Method and system for calculating cruising path of rented vehicle based on expected cruising mileage

Abstract

The invention discloses a rental vehicle cruise path calculation method and system based on expected cruise mileage, and the method comprises the steps: determining a road section according with a driving path, correcting a preliminary track point, and obtaining a corresponding road section number; according to the taxi track data of the road sections corresponding to the taxi track points, determining the taxi transportation force and demand quantity of different road sections; predicting the taxi transport capacity and demand quantity of each road section in the future and the passenger carrying probability of each road section; evaluating the time sequence period rule degree of the transportation capacity and the demand quantity of each road section, and correcting the passenger carrying probability prediction value of each road section; determining a candidate path set according to the information of all the road sections, calculating the expected cruise mileage of each candidate path, and determining a recommended path; and updating the transport capacity data of the road section, and re-determining the recommended path. According to the invention, the cruising efficiency of the taxi is effectively improved, the resource configuration is optimized, the overall operation level is improved, the operation cost is reduced, and the travel experience and satisfaction of passengers are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation systems, and in particular to a method and system for calculating a cruising path of a rental vehicle based on expected cruising mileage. Background Art

[0002] In the modern urban transportation system, taxis, as an important means of public transportation, play a vital role. However, with the acceleration of urbanization and the increase in population, the taxi industry faces a series of challenges that not only affect operational efficiency and economic benefits, but also affect passengers' travel experience and satisfaction. In taxi operations, the problem of matching capacity with demand is one of the key factors affecting operational efficiency and passenger satisfaction. During peak hours, passenger demand in certain specific areas surges, while other areas may have a high taxi idling rate, resulting in resource waste and imbalance between supply and demand. This situation causes passengers in high-demand areas to wait for a long time and find it difficult to get a taxi, while taxis in low-demand areas face the problem of empty driving, which increases operating costs and driver fatigue. Existing taxi dispatch systems usually dispatch based on historical data and experience, which is difficult to adjust in real time and cannot effectively respond to sudden changes and dynamic demands. Due to the lack of accurate prediction and rapid response mechanism for real-time demand, taxi companies find it difficult to quickly dispatch sufficient capacity to areas with concentrated demand during high-demand periods, and they are also unable to effectively reduce the empty driving rate when demand is low. Optimizing the matching of capacity and demand, real-time monitoring and evaluation of the supply and demand of each road section, rationally allocating taxi resources, and improving vehicle utilization and passenger satisfaction are among the problems that need to be solved in the current taxi industry. The route planning problem of taxis should not be ignored either. Traditional route planning methods are mainly based on the principle of the shortest path or the shortest time, and fail to fully consider the dynamic changes of real-time traffic conditions and passenger demand. This simple route planning method often causes taxis to experience a long period of empty mileage during cruising, reducing cruising efficiency. For example, when cruising, drivers may choose a route with traffic congestion, resulting in long periods of low-speed driving, wasting fuel and time, and increasing operating costs. In addition, failure to reasonably plan the cruising route will also increase driver fatigue and affect driving safety. Existing route planning systems usually lack dynamic adjustment capabilities and are difficult to adapt to the complex and changing urban traffic environment. When faced with emergencies, traffic control or road construction, drivers often cannot get timely and effective route adjustment suggestions, resulting in increased driving time and distance, affecting operational efficiency and service quality. To solve these problems, a dynamically adjustable route planning method is needed to intelligently select the optimal route based on real-time traffic data, passenger demand data and historical data, reduce empty mileage and improve cruising efficiency. At the same time, it is also necessary to establish an efficient feedback mechanism that can obtain the driver's location information and driving status in real time, adjust the recommended route in time, and ensure that the taxi can reach the passenger's location and destination by the optimal route. Therefore, there are problems in the existing taxi operation that do not match the capacity with the demand and the unreasonable route planning. These problems not only affect the operating efficiency and service quality of taxis, but also increase the operating costs of drivers and the inconvenience of passengers to a certain extent.Therefore, there is an urgent need for a method that can effectively solve these problems in order to improve the cruising efficiency of taxis, optimize resource allocation, and improve the overall operation level. Summary of the invention

[0003] The present invention solves the problems existing in the above-mentioned prior art and provides a method for calculating the cruising path of a rental vehicle based on expected cruising mileage, which mainly includes:

[0004] The initial trajectory of the taxi is obtained based on the real-time satellite positioning data of the taxi, the distance and angle difference between the taxi trajectory point and the road segment are calculated, the road segment that meets the driving path is determined and the initial trajectory point is corrected, and the corresponding road segment number is obtained;

[0005] According to the taxi trajectory data of the road section corresponding to the taxi trajectory point, the taxi transport capacity and demand of different road sections are determined;

[0006] Based on the historical taxi capacity and demand data of different road sections, the future taxi capacity and demand of each road section are predicted, and based on the road section passenger probability prediction formula, the passenger probability of each road section is predicted;

[0007] According to the historical data of taxi capacity and demand, the degree of temporal regularity of the capacity and demand of each road section is evaluated, and the predicted value of passenger probability of each road section is revised based on the degree of temporal regularity;

[0008] Determine the candidate path set based on all road segment information, calculate the conditional passenger load probability of each road segment in the candidate path based on the conditional passenger load probability formula, and use the path expected cruising mileage calculation formula to calculate the expected cruising mileage of each candidate path, and determine the path with the shortest expected cruising mileage as the recommended path;

[0009] According to the real-time location information of the taxi, it is determined whether the driver is driving according to the recommended route, and the road segment capacity data is updated. Based on the updated road segment capacity data, the path with the shortest expected cruising mileage is re-determined as the recommended route.

[0010] Further, the method of obtaining a preliminary trajectory of the taxi according to the real-time satellite positioning data of the taxi, calculating the distance and angle difference between the taxi trajectory point and the road segment, determining the road segment that meets the driving path and correcting the preliminary trajectory point, and obtaining the corresponding road segment number includes:

[0011] According to the real-time satellite positioning data and timestamp obtained by the GPS device of the taxi, a set of positioning points including longitude and latitude is obtained to obtain the preliminary trajectory of the taxi; through the map data, the potential driving road segments of the taxi are obtained, including the coordinates, direction and connectivity information of the road segments; according to the preliminary trajectory points of the taxi, the Euclidean distance between each trajectory point and the potential driving road segment is calculated, and the road segments with a distance less than a preset distance threshold are selected as candidate road segments; according to the difference between the longitude and latitude of adjacent preliminary trajectory points of the taxi, the angle calculation is performed to determine the movement direction of the taxi trajectory point; the angle difference between the direction of the candidate road segment and the movement direction of the taxi trajectory point is calculated to determine the candidate road segment with an angle difference less than a preset angle difference threshold; through the map data, according to the connectivity of the candidate road segment with an angle difference less than the preset angle difference threshold, the road segment that meets the actual driving path is determined; according to the road segment that meets the actual driving path, the coordinates of the preliminary trajectory points of the taxi are adjusted, and the coordinates of the preliminary trajectory points of the taxi with an offset are corrected to the road segment of the road where the taxi actually travels, the corrected taxi trajectory points are obtained, and the road segment number corresponding to each corrected taxi trajectory point is determined.

[0012] Further, the taxi transport capacity and demand of different road sections are determined according to the taxi trajectory data of the road section corresponding to the taxi trajectory point, including:

[0013] According to the road segment corresponding to the taxi trajectory point, with t minutes as the time granularity, obtain the taxi trajectory data set within the past preset time period, the taxi trajectory data includes time, road segment, license plate number and passenger status, the passenger status includes empty and with passengers; through the taxi trajectory data set within the past preset time period, obtain the taxi license plate number of the same road segment, remove duplicate license plate numbers, and determine the taxi transport capacity of each road segment every t minutes in the past preset time period, the taxi transport capacity is the number of empty taxis; obtain the passenger status corresponding to the taxi license plate number every t minutes in the past preset time period, judge the time point and road segment when the passenger status changes from empty to with passengers, obtain the license plate number set of passengers picked up on each road segment every t minutes, and determine the number of passengers picked up on each road segment every t minutes as the taxi demand.

[0014] Furthermore, the method predicts the future taxi capacity and demand of each road segment based on the historical taxi capacity and demand data of different road segments, and predicts the passenger carrying probability of each road segment based on the road segment passenger carrying probability prediction formula, including:

[0015] Obtain the historical taxi capacity and demand data of each road section, take t minutes as the time granularity, obtain the taxi capacity and demand data of every t minutes in the historical taxi capacity and demand data, and construct the time series data of each road section; according to the time series data of each road section, use the spatiotemporal graph convolutional neural network algorithm to train the model and construct a taxi capacity and demand prediction model; according to the taxi capacity and demand data of each road section every t minutes in the past preset time period, use the taxi capacity and demand prediction model to predict the capacity and demand of each road section in the next t minutes; according to the road section passenger probability prediction formula Predict the passenger carrying probability of each road segment, where P represents the predicted value of the road segment passenger carrying probability, D is the predicted value of the road segment demand, and S is the predicted value of the road segment transport capacity.

[0016] Furthermore, the method of evaluating the time-series regularity of the taxi capacity and demand of each road segment based on the historical data of taxi capacity and demand, and revising the predicted value of the passenger probability of each road segment based on the time-series regularity, includes:

[0017] Obtain the historical data of taxi transport capacity and demand on each road section to obtain a complete time series data set; use the time series period regularity degree formula Evaluate the degree of regularity of the time series of transport capacity and demand of each road section, among which, represents the taxi demand sequence of road segment j in the i-th time period, represents the taxi capacity time series of road segment j in the i-th time period, represents the taxi demand sequence of road segment j in the i+1th time period, represents the taxi capacity time series of road segment j in the i+1th time period, and m represents the number of time periods covered by historical data; according to the regularity of the time series cycle of the capacity and demand of each road segment, the reliability correction coefficient β of the passenger probability prediction of each road segment is determined j ,in, According to the predicted value of passenger probability of each road section, the credibility correction coefficient and credibility correction formula are used The predicted value of the passenger-carrying probability of each road section is corrected to determine the corrected taxi passenger-carrying probability of each road section in the entire road network in the next t minutes; the corrected taxi passenger-carrying probability of each road section in the entire road network in the next t minutes is stored in a memory-based key-value database, where the key value is the road section code and the value value is the corrected predicted value of the taxi passenger-carrying probability of the corresponding road section in the next t minutes The predicted data of the road segment passenger probability in the future t minutes in the cached data is updated every t minutes. Further, the candidate path set is determined according to all the road segment information, the conditional passenger probability of each road segment in the candidate path is calculated based on the conditional passenger probability formula, and the expected cruising mileage calculation formula of the path is used to calculate the expected cruising mileage of each candidate path, and the path with the shortest expected cruising mileage is determined as the recommended path, including:

[0018] According to the route recommendation request initiated by the driver, the starting and ending positions are obtained, and the latitude and longitude coordinates of the starting and ending points are determined; map data is obtained, all road segment information is extracted, and a depth-first path search algorithm is used to obtain all paths with a length of K kilometers starting from the request initiation point, and a candidate path set G is obtained; through the key-value database, the predicted values ​​of the future t-minute passenger probability corresponding to all road segments included in all candidate paths are queried; a candidate path in the candidate path set G is defined as G u =(l 1 ,l 2 ,…,l n ), the passenger carrying probability of each road segment of the candidate path is Where n represents the number of road segments included in the candidate path; according to the conditional passenger probability formula Calculate the conditional passenger loading probability of a single road segment, where represents the cumulative passenger carrying probability, and the cumulative passenger carrying probability is calculated as follows: Get the length of each road segment of the candidate path through the formula Calculate the cumulative cruising mileage of each road segment of the candidate route Where L i For road segment l i length; based on the accumulated cruising mileage and the predicted value of passenger probability The expected cruising mileage calculation formula is adopted Calculate the expected cruising mileage of each candidate path; compare the expected cruising mileages of different candidate paths, and determine the path with the shortest expected cruising mileage as the recommended path with the highest cruising efficiency.

[0019] Further, judging whether the driver is traveling along the recommended route based on the real-time location information of the taxi, and updating the road segment capacity data, and re-determining the route with the shortest expected cruising mileage as the recommended route based on the updated road segment capacity data, includes:

[0020] Obtain the longitude and latitude sequence information of the recommended path, and transmit the longitude and latitude sequence information of the recommended path to the background server through the API interface; use the map rendering API to load the longitude and latitude data of the recommended path, render the recommended path on the map, and obtain a visualized recommended path route; obtain the driver's real-time location information, compare the driver's driving route with the recommended path, and determine whether the driver is driving according to the recommended path; if the driver is driving according to the recommended path, traverse all road segments that the driver drives according to the recommended path, accumulate the road segment capacity of each road segment by 1, and obtain the updated road segment capacity data; according to the updated road segment capacity data, recalculate the updated passenger probability prediction value, and adjust the passenger probability of the corresponding road segment in the key-value database; according to the updated passenger probability prediction value, re-determine the path with the shortest expected cruising mileage as the recommended path with the highest cruising efficiency.

[0021] The second aspect of the present invention provides a system for calculating a cruising path of a rental vehicle based on expected cruising mileage, which mainly includes:

[0022] The satellite positioning data real-time map matching module is used to determine the road segment that matches the taxi's driving path and correct the driving trajectory, and obtain the corresponding road segment number;

[0023] The road segment-level taxi capacity demand calculation module is used to determine the taxi capacity and demand of different road segments according to the taxi trajectory data of the road segment corresponding to the taxi trajectory point;

[0024] The road segment-level taxi passenger probability prediction module is used to predict the future taxi capacity, demand and passenger probability of each road segment based on the historical taxi capacity and demand data of different road segments;

[0025] The taxi passenger carrying probability credibility correction module is used to evaluate the temporal periodic regularity of the transport capacity and demand of each road section, and to correct the predicted value of the passenger carrying probability of each road section based on the temporal periodic regularity;

[0026] A real-time cruise route recommendation module is used to determine a candidate route set based on all road segment information, calculate the expected cruise mileage of each candidate route, and determine the route with the shortest expected cruise mileage as the recommended route;

[0027] The cruising route update module is used to determine whether the driver drives according to the recommended route, update the road segment capacity data, and re-determine the route with the shortest expected cruising mileage as the recommended route.

[0028] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0029] The present invention provides a method and system for calculating the cruising path of a rental vehicle based on expected cruising mileage. The present invention accurately obtains the real-time satellite positioning data of the taxi, calculates the distance and angle difference between the trajectory point and the road segment, and corrects the initial trajectory point, thereby ensuring that each trajectory point is accurately mapped to the number of the road segment actually traveled, effectively reducing the positioning error and improving the accuracy of path planning. By analyzing the trajectory data of the road segment corresponding to the taxi trajectory point, the supply and demand of each road segment is monitored and evaluated in real time, and taxi resources are reasonably allocated to improve vehicle utilization and passenger satisfaction. The present invention can also predict the future capacity and demand of each road segment based on historical data, understand the future supply and demand changes in advance, provide data support for path planning, and avoid the occurrence of supply and demand mismatch. By evaluating the periodic change characteristics of the capacity and demand of each road segment, the prediction result is corrected to improve the accuracy and reliability of the prediction. When determining the candidate path, the present invention can calculate the expected cruising mileage of each path, and determine the recommended path with the highest cruising efficiency, reduce the empty mileage, and improve the operation efficiency. In addition, by obtaining the driver's location information in real time, determining whether the driver is driving along the recommended route, and dynamically updating the road segment capacity data and re-recommending the optimal route, the timeliness and effectiveness of the route recommendation are ensured. The present invention effectively improves the cruising efficiency of taxis, optimizes resource allocation, improves the overall operation level, reduces operating costs, and enhances passengers' travel experience and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A flowchart of a method for calculating a cruising path of a rental vehicle based on expected cruising mileage according to the present invention;

[0031] Figure 2 A flowchart of a system for calculating a cruising path of a rental vehicle based on expected cruising mileage according to the present invention;

[0032] Figure 3 A schematic diagram of a method for calculating a cruising path of a rental vehicle based on expected cruising mileage according to the present invention; DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0034] like Figure 1 and Figure 3 In this embodiment, a method for calculating a cruising path of a rental vehicle based on expected cruising mileage may specifically include:

[0035] Step S101, a satellite positioning data real-time map matching module is used to obtain the preliminary trajectory of the taxi based on the real-time satellite positioning data of the taxi, calculate the distance and angle difference between the taxi trajectory point and the road segment, determine the road segment that meets the driving path and correct the preliminary trajectory point, and obtain the corresponding road segment number.

[0036] According to the real-time satellite positioning data and timestamp obtained by the GPS device of the taxi, a set of positioning points including longitude and latitude is obtained to obtain the preliminary trajectory of the taxi. Through the map data, the potential driving road segments of the taxi are obtained, including the coordinates, direction and connectivity information of the road segments. According to the preliminary trajectory points of the taxi, the Euclidean distance between each trajectory point and the potential driving road segment is calculated, and the road segment with a distance less than the preset distance threshold is selected as the candidate road segment. According to the difference between the longitude and latitude of the adjacent preliminary trajectory points of the taxi, the angle is calculated to determine the movement direction of the taxi trajectory point. The angle difference between the direction of the candidate road segment and the direction of the taxi trajectory point is calculated to determine the candidate road segment with an angle difference less than the preset angle difference threshold. Through the map data, according to the connectivity of the candidate road segment with an angle difference less than the preset angle difference threshold, the road segment that meets the actual driving path is determined. According to the road segment that meets the actual driving path, the coordinates of the preliminary trajectory points of the taxi are adjusted, and the coordinates of the preliminary trajectory points of the taxi with an offset are corrected to the road segment of the road where the taxi actually travels, the corrected taxi trajectory points are obtained, and the road segment number corresponding to each corrected taxi trajectory point is determined.

[0037] For example, during a period of time on June 1, 2020, the GPS data of a taxi with a license plate number of Yue A8B70** at 05:00:28 is 113.32334 east longitude and 23.13161 north latitude. The coordinate information of the potential driving road segment obtained through the map data is that the starting longitude and latitude of road segment 1 is 113.32230 east longitude and 23.13130 ​​north latitude, and the end longitude and latitude are 113.32400 east longitude and 23.13200 north latitude, and the direction is 30 degrees north by east. The starting longitude and latitude of road segment 2 is 113.32300 east longitude and 23.13150 north latitude, and the end longitude and latitude are 113.32450 east longitude and 23.13250 north latitude, and the direction is 45 degrees north by east. The Euclidean distances between the preliminary trajectory point and road segment 1 and road segment 2 are 2 meters and 1.5 meters respectively, both less than the preset distance threshold of 2.5 meters, so road segment 1 and road segment 2 are both selected as candidate road segments. According to the difference in longitude and latitude of adjacent trajectory points, the movement direction of the taxi trajectory point is calculated to be 40 degrees north by east. The angle difference between the direction of the candidate road segment and the direction of the trajectory point is calculated, and it is found that the angle difference of road segment 2 is 5 degrees, which is less than the preset angle difference threshold of 10 degrees, so road segment 2 is selected as the candidate road segment. Through the map data, further according to the connectivity of the road segment, it is determined that the road segment that meets the actual driving path is road segment 2. According to the road segment that meets the actual driving path, the coordinates of the preliminary trajectory point of the taxi are adjusted, and the offset trajectory points are corrected to road segment 2, and the coordinates of the corrected taxi trajectory point are obtained as 113.32301 east longitude and 23.13153 north latitude. According to the map data, the road segment number corresponding to the corrected trajectory point is determined to be number 157840.

[0038] Step S102, a road segment-level taxi capacity demand calculation module is used to determine the taxi capacity and demand of different road segments according to the taxi trajectory data of the road segment corresponding to the taxi trajectory point.

[0039] According to the road segment corresponding to the taxi trajectory point, with t minutes as the time granularity, obtain the taxi trajectory data set within the past preset time period. The taxi trajectory data includes time, road segment, license plate number and passenger status. The passenger status includes empty and with passengers. Through the taxi trajectory data set within the past preset time period, obtain the taxi license plate number of the same road segment, remove duplicate license plate numbers, and determine the taxi transport capacity of each road segment every t minutes in the past preset time period. The taxi transport capacity is the number of empty taxis. Obtain the passenger status corresponding to the taxi license plate number every t minutes in the past preset time period, determine the time point and road segment when the passenger status changes from empty to with passengers, obtain the license plate number set of passengers on each road segment every t minutes, and determine the number of passengers picked up on each road segment every t minutes as the taxi demand.

[0040] For example, within the two hours from 15:00 to 17:00 on June 1, 2020, a set of taxi trajectory data on a certain road section is obtained with a time granularity of 5 minutes. The trajectory data includes time, road section, license plate number and passenger status, and the passenger status includes empty and loaded. In the time period from 15:00 to 15:05, through the acquired trajectory data set, it is obtained that during this period, taxis with license plate numbers of Guangdong A8B70**, Guangdong A8B72** and Guangdong A8B75** are traveling on this road section. The passenger status of the license plate number Guangdong A8B70** is empty, the passenger status of the license plate number Guangdong A8B72** is loaded, and the passenger status of the license plate number Guangdong A8B75** is empty. By processing the trajectory data set and removing duplicate license plate numbers, it is determined that the taxi capacity of each vehicle is the number of empty taxis during the time period from 15:00 to 15:05, that is, the taxi capacity of this road section is 2 empty taxis during the time period from 15:00 to 15:05. During the time period from 15:05 to 15:10, the passenger status of the license plate number Guangdong A8B70** changes to passenger, the license plate number Guangdong A8B75** is still empty, and the newly added license plate number Guangdong A8B78** is empty. After removing duplicate license plate numbers, it is determined that the taxi capacity of this road section is 2 empty taxis during the time period from 15:05 to 15:10. By obtaining the passenger status corresponding to the taxi license plate number every t minutes in the past preset time period, the time point and road section when the passenger status changes from empty to passenger are determined. During the time period from 15:05 to 15:10, the passenger status of the license plate number Yue A8B70** changed from empty to loaded, so the number of pick-ups during this time period was 1. The set of license plate numbers for pick-ups on this road section during the time period from 15:05 to 15:10 is Yue A8B70**, and the number of pick-ups on this road section during this time period is determined to be 1, which is the taxi demand.

[0041] Step S103, a road segment level taxi passenger carrying probability prediction module is used to predict the future taxi carrying capacity and demand of each road segment based on the historical taxi carrying capacity and demand data of different road segments, and to predict the passenger carrying probability of each road segment based on the road segment passenger carrying probability prediction formula.

[0042] Obtain the historical taxi capacity and demand data for each road section. With t minutes as the time granularity, obtain the taxi capacity and demand data for every t minutes in the historical taxi capacity and demand data to construct the time series data for each road section. Based on the time series data of each road section, use the spatiotemporal graph convolutional neural network algorithm to train the model and construct a taxi capacity and demand prediction model. Based on the taxi capacity and demand data for each road section every t minutes in the past preset time period, use the taxi capacity and demand prediction model to predict the capacity and demand of each road section in the next t minutes. According to the road section passenger probability prediction formula Predict the passenger carrying probability of each road segment, where P represents the predicted value of the road segment passenger carrying probability, D is the predicted value of the road segment demand, and S is the predicted value of the road segment transport capacity.

[0043] Exemplarily, the prediction is made with a time of 5 minutes. When the time reaches 17:00 on June 1, 2020, the taxi capacity and demand of the road section in the past 2 hours are obtained, that is, the historical taxi capacity and demand data on a certain section of road are obtained with a time granularity of 5 minutes within the 2 hours from 15:00 to 17:00 on June 1, 2020. It is found that in the time period from 15:00 to 15:05, the capacity of the road section is 3 empty taxis and the demand is 1 passenger. In the time period from 15:05 to 15:10, the capacity of the road section is 4 empty taxis and the demand is 2 passengers. And so on, a time series data of every 5 minutes is constructed. Based on these time series data, the spatiotemporal graph convolutional neural network algorithm is used for model training to construct a prediction model for taxi capacity and demand. Using the taxi capacity and demand data of each road section every 5 minutes in the past 2 hours, the trained spatiotemporal graph convolutional neural network prediction model is used to predict the capacity and demand of each road section in the next 5 minutes. If it is predicted that from 17:00 to 17:05, the capacity of the road section is 5 empty taxis and the demand is 3 passengers. According to the road section passenger probability prediction formula Predict the passenger carrying probability of each road segment, where P represents the predicted value of the road segment passenger carrying probability, D is the predicted value of the road segment demand, and S is the predicted value of the road segment transportation capacity. For the time period from 17:00 to 17:05, the predicted value of the road segment demand is 3, and the predicted value of the transportation capacity is 5, so the predicted value of the passenger carrying probability is P = 3 / 5 = 0.6.

[0044] Step S104, a taxi passenger carrying probability credibility correction module is used to evaluate the temporal periodic regularity of the taxi carrying capacity and demand of each road section based on the historical data of taxi carrying capacity and demand, and to correct the predicted value of the passenger carrying probability of each road section based on the temporal periodic regularity.

[0045] Obtain the historical data of taxi capacity and demand on each road section to obtain a complete time series data set. Evaluate the degree of regularity of the time series of transport capacity and demand of each road section, among which, represents the taxi demand sequence of road segment j in the i-th time period, represents the taxi capacity time series of road segment j in the i-th time period, represents the taxi demand sequence of road segment j in the i+1th time period, represents the taxi capacity time series of road segment j in the i+1th time period, and m represents the number of time periods covered by historical data. According to the regularity of the time series cycle of the capacity and demand of each road segment, the reliability correction coefficient β of the passenger probability prediction of each road segment is determined. j ,in, According to the predicted value of passenger probability of each road section, the credibility correction coefficient and credibility correction formula are used The predicted value of the passenger carrying probability of each road segment is corrected to determine the corrected taxi passenger carrying probability of each road segment in the entire road network in the next t minutes. The corrected taxi passenger carrying probability of each road segment in the entire road network in the next t minutes is stored in a memory-based key-value database, where the key value is the road segment code and the value value is the corrected predicted value of the taxi passenger carrying probability of the corresponding road segment in the next t minutes. And update the predicted data of the passenger probability of the road segment in the future t minutes in the cached data every t minutes.

[0046] For example, the historical taxi capacity and demand data on a certain road section are obtained with a time granularity of 5 minutes. In the time period from 15:00 to 15:05, the capacity of the road section is 3 empty taxis and the demand is 1 passenger. In the time period from 15:05 to 15:10, the capacity of the road section is 4 empty taxis and the demand is 2 passengers. And so on, a time series data of every 5 minutes is constructed. The time series period regularity degree formula is used Get the temporal periodic regularity degree α of each road segment j is 0.4, where represents the taxi demand sequence of road segment j in the i-th time period, represents the taxi capacity time series of road segment j in the i-th time period, represents the taxi demand sequence of road segment j in the i+1th time period, represents the taxi capacity time series of road segment j in the i+1th time period, and m represents the number of time periods covered by historical data. According to the regularity of the time series cycle of the transport capacity and demand of each road segment, the reliability correction coefficient formula based on the passenger probability prediction is Determine the reliability correction coefficient β of passenger load probability prediction for each road section j If the predicted value P of the passenger probability of a road section in the next 5 minutes is 0.6, the credibility correction coefficient and credibility correction formula are used. The corrected predicted passenger probability is The corrected taxi passenger probability for each road section in the entire road network in the next 5 minutes is stored in a memory-based key-value database. The key value in the database is set to the road section code 157840, and the value is set to the corrected taxi passenger probability prediction value of 0.24 for the corresponding road section in the next 5 minutes. The next 5 minutes road section passenger probability prediction data in the cached data is updated every 5 minutes.

[0047] Step S105, a cruising route real-time recommendation module is used to determine a candidate route set based on all road segment information, calculate the conditional passenger carrying probability of each road segment in the candidate route based on the conditional passenger carrying probability formula, and use the path expected cruising mileage calculation formula to calculate the expected cruising mileage of each candidate route, and determine the route with the shortest expected cruising mileage as the recommended route.

[0048] According to the route recommendation request initiated by the driver, the starting and ending positions are obtained, and the latitude and longitude coordinates of the starting and ending points are determined. Map data is obtained, all road segment information is extracted, and the depth-first path search algorithm is used to obtain all paths with the request initiation point as the starting point and a length of K kilometers, and the candidate path set G is obtained. Through the key-value database, the predicted passenger probability values ​​corresponding to all road segments included in all candidate paths in the future t minutes are queried. A candidate path in the candidate path set G is defined as G u =(l 1 ,l 2 ,…,l n ), the passenger carrying probability of each road segment of the candidate path is Where n represents the number of road segments included in the candidate path. According to the conditional passenger probability formula Calculate the conditional passenger loading probability of a single road segment, where represents the cumulative passenger carrying probability, and the cumulative passenger carrying probability is calculated as follows: Get the length of each road segment of the candidate path through the formula Calculate the cumulative cruising mileage of each road segment of the candidate route Where L i For road segment l i According to the accumulated cruising mileage and the predicted value of passenger probability The expected cruising mileage calculation formula is adopted That is, all possible cruising distances are multiplied by the corresponding probability of picking up passengers on each road section to calculate the expected cruising distance of each candidate route. According to statistics, the expected cruising distance reflects the average value of all possible cruising distances if the vehicle repeats the random event of driving along the route during the period many times. The expected cruising distance of each candidate route is quantified using the expected cruising distance. Compare the expected cruising distances of different candidate routes, and determine the route with the shortest expected cruising distance as the recommended route with the highest cruising efficiency. The driver can meet the passengers on the roadside with the shortest empty driving distance along this route.

[0049] For example, if the driver requests a route recommendation, the starting point is point A, the longitude and latitude coordinates are 113.32334 east longitude, 23.13161 north latitude, and the end point is point B, the longitude and latitude coordinates are 113.28101 east longitude, 23.13137 north latitude. Through the depth-first path search algorithm, all candidate paths from point A to point B with a length of 3 kilometers are obtained to obtain a candidate path set, including candidate path 1 and candidate path 2. Candidate path 1 includes road segment 1 to road segment 15, wherein the road segment passenger probability of road segment 1 is 22.0%, and the road segment length is 241 meters. The road segment passenger probability of road segment 2 is 14.0%, and the road segment length is 315 meters. The road segment passenger probability of road segment 3 is 41.0%, and the road segment length is 196 meters. The road segment passenger probability of road segment 4 is 38.0%, and the road segment length is 98 meters. The road segment passenger probability of road segment 5 is 6.0%, and the road segment length is 361 meters. The road segment probability of road segment 6 is 1.3%, and the road segment length is 415 meters. The road segment probability of road segment 7 is 19.0%, and the road segment length is 149 meters. The road segment probability of road segment 8 is 31.0%, and the road segment length is 399 meters. The road segment probability of road segment 9 is 20.0%, and the road segment length is 102 meters. The road segment probability of road segment 10 is 15.0%, and the road segment length is 301 meters. The road segment probability of road segment 11 is 1.0%, and the road segment length is 238 meters. The road segment probability of road segment 12 is 4%, and the road segment length is 512 meters. The road segment probability of road segment 13 is 8%, and the road segment length is 123 meters. The road segment probability of road segment 14 is 15%, and the road segment length is 210 meters. The road segment probability of road segment 15 is 824%, and the road segment length is 347 meters. According to the conditional passenger probability formula Calculate the conditional passenger probability of a single road segment, that is, the probability of meeting a passenger on this road segment without meeting a passenger before, where It represents the cumulative passenger carrying probability, that is, the total passenger carrying probability of vehicles traveling to this road section. The cumulative passenger carrying probability calculation formula is: Get the conditional passenger probability of road segment 1 Therefore, the cumulative passenger carrying probability of road segment 1 is Conditional passenger probability of road segment 2 Therefore, the cumulative passenger carrying probability of road segment 2 is The calculation methods of the conditional passenger carrying probability and cumulative passenger carrying probability of the remaining road segments are similar, and the conditional passenger carrying probability of road segment 3 is 0.275, and the cumulative passenger carrying probability of road segment 3 is 0.6042; the conditional passenger carrying probability of road segment 4 is 0.1504, and the cumulative passenger carrying probability of road segment 4 is 0.7546; the conditional passenger carrying probability of road segment 5 is 0.0196, and the cumulative passenger carrying probability of road segment 5 is 0.7743; the conditional passenger carrying probability of road segment 6 is 0.0113, and the cumulative passenger carrying probability of road segment 6 is 0.7855; the conditional passenger carrying probability of road segment 7 is 0.0407, and the cumulative passenger carrying probability of road segment 7 is 0.8263; the conditional passenger carrying probability of road segment 8 is 0.0539, and the cumulative passenger carrying probability of road segment 8 is 0.8801; and the conditional passenger carrying probability of road segment 9 is 0.0 The conditional passenger probability of road segment 9 is 0.024, the cumulative passenger probability of road segment 9 is 0.9041, the conditional passenger probability of road segment 10 is 0.0144, the cumulative passenger probability of road segment 10 is 0.9185, the conditional passenger probability of road segment 11 is 0.009, the cumulative passenger probability of road segment 11 is 0.9275, the conditional passenger probability of road segment 12 is 0.0029, the cumulative passenger probability of road segment 12 is 0.9304, the conditional passenger probability of road segment 13 is 0.0056, the cumulative passenger probability of road segment 13 is 0.9359, the conditional passenger probability of road segment 14 is 0.0096, the cumulative passenger probability of road segment 14 is 0.9455, the conditional passenger probability of road segment 15 is 0.0131, and the cumulative passenger probability of road segment 15 is 0.9586. Get the distance of each road segment of the candidate path through the formula Calculate the cumulative cruising mileage of each road segment of the candidate route The cumulative cruising mileage when driving to road segment 1 is 241 meters, the cumulative cruising mileage when driving to road segment 2 is 556 meters, the cumulative cruising mileage when driving to road segment 3 is 752 meters, the cumulative cruising mileage when driving to road segment 4 is 850 meters, the cumulative cruising mileage when driving to road segment 5 is 1211 meters, the cumulative cruising mileage when driving to road segment 6 is 1459 meters, the cumulative cruising mileage when driving to road segment 7 is 1599 meters, and the cumulative cruising mileage when driving to road segment 8 is The cumulative cruising mileage is 1998 meters, the cumulative cruising mileage when driving to road section 9 is 2100 meters, the cumulative cruising mileage when driving to road section 10 is 2401 meters, the cumulative cruising mileage when driving to road section 11 is 2639 meters, the cumulative cruising mileage when driving to road section 12 is 3151 meters, the cumulative cruising mileage when driving to road section 13 is 3274 meters, the cumulative cruising mileage when driving to road section 14 is 3484 meters, and the cumulative cruising mileage when driving to road section 15 is 3831 meters. According to the cumulative cruising mileage of each road section Combined with the calculated conditional passenger carrying probability of all road segments, the expected cruising mileage calculation formula is used Calculate the expected cruising mileage of this path as If the expected cruising mileage of path 2 is 1022.15 meters obtained by this method, path 1 with a shorter expected cruising mileage is determined as the recommended path with the highest cruising efficiency.

[0050] Step S106, a cruising route updating module is used to determine whether the driver is driving along the recommended route based on the real-time location information of the taxi, and to update the road segment capacity data. Based on the updated road segment capacity data, the route with the shortest expected cruising mileage is re-determined as the recommended route.

[0051] Get the longitude and latitude sequence information of the recommended route, and transmit the longitude and latitude sequence information of the recommended route to the background server through the API interface. Use the map rendering API to load the longitude and latitude data of the recommended route, render the recommended route on the map, and obtain a visualized recommended route. Get the driver's real-time location information, compare the driver's route with the recommended route, and determine whether the driver is driving along the recommended route. If the driver is driving along the recommended route, traverse all road segments that the driver drives along the recommended route, add 1 to the road segment capacity of each road segment, and obtain the updated road segment capacity data. According to the updated road segment capacity data, recalculate the updated passenger probability prediction value, and adjust the passenger probability of the corresponding road segment in the key-value database. According to the updated passenger probability prediction value, re-determine the path with the shortest expected cruising mileage as the recommended path with the highest cruising efficiency.

[0052] For example, on June 1, 2020, a driver initiated a route recommendation request from point A to point B. The cruising route real-time recommendation module recommended route 1 as the recommended route. Route 1 contains the following longitude and latitude sequence: [(113.32334, 23.13161), (113.32400, 23.13200), (113.32450, 23.13250), ...]. This set of longitude and latitude sequence information is transmitted to the backend server through the API interface. The longitude and latitude data of the recommended route is loaded using the map rendering API, and the recommended route is rendered on the map. The driver can see a clear recommended route in the client map. The driver's location information is obtained in real time, and his driving route is compared with the recommended route to confirm that the driver is driving along the recommended route. When the driver drives along the recommended route, all the road segments driven by the driver are traversed. If the driver drives road segment 1, road segment 2, and road segment 3 in route 1, the capacity of these road segments is accumulated by 1 respectively. The updated road segment capacity data is that the capacity of road segment 1 changes from 3 to 4, the capacity of road segment 2 changes from 4 to 5, and the capacity of road segment 3 changes from 5 to 6. According to the updated road segment capacity data, the passenger carrying probability prediction value of each road segment is recalculated. If the original passenger carrying probability of road segment 1 is 22%, the updated passenger carrying probability is 20%; the passenger carrying probability of road segment 2 is 14%, which is updated to 12%; the passenger carrying probability of road segment 3 is 41%, which is updated to 38%. These updated passenger carrying probabilities are stored in the key-value database, and the passenger carrying probability prediction values ​​of the corresponding road segments in the database are also updated accordingly. According to the updated passenger probability prediction value, the expected cruising distance of each candidate path is recalculated. If the expected cruising distance of path 1 is updated to 680 meters, and the expected cruising distances of path 2 and path 3 are 700 meters and 750 meters respectively. By comparing the updated expected cruising distances, it is determined that path 1 is still the path with the shortest expected cruising distance, so it continues to be the recommended path with the highest cruising efficiency.

[0053] like Figure 2 The present embodiment provides a system for calculating a cruising path of a rental vehicle based on expected cruising mileage, which may specifically include:

[0054] The satellite positioning data real-time map matching module is used to obtain the preliminary trajectory of the taxi based on the real-time satellite positioning data of the taxi, calculate the distance and angle difference between the taxi trajectory point and the road segment, determine the road segment that meets the driving path and correct the preliminary trajectory point, and obtain the corresponding road segment number.

[0055] The road segment-level taxi capacity demand calculation module is used to determine the taxi capacity and demand of different road segments based on the taxi trajectory data of the road segment corresponding to the taxi trajectory point.

[0056] The road segment-level taxi passenger load probability prediction module is used to predict the future taxi load and demand of each road segment based on the historical taxi load and demand data of different road segments, and to predict the passenger load probability of each road segment based on the road segment passenger load probability prediction formula.

[0057] The taxi passenger carrying probability credibility correction module is used to evaluate the temporal periodic regularity of the taxi carrying capacity and demand of each road section based on the historical data of taxi carrying capacity and demand, and to correct the predicted value of the passenger carrying probability of each road section based on the temporal periodic regularity.

[0058] The real-time cruising route recommendation module is used to determine the candidate route set based on all road segment information, calculate the conditional passenger load probability of each road segment in the candidate route based on the conditional passenger load probability formula, and use the path expected cruising mileage calculation formula to calculate the expected cruising mileage of each candidate route, and determine the path with the shortest expected cruising mileage as the recommended path.

[0059] The cruising route updating module is used to determine whether the driver is driving along the recommended route based on the real-time location information of the taxi, and to update the road segment capacity data. Based on the updated road segment capacity data, the route with the shortest expected cruising mileage is re-determined as the recommended route.

[0060] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the concept of the present application. For example, the above features are replaced with the technical features with similar functions disclosed in the present application (but not limited to) to form a technical solution.

Claims

1. A method for calculating a cruising path of a rental vehicle based on expected cruising mileage, characterized in that: The method comprises: The initial trajectory of the taxi is obtained based on the real-time satellite positioning data of the taxi, the distance and angle difference between the taxi trajectory point and the road segment are calculated, the road segment that meets the driving path is determined and the initial trajectory point is corrected, and the corresponding road segment number is obtained; According to the taxi trajectory data of the road section corresponding to the taxi trajectory point, the taxi transport capacity and demand of different road sections are determined; Based on the historical taxi capacity and demand data of different road sections, the future taxi capacity and demand of each road section are predicted, and based on the road section passenger probability prediction formula, the passenger probability of each road section is predicted; According to the historical data of taxi capacity and demand, the degree of temporal regularity of the capacity and demand of each road section is evaluated, and the predicted value of passenger probability of each road section is revised based on the degree of temporal regularity; Determine the candidate path set based on all road segment information, calculate the conditional passenger load probability of each road segment in the candidate path based on the conditional passenger load probability formula, and use the path expected cruising mileage calculation formula to calculate the expected cruising mileage of each candidate path, and determine the path with the shortest expected cruising mileage as the recommended path; According to the real-time location information of the taxi, it is determined whether the driver is driving according to the recommended route, and the road segment capacity data is updated. Based on the updated road segment capacity data, the path with the shortest expected cruising mileage is re-determined as the recommended route.

2. The method according to claim 1, wherein: The method of obtaining a preliminary trajectory of the taxi according to the real-time satellite positioning data of the taxi, calculating the distance and angle difference between the taxi trajectory point and the road segment, determining the road segment that meets the driving path and correcting the preliminary trajectory point, and obtaining the corresponding road segment number, includes: According to the real-time satellite positioning data and timestamp obtained by the GPS device of the taxi, a set of positioning points including longitude and latitude is obtained to obtain the preliminary trajectory of the taxi; through the map data, the potential driving road segments of the taxi are obtained, including the coordinates, direction and connectivity information of the road segments; according to the preliminary trajectory points of the taxi, the Euclidean distance between each trajectory point and the potential driving road segment is calculated, and the road segments with a distance less than a preset distance threshold are selected as candidate road segments; according to the difference between the longitude and latitude of adjacent preliminary trajectory points of the taxi, the angle calculation is performed to determine the movement direction of the taxi trajectory point; the angle difference between the direction of the candidate road segment and the movement direction of the taxi trajectory point is calculated to determine the candidate road segment with an angle difference less than a preset angle difference threshold; through the map data, according to the connectivity of the candidate road segment with an angle difference less than the preset angle difference threshold, the road segment that meets the actual driving path is determined; according to the road segment that meets the actual driving path, the coordinates of the preliminary trajectory points of the taxi are adjusted, and the coordinates of the preliminary trajectory points of the taxi with an offset are corrected to the road segment of the road where the taxi actually travels, the corrected taxi trajectory points are obtained, and the road segment number corresponding to each corrected taxi trajectory point is determined.

3. The method according to claim 1, wherein: The method of determining the taxi transport capacity and demand of different road sections according to the taxi trajectory data of the road section corresponding to the taxi trajectory point includes: According to the road segment corresponding to the taxi trajectory point, with t minutes as the time granularity, obtain the taxi trajectory data set within the past preset time period, the taxi trajectory data includes time, road segment, license plate number and passenger status, the passenger status includes empty and with passengers; through the taxi trajectory data set within the past preset time period, obtain the taxi license plate number of the same road segment, remove duplicate license plate numbers, and determine the taxi transport capacity of each road segment every t minutes in the past preset time period, the taxi transport capacity is the number of empty taxis; obtain the passenger status corresponding to the taxi license plate number every t minutes in the past preset time period, judge the time point and road segment when the passenger status changes from empty to with passengers, obtain the license plate number set of passengers picked up on each road segment every t minutes, and determine the number of passengers picked up on each road segment every t minutes as the taxi demand.

4. The method according to claim 1, wherein: The method predicts the future taxi capacity and demand of each road segment based on the historical taxi capacity and demand data of different road segments, and predicts the passenger carrying probability of each road segment based on the road segment passenger carrying probability prediction formula, including: Obtain the historical taxi capacity and demand data of each road section, take t minutes as the time granularity, obtain the taxi capacity and demand data of every t minutes in the historical taxi capacity and demand data, and construct the time series data of each road section; according to the time series data of each road section, use the spatiotemporal graph convolutional neural network algorithm to train the model and construct a taxi capacity and demand prediction model; according to the taxi capacity and demand data of each road section every t minutes in the past preset time period, use the taxi capacity and demand prediction model to predict the capacity and demand of each road section in the next t minutes; according to the road section passenger probability prediction formula Predict the passenger carrying probability of each road segment, where P represents the predicted value of the road segment passenger carrying probability, D is the predicted value of the road segment demand, and S is the predicted value of the road segment transport capacity.

5. The method according to claim 1, wherein: The method of evaluating the time series regularity of the taxi capacity and demand of each road segment based on the historical data of taxi capacity and demand, and revising the predicted value of the passenger probability of each road segment based on the time series regularity, includes: Obtain the historical data of taxi transport capacity and demand on each road section to obtain a complete time series data set; use the time series period regularity degree formula Evaluate the degree of regularity of the time series of transport capacity and demand of each road section, among which, represents the taxi demand sequence of road segment j in the i-th time period, represents the taxi capacity time series of road segment j in the i-th time period, represents the taxi demand sequence of road segment j in the i+1th time period, represents the taxi capacity time series of road segment j in the i+1th time period, and m represents the number of time periods covered by historical data; according to the regularity of the time series cycle of the capacity and demand of each road segment, the reliability correction coefficient β of the passenger probability prediction of each road segment is determined j ,in, According to the predicted value of passenger probability of each road section, the credibility correction coefficient and credibility correction formula are used The predicted value of the passenger-carrying probability of each road section is corrected to determine the corrected taxi passenger-carrying probability of each road section in the entire road network in the next t minutes; the corrected taxi passenger-carrying probability of each road section in the entire road network in the next t minutes is stored in a memory-based key-value database, where the key value is the road section code and the value value is the corrected predicted value of the taxi passenger-carrying probability of the corresponding road section in the next t minutes And update the predicted data of the passenger probability of the road segment in the future t minutes in the cached data every t minutes.

6. The method according to claim 5, wherein: The method of determining a candidate path set according to all road segment information, calculating the conditional passenger carrying probability of each road segment in the candidate path based on the conditional passenger carrying probability formula, and calculating the expected cruising mileage of each candidate path using the path expected cruising mileage calculation formula, and determining the path with the shortest expected cruising mileage as the recommended path includes: According to the route recommendation request initiated by the driver, the starting point and the end point are obtained, and the latitude and longitude coordinates of the starting point and the end point are determined; Obtain map data, extract all road segment information, use the depth-first path search algorithm to obtain all paths with a length of K kilometers starting from the request initiation point, and obtain the candidate path set G; query the predicted passenger probability of all road segments included in all candidate paths in the future t minutes through the key-value database; define a candidate path in the candidate path set G as G u =(l1,l2,…,l n ), the passenger carrying probability of each road segment of the candidate path is Where n represents the number of road segments contained in the candidate path; According to the conditional passenger probability formula Calculate the conditional passenger loading probability of a single road segment, where represents the cumulative passenger carrying probability, and the cumulative passenger carrying probability is calculated as follows: Get the length of each road segment of the candidate path through the formula Calculate the cumulative cruising mileage of each road segment of the candidate route Where L i For road segment l i length; based on the accumulated cruising mileage and the predicted value of passenger probability The expected cruising mileage calculation formula is adopted Calculate the expected cruising mileage of each candidate path; compare the expected cruising mileages of different candidate paths, and determine the path with the shortest expected cruising mileage as the recommended path with the highest cruising efficiency.

7. The method according to claim 1, wherein: The method of determining whether the driver is traveling along the recommended route based on the real-time location information of the taxi, updating the road segment capacity data, and re-determining the route with the shortest expected cruising mileage as the recommended route based on the updated road segment capacity data includes: Obtain the longitude and latitude sequence information of the recommended path, and transmit the longitude and latitude sequence information of the recommended path to the backend server through the API interface; Use the map rendering API to load the latitude and longitude data of the recommended path, render the recommended path on the map, and obtain a visualized recommended path route; The real-time location information of the driver is obtained, and the driver's driving route is compared with the recommended route to determine whether the driver is driving along the recommended route; if the driver is driving along the recommended route, all road segments on which the driver drives along the recommended route are traversed, and the road segment capacity of each road segment is accumulated by 1 to obtain updated road segment capacity data; based on the updated road segment capacity data, the updated passenger loading probability prediction value is recalculated, and the passenger loading probability of the corresponding road segment in the key-value database is adjusted; based on the updated passenger loading probability prediction value, the path with the shortest expected cruising mileage is re-determined as the recommended path with the highest cruising efficiency.

8. A system for calculating the cruising path of a rental vehicle based on expected cruising mileage, characterized in that: The system comprises: The satellite positioning data real-time map matching module is used to determine the road segment that matches the taxi's driving path and correct the driving trajectory, and obtain the corresponding road segment number; The road segment-level taxi capacity demand calculation module is used to determine the taxi capacity and demand of different road segments according to the taxi trajectory data of the road segment corresponding to the taxi trajectory point; The road segment-level taxi passenger probability prediction module is used to predict the future taxi capacity, demand and passenger probability of each road segment based on the historical taxi capacity and demand data of different road segments; The taxi passenger carrying probability credibility correction module is used to evaluate the temporal periodic regularity of the transport capacity and demand of each road section, and to correct the predicted value of the passenger carrying probability of each road section based on the temporal periodic regularity; A real-time cruise route recommendation module is used to determine a candidate route set based on all road segment information, calculate the expected cruise mileage of each candidate route, and determine the route with the shortest expected cruise mileage as the recommended route; The cruising route update module is used to determine whether the driver drives according to the recommended route, update the road segment capacity data, and re-determine the route with the shortest expected cruising mileage as the recommended route.

Images