed to obtain the analysis results. The dynamic traffic manifold is optimized based on the analysis results to achieve the purpose of optimization evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a schematic diagram of the execution flow of an artificial intelligence-based commuting network evaluation method provided by an embodiment of the present invention.

[0016] Figure 2 Schematic diagram of an artificial intelligence-based commuting network evaluation system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The present invention will be described in detail below with reference to the accompanying drawings. Figure 1 This is a schematic diagram of the execution flow of an artificial intelligence-based commuting network evaluation method provided by an embodiment of the present invention. The artificial intelligence-based commuting network evaluation method is introduced in detail below.

[0018] Step S1: Obtain a commuting network and map the commuting network to a Lie group space to generate a smooth manifold of the commuting network in the Lie group space.

[0019] The commuting network comprises commuting nodes and commuting routes, the commuting nodes and the commuting routes are mapped to a Lie group space, the commuting nodes correspond to elements in the Lie group space, and the commuting routes correspond to group action operators in the Lie group space, and a smooth manifold is generated based on the commuting nodes and the commuting routes.

[0020] It should be noted that the Lie group space is a kind of mathematical object with special structure, which combines the characteristics of "group" and "manifold". Specifically, the Lie group is a set that satisfies the operation rules of the group and has the structure of the smooth manifold. Then, the complex commuting network can be operated on the Lie group for calculus.

[0021] Specifically, the commuting nodes and routes in the commuting network are mapped to the Lie group space, the intersections are regarded as the commuting nodes in the commuting network, and the roads are regarded as the commuting routes in the commuting network. Each commuting node corresponds to an element in the Lie group, represents a position or transformation in space, each commuting route corresponds to a group action operator in the Lie group, describes the transformation relationship between nodes, and through the mapping relationship, the entire commuting network forms a smooth manifold structure in the Lie group space, so that the analysis and research on the commuting behavior and its changes have the mathematical basis of continuity and differentiation.

[0022] In step S2, real-time traffic flow data, traffic light phase timing and road topology metadata are obtained, and a dynamic traffic manifold is constructed based on the real-time traffic flow data, the traffic light phase timing, the road topology metadata and the smooth manifold.

[0023] In this embodiment, step S2 comprises: In step S21, real-time traffic flow data is obtained, and the sampling frequency is 1 time per second. The real-time traffic data includes vehicle position and motion state.

[0024] Specifically, the real-time traffic flow data is obtained by the obtaining module 101, and the sampling frequency of the obtaining module 101 is 1 time per second. The real-time traffic flow data includes vehicle position and motion state. For example, the vehicle xx is driving on xx avenue at a speed of 60 kilometers per hour, and its longitude and latitude is represented as: longitude 116.4074, latitude 39.9042 degrees north.

[0025] The traffic light phase timing is obtained by the obtaining module 101, and the time resolution is 0.1 seconds. The traffic light phase timing includes the change time and state of the traffic signal.

[0026] Specifically, the traffic light phase timing is obtained by the obtaining module 101, and the time resolution is 0.1 seconds. The traffic light phase timing includes the change time and state of the traffic signal. For example, the red light of the red traffic light on xx avenue is on for 20 seconds, the green light is on for 15 seconds, and the yellow light is on for 3 seconds.

[0027] Obtain road topology metadata, where the road topology metadata includes the number of lanes, slope, and curvature radius of the road.

[0028] Specifically, the road topology metadata is obtained by the acquisition module 101, wherein the road topology metadata includes the number of lanes, slope, and curvature radius of the road. For example, xx Avenue includes five lanes, a slope of 15 degrees, and a curvature radius of 150 meters.

[0029] A commuting differential equation is constructed based on real-time traffic flow data, traffic light phase sequence, and road topology metadata, and a dynamic traffic manifold is constructed based on the commuting differential equation and the smooth manifold.

[0030] Step S22, obtain the commuting position of the vehicle on the smooth manifold, obtain the traffic flow tensor based on the real-time traffic flow data and the road topology metadata, the traffic flow tensor is represented as the flow state of the traffic flow, and the position and the traffic flow tensor are combined into a commuting vector.

[0031] Obtain traffic signal changes based on traffic light phase timing.

[0032] The commuting vector is differentiated over the changing time to obtain a commuting differential equation, and a dynamic traffic manifold is constructed with the smooth manifold. The changing time is represented by the time corresponding to before and after the traffic signal changes.

[0033] In this embodiment, the commuting position of the vehicle on the smooth manifold is obtained by the acquisition module 101, and the flow state of the real-time traffic flow is obtained by analysis based on the above-mentioned real-time traffic flow data and road topology metadata, and the position and the flow state of the real-time traffic flow are combined into a commuting vector. At the same time, the traffic signal change is obtained through the traffic light phase sequence.

[0034] Furthermore, the above-mentioned positions, the flow state of the real-time traffic flow and the traffic signal changes are input into the smooth manifold, and the dynamic traffic manifold is constructed by the construction module 103 .

[0035] Specifically, the commuting vector is differentiated over the changing time to obtain a commuting differential equation, which is expressed as: ; ; in, Expressed as a commuting vector, Expressed as the position coordinates of the vehicle on the dynamic traffic manifold, Represented as the flow state of real-time traffic flow, It is represented as an external stimulus, and the external stimulus is represented as a traffic signal change.

[0036] Step S3: Build a multi-agent forest model to obtain the commuting start and end points input by the user. The agent forest model generates a commuting score vector based on the commuting start and end points and the dynamic traffic manifold. The commuting score vector includes a travel time score, an energy consumption score, and a probability score for the impact of abnormal events.

[0037] In this embodiment, step S3 includes: In step S31, the user inputs the commuting start point and commuting end point into the mobile user terminal. The mobile user terminal analyzes the commuting start point, commuting end point, and dynamic traffic manifold based on the multi-agent forest model to generate a commuting score vector.

[0038] Specifically, the user inputs the commuting start point and commuting end point into the mobile user terminal. The multi-agent forest model in the mobile client segment analyzes the user's commuting start point and commuting end point based on the dynamic traffic manifold, and generates a commuting score vector according to the generation module 102.

[0039] Step S32: The multi-agent forest model includes a time agent, an energy consumption agent, and a risk agent.

[0040] It should be noted that the above-mentioned multi-agent forest model includes time agents, energy consumption agents and risk agents. Users pay more attention to commuting time, commuting energy consumption and commuting risks when commuting. The commuting score composed of commuting time, commuting energy consumption and commuting risks focuses more on user experience.

[0041] The time agent predicts the travel time based on the LSTM to obtain the predicted travel time, and scores the predicted travel time to obtain a travel time score.

[0042] It's important to note that the LSTM model, short for long-short-term memory, is a recurrent neural network used in deep learning for processing and predicting time series data. By learning from past historical data, the LSTM can capture temporal dependencies and periodicity in the data, enabling predictions for future time points.

[0043] In this embodiment, LSTM can predict traffic conditions in the next few seconds to minutes based on historical traffic flow and vehicle status, and predict the travel time of each vehicle plan. At the same time, it scores the travel time of each plan to obtain a travel time score.

[0044] For example, the travel time of Plan 1 from A to B is 15 minutes, the travel time of Plan 2 from A to B is 14 minutes, and the travel time of Plan 1 from A to B is 16 minutes. Plans 1, 2, and 3 are scored, with Plan 1 scoring 88 points, Plan 2 scoring 90 points, and Plan 3 scoring 85 points.

[0045] The energy consumption agent solves energy based on a dynamic equation to obtain commuting energy consumption, and scores the consumed energy to obtain an energy consumption score.

[0046] In this embodiment, the acquisition module 101 acquires and solves the real-time speed of the vehicle, the distance of the trip, and the traffic signal changes to obtain commuting energy consumption, and scores the energy consumption to obtain an energy consumption score.

[0047] For example, the vehicle's power consumption in Plan 1 from A to B is less than 1%, the vehicle's power consumption in Plan 2 from A to B is 1%, and the vehicle's power consumption in Plan 3 from A to B is 2%. Plans 1, 2, and 3 are scored, with Plan 1 receiving 95 points, Plan 2 receiving 90 points, and Plan 3 receiving 85 points.

[0048] The risk agent analyzes the impact of abnormal events based on the Bayesian network to obtain the impact probability of the abnormal events, and scores the impact probability of the abnormal events to obtain the impact probability score of the abnormal events.

[0049] It should be noted that the Bayesian network is a probabilistic graphical model used to describe the conditional dependencies between variables. When analyzing the impact of abnormal events, the Bayesian network can help us understand the extent of the impact of abnormal events on other variables in the system. By inferring the conditional probabilities of related nodes in the network, we can evaluate the changes and amplitudes of various variables in the system after the abnormal event occurs, thereby identifying the causes and potential impacts of the abnormality.

[0050] In this embodiment, the acquisition module 101 acquires the abnormal event impact probability, and scores the abnormal event impact probability to acquire the abnormal event impact probability score.

[0051] For example, the probability of an abnormal event affecting Plan 1 from A to B is 10%, the probability of an abnormal event affecting Plan 2 from A to B is 20%, and the probability of an abnormal event affecting Plan 3 from A to B is 50%. Plans 1, 2, and 3 are scored, with Plan 1 receiving 90 points, Plan 2 receiving 85 points, and Plan 3 receiving 50 points.

[0052] Step S4: Generate a rating relationship equation based on the user's historical preferences, integrate the commuting rating vectors based on the rating relationship equation, obtain a commuting path rating set, and mark the three commuting paths with the highest ratings in the path rating set and push them to the user.

[0053] In this embodiment, step S4 includes: Step S41 : obtaining the user's historical preferences based on the historical data, performing rating and proportioning on the travel time score, the energy consumption score, and the abnormal event impact probability score based on the user's historical preferences, and generating a rating relationship equation.

[0054] It should be noted that based on the historical data obtained by the acquisition module 101, the historical preferences of each user are analyzed through the historical data, and the travel time score, energy consumption score and abnormal event impact probability score are rated and proportioned according to the user's historical preferences, and the scoring relationship equation is generated by the generation module 102.

[0055] For example, if a user's historical data shows that the user's commute tends to be shorter, the travel time score, energy consumption score, and abnormal event impact probability score are allocated in a ratio of 6:3:1. If the user's commute tends to be lower in energy consumption, the travel time score, energy consumption score, and abnormal event impact probability score are allocated in a ratio of 3:6:1.

[0056] Specifically, the scoring relationship equation is expressed as: ; in, It is expressed as the travel time rating ratio, Expressed as the energy consumption score ratio, It is expressed as the rating ratio of the probability score of abnormal events. Expressed as a travel time score, Expressed as an energy consumption score, Expressed as the probability score of abnormal event impact, Expressed as a commute score.

[0057] The travel time score, energy consumption score, and abnormal event impact probability score are input into the scoring relationship equation to generate a commuting path score set. The three commuting paths with the highest scores in the commuting path score set are extracted and marked, and the three commuting paths with the highest scores are pushed to the user.

[0058] For example, the travel time score, energy consumption score, and abnormal event impact probability score obtained are 90 points, 80 points, and 80 points respectively. Historical data shows that the user's commute tends to have a shorter travel time. The travel time score, energy consumption score, and abnormal event impact probability score are rated at 6:3:1, and the calculated commuting score is 86 points. If the travel time score, energy consumption score, and abnormal event impact probability score obtained are 90 points, 90 points, and 90 points respectively, historical data shows that the user's commute tends to have lower energy consumption. The travel time score, energy consumption score, and abnormal event impact probability score are rated at 3:6:1, and the calculated commuting score is 90 points. Similarly, the scores of multiple commuting paths are obtained to generate a commuting path score set. The three paths with the highest commuting scores are filtered out from the commuting path score set, marked, and pushed to the user.

[0059] It should be noted that if the abnormal event impact probability score is lower than 60 points, indicating that the probability of an abnormal event is extremely high, the travel time score, energy consumption score, and abnormal event impact probability score are allocated in a ratio of 2:2:6.

[0060] In step S42, the mark includes a time mark, an energy consumption mark and a comprehensive mark.

[0061] If the travel time score corresponding to the extracted commuting route is the highest, the extracted commuting route is marked as having the shortest travel time.

[0062] If the energy consumption score corresponding to the extracted commuting route is the highest, the extracted commuting route is marked as having the lowest energy consumption.

[0063] If the extracted commuting route has the highest score, the extracted commuting route is marked as having the highest comprehensive score.

[0064] The three commuting routes and their corresponding scores after being marked are pushed to the user.

[0065] For example, the three paths obtained are path 1, path 2, and path 3, and the commuting scores of path 1, path 2, and path 3 are 88 points, 90 points, and 92 points respectively. Path 1 has the shortest commuting time, so path 1 is marked as the shortest time. Path 2 has the lowest commuting energy consumption, so path 2 is marked as the lowest energy consumption. Path 3 has the highest comprehensive score, so path 3 is marked as the highest score.

[0066] Step S5: Obtain the user's final commuting path, trace the final commuting path, and generate a tracing report.

[0067] Based on the user selection, the final commuting path and the corresponding score are obtained, the final commuting path is represented as the user's final selected commuting path, the corresponding commuting score vector is obtained, the sensitivity of the commuting score vector is obtained by partial derivation of the corresponding commuting score vector, and the commuting influence matrix is obtained.

[0068] For example, the user selects the final commuting path and the corresponding score: path 1, 88 points, the commuting score vector of path 1 is (90, 85, 80), the sensitivity of the commuting score vector is obtained by partial derivation of the commuting score vector, and the commuting influence matrix [time contribution 72%, energy consumption contribution 20%, abnormal event contribution 10%] is obtained.

[0069] Based on the commuting influence matrix, the key feature screening is performed to obtain the commuting influence list.

[0070] For example, by performing key feature screening on the above commuting influence matrix, the time contribution is screened as a key feature, the time contribution includes the waiting time of red light and the time of intersection congestion, the waiting time of red light and the time of intersection congestion are obtained, and the commuting influence list [intersection A red light extension waiting time 18%, xx intersection congestion time 9%] is generated.

[0071] Based on the commuting influence list, a dynamic decision tree is constructed, and a traceability report is generated based on the dynamic decision tree.

[0072] For example, by constructing a dynamic decision tree based on the above commuting influence list, a traceability report [xx intersection green light + 24, xxx section smooth + 12] is generated based on the dynamic decision tree.

[0073] Step S6, based on the user feedback and the traceability report, the dynamic traffic flow is optimized, and returns to step S2.

[0074] Specifically, the user feedback and the traceability report are analyzed to optimize the dynamic traffic flow, and the optimized dynamic traffic flow is returned to step S2.

[0075] Further, the factors affecting the score are found by the traceability method, the influencing factors are analyzed, the analysis results are obtained, and the dynamic traffic flow is optimized according to the analysis results to achieve the purpose of optimizing the evaluation.

[0076] Figure 2 A schematic diagram of a commuting network evaluation system based on artificial intelligence is shown, which can realize the idea of the present application.

[0077] Specifically, a commuting network evaluation system based on artificial intelligence comprises: Acquisition module 101, which is used to acquire the commuting network, obtain real-time traffic flow data, traffic light phase sequence and road topology metadata, and obtain the user's final commuting path; A generation module 102 is configured to map the commuting network to a Lie group space, generating a smooth manifold of the commuting network in the Lie group space. The agent forest model generates a commuting score vector based on the starting point, the end point, and the dynamic traffic manifold, and generates a score relationship equation based on the user's historical preferences. A construction module 103 is configured to construct a dynamic traffic manifold based on real-time traffic flow data, traffic light phase sequence, road topology metadata, and the smooth manifold, and to construct a multi-agent forest model; An integration module 104 integrates the commuting score vectors based on the score relationship equation to obtain a commuting path score set, and marks the three commuting paths with the highest scores in the path score set and pushes them to the user; A tracing module 105 is used to trace the final commuting path and generate a tracing report; The optimization module 106 optimizes the dynamic traffic flow shape based on user feedback and traceability reports.

[0078] The specific usage and function of this embodiment are described below: First, the commuting network is obtained through step S1, and the commuting network is mapped to the Lie group space to generate a smooth manifold of the commuting network on the Lie group space. Then, according to step S2, real-time traffic flow data, traffic light phase sequence and road topology metadata are obtained. A dynamic traffic manifold is constructed through the real-time traffic flow data, traffic light phase sequence, road topology metadata and the smooth manifold. Then, a multi-agent forest model is constructed through step S3 to obtain the commuting start point and commuting end point input by the user. The agent forest model generates a commuting score vector based on the commuting start point, commuting end point and dynamic traffic manifold. The commuting score vector includes travel time score, energy consumption score, and vehicle weight. The source consumption score and the probability score of the impact of abnormal events are then generated through step S4. The commuting score vectors are integrated according to the scoring relationship equation to obtain the commuting path score set, and the three commuting paths with the highest scores in the path score set are marked and pushed to the user. By mapping the commuting network to the Lie group space, the travel time and energy consumption are converted into commuting differential equations and a dynamic traffic manifold is constructed. The paths generated by the dynamic traffic manifold are scored through the multi-agent forest model, taking various commuting factors into consideration to make the final commuting score more accurate. At the same time, the user's historical data is added to better meet the user's needs.

[0079] The user's final commuting path is obtained through step S5, and the final commuting path is traced back, and a traceability report is generated. Finally, the dynamic traffic manifold is optimized according to step S6, and the process returns to step S2. The factors affecting the score are found through the traceability method, and the influencing factors are analyzed to obtain the analysis results. The dynamic traffic manifold is optimized based on the analysis results to achieve the purpose of optimization evaluation.

[0080] In addition, an embodiment of the present invention further provides an electronic device, including: At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method proposed in the first embodiment of the present invention.

[0081] The following is a detailed introduction to the various components of electronic equipment: The term "processor" is the control center of an electronic device and can be a single processor or a collective term for multiple processing elements. For example, the processor can be one or more central processing units (CPUs), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the first embodiment of the present invention, such as one or more digital signal processors (DSPs) or one or more field programmable gate arrays (FPGAs).

[0082] The processor can execute various functions of the electronic device by running or executing software programs stored in the memory and calling data stored in the memory.

[0083] The memory is used to store the software program for executing the solution of the present invention, and the execution is controlled by the processor. The specific implementation method can refer to the above method embodiment and will not be repeated here.

[0084] The memory may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, a random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only (CD-ROM), or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. The memory may be integrated with the processor or exist independently and be coupled to the processor via an interface circuit of the electronic device, and this is not specifically limited in the embodiments of the present invention.

[0085] The above embodiments can be implemented in whole or in part via software, hardware (e.g., circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. When loaded or executed on a computer, the processes or functions described in accordance with the embodiments of the present invention are fully or partially performed. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another. For example, the computer instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wireless communication (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium accessible by a computer, or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0086] It should be understood that the term "and / or" as used herein simply describes an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent the existence of A alone, the existence of both A and B, or the existence of B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the related objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0087] It should be understood that in the embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0088] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A commuting network evaluation method based on artificial intelligence, characterized in that: The method comprises: S1: Obtain the commuting network and map it to the Lie group space to generate a smooth manifold of the commuting network on the Lie group space; S2: acquiring real-time traffic flow data, traffic light phase sequence, and road topology metadata, and constructing a dynamic traffic manifold based on the real-time traffic flow data, traffic light phase sequence, road topology metadata, and the smooth manifold; S3: Build a multi-agent forest model to obtain the commuting start and end points input by the user. The agent forest model generates a commuting score vector based on the commuting start and end points and the dynamic traffic manifold. The commuting score vector includes a travel time score, an energy consumption score, and a probability score of the impact of abnormal events. S4: Generate a rating relationship equation based on the user's historical preferences, integrate the commuting rating vectors based on the rating relationship equation, obtain a commuting path rating set, and mark the three commuting paths with the highest ratings in the commuting path rating set and push them to the user; S5: Obtain the user's final commuting route, trace the final commuting route, and generate a traceability report; S6: Optimize the dynamic traffic manifold based on user feedback and traceability reports, and return to step S2.

2. The commuting network evaluation method based on artificial intelligence according to claim 1, characterized in that: The method of obtaining the commuting network and mapping the commuting network to the Lie group space to generate a smooth manifold of the commuting network on the Lie group space includes: The commuting network includes commuting nodes and commuting routes. The commuting nodes and commuting routes are mapped to the Lie group space. The commuting nodes correspond to elements in the Lie group space, and the commuting routes correspond to group action operators in the Lie group space. A smooth manifold is generated based on the commuting nodes and commuting routes.

3. The commuting network evaluation method based on artificial intelligence according to claim 1, characterized in that: The acquiring of real-time traffic flow data, traffic light phase sequence, and road topology metadata, and constructing a dynamic traffic manifold based on the real-time traffic flow data, traffic light phase sequence, road topology metadata, and the smooth manifold includes: Acquire real-time traffic flow data with a sampling frequency of once per second, including vehicle location and motion status; Obtaining a traffic light phase sequence with a time resolution of 0.1 seconds, wherein the traffic light phase sequence includes the change time and state of the traffic signal; Obtaining road topology metadata, wherein the road topology metadata includes the number of lanes, slope, and curvature radius of the road; A commuting differential equation is constructed based on real-time traffic flow data, traffic light phase sequence, and road topology metadata, and a dynamic traffic manifold is constructed based on the commuting differential equation and the smooth manifold.

4. The commuting network evaluation method based on artificial intelligence according to claim 3, characterized in that: The method of constructing a commuting differential equation based on real-time traffic flow data, traffic light phase sequence, and road topology metadata, and constructing a dynamic traffic manifold based on the commuting differential equation and the smooth manifold, includes: Obtaining a commuting position of a vehicle on a smooth manifold, obtaining a traffic flow tensor based on real-time traffic flow data and road topology metadata, wherein the traffic flow tensor represents a flow state of traffic flow, and composing the position and the traffic flow tensor into a commuting vector; Obtain traffic signal changes based on traffic light phase timing; The commuting vector is differentiated over the changing time to obtain a commuting differential equation, and a dynamic traffic manifold is constructed with the smooth manifold. The changing time is represented by the time corresponding to before and after the traffic signal changes.

5. The commuting network evaluation method based on artificial intelligence according to claim 1, characterized in that: The multi-agent forest model is constructed to obtain the commuting start point and commuting destination input by the user. The agent forest model generates a commuting score vector based on the commuting start point, commuting destination, and dynamic traffic manifold. The commuting score vector includes a travel time score, an energy consumption score, and an abnormal event impact probability score, including: The user enters the commuting start and end points into the mobile user terminal. The mobile user terminal analyzes the commuting start and end points and the dynamic traffic manifold based on the multi-agent forest model to generate a commuting score vector. The multi-agent forest model includes a time agent, an energy consumption agent and a risk agent.

6. The commuting network evaluation method based on artificial intelligence according to claim 5, characterized in that: The method further comprises: The time agent predicts the travel time based on the LSTM to obtain the predicted travel time, and scores the predicted travel time to obtain a travel time score; The energy consumption agent solves energy based on a dynamic equation to obtain commuting energy consumption, and scores the energy consumption to obtain an energy consumption score; The risk agent analyzes the impact of abnormal events based on the Bayesian network to obtain the impact probability of the abnormal events, and scores the impact probability of the abnormal events to obtain the impact probability score of the abnormal events.

7. The commuting network evaluation method based on artificial intelligence according to claim 1, characterized in that: The method generates a rating relationship equation based on the user's historical preferences, integrates the commuting rating vectors based on the rating relationship equation, obtains a commuting path rating set, and marks the three commuting paths with the highest ratings in the commuting path rating set and pushes them to the user, including: Obtaining user historical preferences based on historical data, evaluating and proportioning travel time scores, energy consumption scores, and abnormal event impact probability scores based on user historical preferences, and generating a scoring relationship equation; The travel time score, energy consumption score, and abnormal event impact probability score are input into the scoring relationship equation to generate a commuting path score set. The three commuting paths with the highest scores in the commuting path score set are extracted and marked, and the three commuting paths with the highest scores are pushed to the user.

8. The commuting network evaluation method based on artificial intelligence according to claim 7, characterized in that: The extracting and marking the three commuting routes with the highest scores from the commuting route score set, and pushing the three commuting routes with the highest scores to the user, includes: The marks include time marks, energy consumption marks and comprehensive marks; If the travel time score corresponding to the extracted commuting route is the highest, the extracted commuting route is marked as having the shortest travel time; If the energy consumption score corresponding to the extracted commuting route is the highest, the extracted commuting route is marked as having the lowest energy consumption; If the extracted commuting route has the highest score, the extracted commuting route is marked as having the highest comprehensive score; The three commuting routes and their corresponding scores after being marked are pushed to the user.

9. The commuting network evaluation method based on artificial intelligence according to claim 1, characterized in that: The obtaining of the user's final commuting route, tracing the final commuting route, and generating a tracing report include: Obtaining a final commuting route and a corresponding score based on the user's selection, wherein the final commuting route is represented as the commuting route ultimately selected by the user, obtaining a commuting score vector corresponding to the score, and performing a partial derivative of the corresponding commuting score vector to obtain a sensitivity of the commuting score vector, thereby obtaining a commuting impact matrix; Screen key features based on the commuting impact matrix to obtain a commuting impact list; Build a dynamic decision tree based on the commuting impact list, and generate a traceability report based on the dynamic decision tree.

10. An artificial intelligence-based commuting network evaluation system, characterized in that: include: An acquisition module, which is used to acquire the commuting network, obtain real-time traffic flow data, traffic light phase timing, and road topology metadata, and obtain the user's final commuting path; A generation module is configured to map the commuting network to a Lie group space, generating a smooth manifold of the commuting network in the Lie group space. The agent forest model generates a commuting score vector based on the starting point, the end point, and the dynamic traffic manifold, and generates a score relationship equation based on the user's historical preferences. A construction module, wherein the construction module constructs a dynamic traffic manifold based on real-time traffic flow data, traffic light phase time series, road topology metadata, and the smooth manifold, and constructs a multi-agent forest model; An integration module, which integrates the commuting score vectors based on the score relationship equation to obtain a commuting path score set, and marks the three commuting paths with the highest scores in the commuting path score set and pushes them to the user; A tracing module, which is used to trace the final commuting path and generate a tracing report; An optimization module optimizes the dynamic traffic flow shape based on user feedback and traceability reports.

Images