A method, device, equipment and medium for risk detection in ride-hailing operations

Abstract

This invention relates to the field of ride-hailing risk monitoring, specifically to a method, device, equipment, and medium for risk detection in ride-hailing operations. This application simultaneously acquires real-time order data and associated trip-aware data, overcoming the limitations of a single data source. By fusing multi-dimensional data, it mines comprehensive risk control information, solving the problem of insufficient data source coverage. Secondly, based on the comprehensive analysis of these two types of data, the operational risk situation is analyzed more accurately than with limited data, enabling comprehensive risk identification. Then, an initial control strategy is matched according to the risk situation to ensure that the strategy is adapted to the risk. Finally, the strategy is adjusted in conjunction with the overall regional operational pressure, so that control measures both consider risk prevention and control and adapt to real-time operational scenarios, avoiding mechanical risk control and addressing the shortcomings of existing technologies in adapting to complex scenarios and inaccurate risk identification, thereby improving the effectiveness of risk control and scenario adaptability.

Description

Technical Field

[0001] This invention relates to the field of ride-hailing risk monitoring, specifically to a risk detection method, device, equipment, and medium for ride-hailing operations. Background Technology

[0002] Against the backdrop of the continuous expansion of the ride-hailing industry's operational scale and increasing operational complexity, ensuring driver and passenger safety, maintaining platform operational order, and preventing various operational risks have become core requirements for supporting the industry's stable development. In the development of ride-hailing risk control technology, early risk control solutions relied on simple rule-based judgment models, identifying abnormal orders by setting fixed thresholds for a few parameters such as order amount and trip time. With technological advancements, machine learning algorithms have been introduced into ride-hailing risk control scenarios, enabling risk prediction based on multi-dimensional order data; however, these related technical solutions still suffer from significant adaptability deficiencies.

[0003] Existing technical solutions mostly build risk control models based on single or limited types of data sources. Some systems rely solely on order information and identify abnormal orders through preset rules, such as classifying orders with amounts exceeding a multiple of the regional time period average or orders with inconsistent origin and destination logic as risky orders. Other solutions analyze order data and driver and passenger historical behavior characteristics through machine learning models, but fail to fully explore key dimensions such as real-time vehicle driving status, surrounding road conditions, and third-party credit information. This makes it difficult for risk control models to adapt to complex and ever-changing operational scenarios and achieve accurate and comprehensive risk identification. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a risk detection method, device, equipment and medium for ride-hailing operations to solve the problems of insufficient data source coverage of the risk control model in the prior art, inability to fully mine multi-dimensional risk control related information, difficulty in adapting to complex operating scenarios, and inability to accurately and comprehensively identify ride-hailing operation risks.

[0005] In a first aspect, embodiments of the present invention provide a risk detection method for ride-hailing operations, the method comprising: Obtain real-time order data corresponding to the current ride-hailing order and the trip perception data associated with the current order; By analyzing the real-time order data and the trip perception data, the current operational risk status of the ride-hailing vehicle can be obtained; Obtain an initial operational control strategy that matches the aforementioned operational risk situation; The initial operation and control strategy is adjusted based on the overall operational pressure in the area where the ride-hailing vehicle is located to obtain the target operation and control strategy, and the ride-hailing vehicle is managed in accordance with the target operation and control strategy.

[0006] Furthermore, the acquisition of real-time order data corresponding to the real-time ride-hailing order and the trip-aware data associated with the real-time order includes: Retrieve the current order of the ride-hailing vehicle, and obtain the real-time price data and operating time data of the ride-hailing vehicle in the current order, and use the real-time price data and the operating time data as the real-time data of the order; Obtain the real-time audio recording data and vehicle driving data generated by the ride-hailing vehicle in the current order; Obtain real-time traffic data associated with the ride-hailing vehicle's trip in the current order; The trip perception data is based on the real-time audio recording data, the vehicle driving data, and the real-time traffic data.

[0007] Furthermore, the analysis of the real-time order data and the trip perception data to obtain the current operational risk status of the ride-hailing service includes: Detect whether there are any abnormal data in the real-time price data and the operating time data; If there are corresponding abnormal data in the real-time price data and the operating time data, then a first risk quantification value is calculated based on the abnormal data in the real-time price data and the operating time data; Detect whether there are any abnormal data in the real-time audio recording data, the vehicle driving data, and the real-time traffic data; If there are any abnormal data in the real-time price data and the operating time data, a second risk quantification value is calculated based on the abnormal data in the real-time recording data, the vehicle driving data and the real-time traffic data. The first risk quantification value and the second risk quantification value are obtained to determine the current operational risk status of the ride-hailing vehicle.

[0008] Furthermore, the step of obtaining the first risk quantification value and the second risk quantification value to determine the current operational risk status of the ride-hailing vehicle includes: The target risk quantification value is obtained by weighted summation of the first risk quantification value and the second risk quantification value. Obtain the numerical range into which the target risk quantification value falls, and or region the risk level corresponding to the numerical range; Based on the second quantitative risk value, it is determined that the driver of the ride-hailing vehicle is currently engaging in risky driving behavior; The risk level and the driving risk behavior are considered as the operational risk situation.

[0009] Furthermore, the acquisition of an initial operational control strategy that matches the operational risk situation includes: Extract risk levels and driving risk behaviors from the described operational risk conditions; If the risk level is higher than the preset level, then obtain the order intervention strategy corresponding to the risk level; Based on the aforementioned driving risk behaviors, determine the corresponding driving intervention strategies; The initial operation control strategy is generated based on the order intervention strategy and the driving intervention strategy.

[0010] Furthermore, adjusting the initial operation control strategy based on the overall operational pressure in the area where the ride-hailing vehicle is located to obtain the target operation control strategy includes: Analyze the overall operational pressure to determine the current operational demand in the area where the ride-hailing service is located; Based on the operational needs, determine the adjustable sub-control strategies in the initial operational control strategy, as well as the priorities corresponding to the control strategies; The adjustable sub-control strategies are adjusted according to the priority to obtain the adjusted sub-control strategies, and the changing trend of the overall operational pressure is obtained. By utilizing the aforementioned trend, the execution trigger conditions and execution time of the adjusted sub-control strategy are determined, thus obtaining the target operation control strategy.

[0011] Furthermore, the method also includes: After managing the ride-hailing vehicles in accordance with the target operation and management strategy, obtain the corresponding management feedback. If the feedback from the control measures indicates that the operational risk of the ride-hailing vehicle has decreased, then other ride-hailing vehicles with the same risk in the same area as the ride-hailing vehicle will be subject to synchronized operational control in accordance with the target operational control strategy.

[0012] Secondly, embodiments of the present invention provide a risk detection device for ride-hailing operations, the device comprising: The data acquisition module is used to acquire real-time order data corresponding to the real-time orders of ride-hailing vehicles and trip perception data associated with the real-time orders; The analysis module is used to analyze the real-time order data and the trip perception data to obtain the current operational risk status of the ride-hailing vehicle; The acquisition module is used to acquire an initial operational control strategy that matches the operational risk situation. The adjustment module is used to adjust the initial operation control strategy based on the overall operational pressure of the area where the ride-hailing vehicle is located, to obtain the target operation control strategy, and to control the ride-hailing vehicle in accordance with the target operation control strategy.

[0013] Thirdly, embodiments of the present invention provide a computer device, including: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the method described in the first aspect or any corresponding embodiment thereof.

[0014] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer instructions that cause a computer to perform the method described in the first aspect or any of its corresponding embodiments.

[0015] This application simultaneously acquires real-time order data and associated trip-aware data, breaking through the limitations of a single data source. By integrating multi-dimensional data, it mines comprehensive risk control information and solves the problem of insufficient data source coverage. Secondly, based on the comprehensive analysis of operational risks using both types of data, the assessment is more accurate than with limited data and can comprehensively identify risks. Then, initial control strategies are matched according to the risk situation to ensure that the strategies are adapted to the risks. Finally, the strategies are adjusted in conjunction with the overall regional operational pressure, so that control measures take into account both risk prevention and control and adapt to real-time operational scenarios, avoiding mechanical risk control and solving the shortcomings of existing technologies in adapting to complex scenarios and inaccurate and comprehensive risk identification, thereby improving the effectiveness of risk control and scenario adaptability. Attached Figure Description

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a risk detection method for ride-hailing operations according to some embodiments of the present invention; Figure 2 This is a flowchart illustrating another risk detection method for ride-hailing operations according to some embodiments of the present invention; Figure 3 This is a structural block diagram of a risk detection device for ride-hailing operations according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] According to embodiments of the present invention, a risk detection method, apparatus, device, and medium for ride-hailing operations are provided. It should be noted that the steps shown in the flowcharts in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowcharts, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0020] This embodiment provides a risk detection method for ride-hailing operations. Figure 1 This is a flowchart of a risk detection method for ride-hailing operations according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain real-time order data corresponding to the current ride-hailing order and trip perception data associated with the current order.

[0021] First, the system connects to the ride-hailing platform's order data interface, vehicle sensor data interface, trip recording storage node, and third-party data synchronization interface to trigger a data retrieval command. This retrieves real-time order data for the current order, including the order's origin, destination, real-time trip duration, incurred fare, driver's current order acceptance status, and passenger's current location information. Simultaneously, it acquires associated trip-awareness data, including sensor data such as vehicle speed, acceleration, and braking frequency, real-time segments of trip recordings, traffic congestion data for the current road segment, and real-time credit data snapshots for both the driver and passenger. The acquired data undergoes format validation, and upon successful validation, it is temporarily stored in a memory cache according to the "Order ID - Data Type" structure, awaiting subsequent analysis.

[0022] In this embodiment of the application, obtaining real-time order data corresponding to real-time ride-hailing orders and trip-aware data associated with real-time orders includes: Step A1: Retrieve the current order of the ride-hailing vehicle, and obtain the real-time price data and operating time data of the ride-hailing vehicle in the current order, and use the real-time price data and operating time data as the real-time data of the order.

[0023] First, the interface of the ride-hailing order management module is called, and the unique identifier ID of the current order is input to retrieve the corresponding current order data. Real-time price-related fields are extracted from the order data, including the current trip cost, the real-time fare per unit, and the cost change records. This data is then organized into real-time price data. Next, operating time-related fields are extracted, including the order acceptance time, the current driving time, and the remaining estimated driving time. This data is then organized into operating time data. The real-time price data and operating time data are then structured and integrated according to field names, and the data collection timestamp is marked to form a real-time order dataset, which is stored in a temporary data queue for subsequent association and retrieval.

[0024] Step A2: Obtain real-time audio recordings and vehicle driving data generated by the ride-hailing service in the current order.

[0025] The system interfaces with the in-vehicle recording device for real-time transmission. Based on the current order ID, it matches the corresponding recording stream, extracts audio segments from the order acceptance time to the current time, and converts these audio segments into standardized audio format files as real-time recording data. It also calls the real-time data interface of the vehicle sensors to obtain second-level data collected by the speed sensor, acceleration sensor, and brake sensor during the current order's journey, including vehicle speed, acceleration values, and brake trigger status at each collection point. This data is then organized according to time sequence to form vehicle driving data. Finally, it adds an order ID identifier to both the real-time recording data and the vehicle driving data, and stores them synchronously in the trip data storage partition.

[0026] Step A3: Obtain real-time traffic data associated with the ride-hailing trip in the current order.

[0027] The system reads the real-time driving location data of the current order, connects to the interface of the cooperating map navigation platform, inputs the latitude and longitude information of the current location, and obtains real-time traffic data for the location and the surrounding 5-kilometer range. It extracts congestion-related fields from the traffic data, including the congestion level, duration of congestion, estimated travel time, location of construction sections, and the scope of construction impact. Simultaneously, it extracts weather-related traffic impact data, including the weather conditions of the current road section and the impact of weather on the road surface friction coefficient. This data is then labeled with the order ID and collection time, and compiled into a real-time traffic dataset associated with the current order's journey.

[0028] Step A4: Use real-time audio recording data, vehicle driving data, and real-time traffic data as trip perception data.

[0029] The system reads real-time audio recordings and vehicle driving data marked with the current order ID from the trip data storage partition, and reads the corresponding real-time traffic data from the traffic data storage partition. It then calls a data association algorithm to align the time nodes of the real-time audio recordings and vehicle driving data with the time nodes of the real-time traffic data, using the order's time series as a benchmark. The aligned three types of data are then structurally concatenated, unifying the timestamp format and removing duplicate time nodes. Finally, the concatenated data is marked as trip-aware data and stored in the feature data buffer.

[0030] Step S102: Analyze real-time order data and trip perception data to obtain the current operational risk situation of ride-hailing services.

[0031] The system reads real-time order data and trip-aware data from the memory cache, calls preset feature extraction rules, and extracts the following from the real-time order data: order amount fluctuation range, order travel time deviation from the average percentage, and passenger order time characteristics. From the trip-aware data, it extracts the number of rapid accelerations per unit time, the percentage of speeding time, the frequency of abnormal keywords in the recording, and the risk level of the current road segment. The extracted features are input into a trained dynamic risk assessment model, which outputs the risk probability value of the current order through feature matching and probability calculation. The risk probability value is matched with a preset risk level range to obtain the corresponding operational risk situation, such as a risk probability of 0.7-0.8 corresponding to moderate risk, and above 0.8 corresponding to high risk.

[0032] In this embodiment of the application, real-time order data and trip perception data are analyzed to obtain the current operational risk status of the ride-hailing service, including: Step B1: Check for any abnormal data in the real-time price data and operating time data.

[0033] The system loads preset rules for judging anomalies in real-time price data and operating time data. The rules for real-time price data include price fluctuations exceeding 30% of the average for the same route and time period, and real-time unit prices exceeding the platform's pricing rule range. The rules for operating time data include travel time exceeding 50% of the estimated travel time, and a negative difference between the order acceptance time and the current time. The system iterates through the fields of the real-time order data, comparing the real-time price data and operating time data with the corresponding rules. Fields exceeding the rule thresholds are marked as abnormal data. The system outputs the field name of the abnormal data, the difference between the actual value and the threshold, and generates the real-time order data anomaly detection results.

[0034] Step B2: If there are corresponding abnormal data in the real-time price data and operating time data, calculate the first risk quantification value based on the abnormal data in the real-time price data and operating time data.

[0035] Read the anomaly detection results of real-time order data and extract the difference between the actual value of the anomaly data and the threshold; call the preset first risk quantification calculation rule to normalize the difference of each anomaly field and map it to the [0,1] interval; based on the risk weight of each anomaly field (preset price data weight 0.6, operating time data weight 0.4), multiply the normalized value by the weight and sum them to obtain the first risk quantification value; keep the first risk quantification value to three decimal places, mark the corresponding anomaly data type, and store it in the risk quantification result set.

[0036] Step B3: Detect whether there are any abnormal data in the real-time audio recording data, vehicle driving data, and real-time traffic data.

[0037] Anomaly detection rules are loaded for real-time audio recording data, vehicle driving data, and real-time traffic data: the anomaly rule for real-time audio recording data is the presence of keywords related to arguments or threats; the anomaly rule for vehicle driving data is exceeding the speed limit by 20% or having more than 5 instances of rapid acceleration within 10 minutes; and the anomaly rule for real-time traffic data is a traffic congestion level exceeding level 4 and a dwell time exceeding 10 minutes. The fields or contents of the three types of data are traversed, the data is compared with the corresponding rules, and the content that meets the anomaly rule is marked as abnormal data. The type and occurrence time of the abnormal data are output, and the trip perception data anomaly detection results are generated.

[0038] Step B4: If there are corresponding abnormal data in the real-time price data and operating time data, calculate the second risk quantification value based on the abnormal data in the real-time recording data, vehicle driving data and real-time traffic data.

[0039] The system reads the anomaly detection results from the trip perception data and extracts the quantitative parameters of each anomaly, such as the frequency of abnormal keywords in the recording, the duration of vehicle speeding, and the duration of traffic congestion. It then calls a preset second risk quantification calculation rule to normalize each anomaly parameter, mapping it to the [0,1] interval. Based on the risk weights of each data type (0.3 for recording data, 0.5 for vehicle data, and 0.2 for traffic data), the normalized values ​​are multiplied by their weights and summed to obtain the second risk quantification value. The second risk quantification value is then rounded to three decimal places, the corresponding anomaly data type is marked, and the result is stored in the risk quantification result set.

[0040] Step B5: Obtain the first risk quantification value and the second risk quantification value to determine the current operational risk situation of the ride-hailing service.

[0041] In this embodiment of the application, obtaining a first risk quantification value and a second risk quantification value to determine the current operational risk status of a ride-hailing vehicle includes: performing a weighted summation based on the first risk quantification value and the second risk quantification value to obtain a target risk quantification value; obtaining the numerical range into which the target risk quantification value falls, and the risk level corresponding to the numerical range; determining that the ride-hailing driver currently exhibits risky driving behavior based on the second quantification risk value; and using the risk level and risky driving behavior as the operational risk status.

[0042] The system retrieves the first and second risk quantification values ​​from the risk quantification result set, loads the preset weighting coefficient configuration (first risk weight 0.4, second risk weight 0.6), and performs a weighted summation calculation using the formula "target risk quantification value = first risk quantification value × 0.4 + second risk quantification value × 0.6", retaining three decimal places. It then calls the risk level interval mapping table, which presets interval rules: [0, 0.5) corresponds to low risk, [0.5, 0.8) to medium risk, and [0.8, 1] to high risk. The target risk quantification value is substituted into the interval for comparison to determine the corresponding risk level. A threshold of 0.5 is set for the second risk quantification value. If the value exceeds the threshold, combined with abnormal records in the trip perception data, the type of driving risk behavior (such as rapid acceleration, speeding, voice conflict) and its occurrence parameters are identified; if it does not exceed the threshold, it is marked as "no significant driving risk behavior". Finally, the determined risk levels and driving risk behaviors are integrated into structured operational risk data, including fields such as level identifier, behavior type, and risk quantification basis.

[0043] As an example, in the driver's current order, the first risk quantification value is 0.6 (due to a 40% surge in order fees within 10 minutes), and the second risk quantification value is 0.75 (due to three instances of rapid acceleration and two minutes of speeding within 5 minutes). The target risk quantification value, calculated using a weighted average, is 0.6 × 0.4 + 0.75 × 0.6 = 0.69, which falls within the range of [0.5, 0.8), corresponding to a medium risk level. Since the second risk quantification value of 0.75 exceeds the 0.5 threshold, combined with abnormal records, the driving risk behavior is determined to be "frequent rapid acceleration + short-term speeding." The final operational risk situation is: medium risk level, driving risk behavior "three instances of rapid acceleration within 5 minutes and speeding exceeding the limit by 20% for two minutes," accompanied by two risk quantification values ​​and their calculation basis.

[0044] Step S103: Obtain an initial operational control strategy that matches the operational risk situation.

[0045] The system retrieves a pre-defined risk control strategy library, which stores the mapping relationship between risk levels and initial operational control strategies. It then inputs the current operational risk level and matches the corresponding initial operational control strategy: for mild risk, it matches the "driver voice reminder + passenger risk warning" strategy; for moderate risk, it matches the "order verification + order acceptance range restriction" strategy; and for high risk, it matches the "order suspension + remote vehicle monitoring" strategy. The system extracts the parameters, execution conditions, and operation instructions of the matched initial operational control strategy and stores them in the strategy execution buffer for subsequent adjustments.

[0046] In this embodiment of the application, obtaining an initial operational control strategy that matches the operational risk situation includes: Step C1: Extract the risk level and driving risk behavior from the operational risk situation.

[0047] Read the structured data on operational risks, parse the field identifiers in the data, and locate the risk level field and the driving risk behavior field; extract the numerical or labeling information of the risk level field, such as the text labels "moderate risk" and "high risk", or the corresponding numerical ranges of 0.7-0.8 and above 0.8; extract the specific content of the driving risk behavior field, including the type of risk behavior, such as rapid acceleration, speeding, and voice conflict, as well as parameters such as the frequency and duration of the behavior; organize the extracted risk levels and driving risk behaviors according to the structure of "risk level - risk behavior type - behavior parameter", and store them in the intervention strategy matching dataset, waiting for subsequent strategy matching.

[0048] Step C2: If the risk level is higher than the preset level, obtain the order intervention strategy corresponding to the risk level.

[0049] The system loads a preset risk level threshold and compares the extracted risk level with the threshold. If the risk level value or identifier is higher than the preset level (e.g., the preset level is medium risk and the current risk level is high risk), the system retrieves the order intervention strategy library. The strategy library stores the mapping relationship between risk levels and order intervention strategies. Inputting the current risk level matches the corresponding order intervention strategy. For example, high risk matches the "order suspension + identity verification" strategy, and medium risk matches the "order warning + fee freeze" strategy. The system extracts the execution conditions, operation instructions, and triggering time of the matched order intervention strategy and stores these parameters in the strategy buffer.

[0050] Step C3: Determine the corresponding driving intervention strategy based on driving risk behaviors.

[0051] The system reads driving risk behaviors from the intervention strategy matching dataset, extracts the type and parameters of the behaviors, such as the frequency of rapid acceleration and the duration of speeding; it then retrieves the driving intervention strategy library, which stores the mapping relationship between driving risk behavior types and intervention strategies. By inputting a driving risk behavior type, the system matches the corresponding driving intervention strategy, such as matching rapid acceleration with a "real-time voice reminder" strategy, speeding with a "speed limit + voice warning" strategy, and voice conflict with a "remote monitoring + early warning push" strategy; and adjusts the execution intensity of the strategy based on the parameters of the driving risk behavior. For example, if the frequency of rapid acceleration exceeds 8 times / 10 minutes, the "single voice reminder" is adjusted to "two consecutive voice reminders," thus obtaining a suitable driving intervention strategy.

[0052] As an example, in the structured data of the current order's operational risk status, the risk level field is identified as "high risk" (corresponding to a value range of 0.85), and the driving risk behavior field contains "speeding (speed exceeding the limit by 25% for 3 minutes) + rapid acceleration (6 times within 10 minutes)". First, these two pieces of information are extracted, organized according to "risk level - risk behavior type - behavior parameters", and stored in the intervention strategy matching dataset. The risk level is compared with the preset "moderate risk" threshold. Because it is higher than the threshold, the order intervention strategy library is retrieved, and the "order suspension + identity verification" strategy is matched. Its execution conditions (risk level ≥ 0.8) and operation instructions (pause order, push identity verification interface) are extracted and stored in the strategy buffer. Next, the driving risk behavior is read, and after matching with the driving intervention strategy library, the "speed limit + continuous voice reminder" strategy is obtained (due to 6 instances of rapid acceleration, the single reminder is adjusted to two consecutive reminders). Finally, the strategy fusion algorithm is invoked to align the triggering times of the two strategies (speed limit and voice reminder to be executed immediately after the order is suspended). After verifying that there are no logical conflicts, the initial operation control strategy is generated: "Triggering condition: Risk level ≥ 0.8; Execution steps: 1. Suspend the current order; 2. Push identity verification request to the driver and passengers; 3. Limit the vehicle speed to the speed limit; 4. Push two consecutive voice reminders 'Please drive smoothly, speed limit has been set'; Operation target: Mr. Zhang's current order and vehicle terminal."

[0053] Step C4: Generate an initial operation control strategy based on the order intervention strategy and the driving intervention strategy.

[0054] The system reads the order intervention strategy and driving intervention strategy from the strategy buffer, calls the strategy fusion algorithm to align the execution instructions and triggering times of the two strategies, merges duplicate operation instructions, and adjusts the execution order of the strategies. For example, it executes the "order warning" in the order intervention strategy first, and then executes the "voice reminder" in the driving intervention strategy. The system performs parameter verification on the fused strategy to check whether the execution conditions of the strategies conflict. For example, it checks whether there is a logical contradiction between the order suspension instruction and the speed limit instruction. If so, it adjusts according to the "order priority" rule. After the verification passes, the fused strategy is marked as the initial operation control strategy, and the execution steps, operation objects, and triggering conditions of the strategy are output.

[0055] Step S104: Adjust the initial operation and control strategy based on the overall operational pressure in the area where the ride-hailing vehicle is located to obtain the target operation and control strategy, and manage the ride-hailing vehicle in accordance with the target operation and control strategy.

[0056] Connect to regional operation data nodes to obtain real-time order volume, number of online drivers, and order completion rate data for the ride-hailing area. Calculate the overall regional operational pressure value: Operational pressure value = (Real-time order volume / Number of online drivers) × Order completion rate deviation. Read the adjustment rules of the initial operation control strategy. If the operational pressure value is higher than the preset threshold, increase the execution intensity of the initial strategy, such as changing "order verification" to "order verification + secondary identity verification". If the operational pressure value is lower than the preset threshold, decrease the execution intensity of the initial strategy, such as changing "order suspension" to "order warning + continuous monitoring". Generate the adjusted target operation control strategy and send the strategy instructions to the corresponding execution nodes. The execution nodes perform control operations on ride-hailing orders and vehicle status according to the instructions.

[0057] In this embodiment of the application, the initial operation control strategy is adjusted based on the overall operational pressure in the area where the ride-hailing service is located to obtain the target operation control strategy, including: Step D1: Analyze the overall operational pressure and determine the current operational needs in the area where the ride-hailing service is located.

[0058] Real-time operational data for the ride-hailing service's location is retrieved, including four core indicators: the number of currently online drivers, the number of pending orders, the order completion rate, and the average driver idle time. The overall operational pressure value is calculated using the formula: "Overall operational pressure = (number of pending orders / number of online drivers) × (1 - order completion rate) + average driver idle time / preset baseline time". The overall operational pressure value is then matched with a preset pressure range to determine operational needs: if the pressure value is in the range [0, 0.3), the operational need is "improving order flow efficiency"; if it is in the range [0.3, 0.7), the operational need is "balancing order allocation and risk management"; and if it is in the range [0.7, 1], the operational need is "prioritizing order completion". These operational needs are then marked and stored in the strategy adjustment dataset.

[0059] Step D2: Based on operational needs, determine the adjustable sub-control strategies in the initial operational control strategy, as well as the priorities of the control strategies.

[0060] The system reads the structured data of the initial operational control strategy and breaks it down into three sub-control strategies: order intervention, driving intervention, and credit control. Adjustable sub-control strategies are then matched to current operational needs. For example, if the operational need is to improve order flow efficiency, the sub-control strategy can be adjusted to "order acceptance range restriction" or "order pause trigger threshold" for order intervention. If the need is to balance allocation and control, the sub-control strategy can be adjusted to "voice reminder frequency" or "vehicle speed limit intensity" for driving intervention. Simultaneously, preset strategy priority rules are loaded, with order intervention priority > driving intervention priority > credit control priority. Each adjustable sub-control strategy is marked with a corresponding priority value, and the relationship between priority and sub-control strategy is stored.

[0061] Step D3: Adjust the adjustable sub-control strategies according to priority to obtain the adjusted sub-control strategies and obtain the overall operational pressure change trend.

[0062] The parameters of the adjustable sub-control strategies are adjusted sequentially according to their priority from high to low: If the operational requirement is to prioritize order completion, the "order suspension trigger threshold" for high-priority order intervention sub-control strategies is increased from 0.8 to 0.9; for low-priority driving intervention sub-control strategies, the "speed limit trigger duration" is extended from 1 minute to 3 minutes. After the adjustment, the adjusted sub-control strategies are simulated and executed. The real-time update interface of regional operational data is retrieved, and the changes in online driver order volume and pending order volume within 5 minutes after the adjustment are collected. The magnitude and direction of the overall operational pressure change are calculated to obtain the trend. If the pressure value decreases, it is marked as a "positive trend," and if it increases, it is marked as a "negative trend."

[0063] Step D4: Utilize the changing trends to determine the execution trigger conditions and execution time of the adjusted sub-control strategies, thereby obtaining the target operational control strategy.

[0064] Read the overall operational pressure trend. If it's a positive trend, set the execution trigger condition of the adjusted sub-control strategy to "trigger when the real-time risk quantification value reaches the adjusted threshold," and the execution duration to "continue until the current order ends." If it's a negative trend, fine-tune the parameters of the sub-control strategy and simulate execution again until a positive trend is obtained. Integrate the adjusted sub-control strategy, execution trigger condition, and execution duration, replacing the corresponding content in the initial operational control strategy to form the target operational control strategy. Simultaneously, mark the basis for strategy adjustments and the verification results of trend changes to ensure strategy traceability.

[0065] As an example, during a certain evening rush hour, the overall operational pressure in the area where ride-hailing services operate is 0.8, and the operational requirement is "prioritizing order completion." The initial operational control strategy includes adjustable sub-control strategies: "Order Suspension Trigger Threshold" (priority 1) for order intervention and "Voice Reminder Frequency" (priority 2) for driving intervention. First, the order suspension trigger threshold was increased from 0.8 to 0.9, and then the voice reminder frequency was adjusted from once every 2 minutes to once every 5 minutes. After simulation, the number of pending orders in the area decreased by 20% within 10 minutes, and the overall operational pressure showed a positive downward trend. Finally, the adjusted sub-control strategy was determined to trigger order suspension when the risk quantification value is ≥ 0.9, with an execution duration of "the entire current order," thus forming the target operational control strategy.

[0066] In this embodiment of the application, the method further includes: after managing the ride-hailing vehicle in accordance with the target operation and control strategy, obtaining the corresponding management and control feedback; if the management and control feedback indicates that the operational risk of the ride-hailing vehicle has been reduced, then simultaneously managing and controlling other ride-hailing vehicles with the same risk in the area where the ride-hailing vehicle is located in accordance with the target operation and control strategy.

[0067] The system interfaces with ride-hailing platforms to extract order status data, driver behavior data, and risk quantification update data after the implementation of the target operational control strategy. This data is then integrated into a control feedback dataset. Risk change analysis rules are invoked to compare the target risk quantification value and the frequency of driving risk behaviors before and after control measures. If the target risk quantification value decreases by ≥20% and driving risk behaviors cease, the control feedback is deemed "operational risk reduced." If the result is a risk reduction, the system retrieves the order list for the current ride-hailing vehicle's region, filters out other ride-hailing vehicles with the same risk level and driving risk behaviors, and extracts their order IDs and driver IDs. The control strategy synchronization interface is then invoked to push the execution instructions and trigger conditions of the target operational control strategy to the corresponding execution nodes of these ride-hailing vehicles, initiating the synchronization control process. Simultaneously, the number of ride-hailing vehicles synchronized and the execution time are recorded for subsequent effect tracking.

[0068] As an example, after implementing the target control strategy (order alert + driving voice reminder every 3 minutes) for a driver's order that corresponds to "medium risk + high frequency of rapid acceleration," the control feedback data showed that the target risk quantification value dropped from 0.69 to 0.45, the rapid acceleration behavior stopped, and it was judged as "reduced operational risk." Subsequently, the order list of the area was retrieved, and the orders of 5 drivers with a risk level of medium risk and high frequency of rapid acceleration were selected. The control strategy of "order alert + driving voice reminder every 3 minutes" was simultaneously pushed to the order receiving terminal and vehicle voice system of these 5 drivers, and synchronous control was initiated. After 10 minutes, the tracking data showed that the frequency of rapid acceleration behavior of these 5 drivers all decreased by more than 60%, and the risk quantification value dropped to below 0.5 on average.

[0069] As a complete embodiment, such as Figure 2 As shown, firstly, ride-hailing order data (including order context, passenger and driver information), trip recordings, vehicle sensor data (speed, acceleration, etc. recorded in seconds), real-time traffic data (congestion, construction, weather impact), and third-party credit data (driver / passenger credit scores, consumption records) are acquired synchronously and then collected at the data collection node.

[0070] Next, data cleaning is completed by deduplication and removal of invalid values; then, order amounts are normalized and vehicle speeds are converted into speed limit ratios to achieve data standardization; missing data are filled with mean values ​​of the same dimension, and regression prediction is used to supplement time series data; the processed data flows into the dynamic risk assessment model stage.

[0071] Then, features such as order fluctuations, driving behavior, and credit changes are extracted from the preprocessed data; a model combining LSTM, CNN, and attention mechanism is used to complete the training by inputting historical normal / risk order data; when a new order is generated, real-time data is input into the model, and the order risk probability value is output.

[0072] Next, the strategy library is invoked to match order restrictions, driving intervention, and credit control strategies based on risk probability; strategies are adjusted in conjunction with the current regional operational pressure, such as relaxing risk thresholds when order volume is tight and tightening rules when order volume is sufficient; at the same time, feedback data from risk warning responses is received to optimize strategy parameters.

[0073] Finally, when the risk probability exceeds the warning threshold, a real-time warning is issued to drivers, passengers, and platform managers via SMS, APP push, and voice; responses are tiered according to risk level (mild voice reminder, moderate identity verification, severe remote vehicle control + emergency rescue notification); after the response, the results are tracked and the feedback data is sent back to the strategy generation stage to continuously iterate the risk control system.

[0074] This embodiment also provides a risk detection device for ride-hailing operations. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0075] This embodiment provides a risk detection device for ride-hailing operations, such as... Figure 3 As shown, it includes: The data acquisition module 301 is used to acquire real-time order data corresponding to the real-time orders of ride-hailing vehicles and trip perception data associated with the real-time orders; Analysis module 302 is used to analyze the real-time order data and the trip perception data to obtain the current operational risk status of the ride-hailing vehicle; Module 303 is used to acquire initial operational control strategies that match the operational risk situation; The adjustment module 304 is used to adjust the initial operation and control strategy based on the overall operation pressure of the area where the ride-hailing vehicle is located, to obtain the target operation and control strategy, and to manage the ride-hailing vehicle in accordance with the target operation and control strategy.

[0076] In this embodiment, the data acquisition module 301 is used to retrieve the current order of the ride-hailing vehicle, and obtain the real-time price data and operating time data of the ride-hailing vehicle in the current order, and use the real-time price data and operating time data as real-time order data; obtain the real-time audio recording data and vehicle driving data generated by the ride-hailing vehicle in the current order; obtain the real-time traffic data associated with the order trip of the ride-hailing vehicle in the current order; and use the real-time audio recording data, vehicle driving data and real-time traffic data as trip perception data.

[0077] In this embodiment, the analysis module 302 is used to detect whether there are any abnormal data in the real-time price data and operating time data; if there are any abnormal data in the real-time price data and operating time data, a first risk quantification value is calculated based on the abnormal data in the real-time price data and operating time data; detect whether there are any abnormal data in the real-time recording data, vehicle driving data, and real-time traffic data; if there are any abnormal data in the real-time price data and operating time data, a second risk quantification value is calculated based on the abnormal data in the real-time recording data, vehicle driving data, and real-time traffic data; and obtain the first risk quantification value and the second risk quantification value to determine the current operational risk status of the ride-hailing vehicle.

[0078] In this embodiment of the application, the analysis module 302 is used to perform a weighted summation based on the first risk quantification value and the second risk quantification value to obtain the target risk quantification value; obtain the numerical range into which the target risk quantification value falls, and the risk level corresponding to the numerical range of the region; determine that the driver of the ride-hailing vehicle currently has driving risk behavior based on the second quantification risk value; and take the risk level and driving risk behavior as the operational risk situation.

[0079] In this embodiment of the application, the acquisition module 303 is used to extract the risk level and driving risk behavior from the operational risk situation; if the risk level is higher than the preset level, the order intervention strategy corresponding to the risk level is acquired; the corresponding driving intervention strategy is determined based on the driving risk behavior; and an initial operation control strategy is generated based on the order intervention strategy and the driving intervention strategy.

[0080] In this embodiment, the adjustment module 304 is used to analyze the overall operational pressure and determine the current operational demand in the area where the ride-hailing service is located; based on the operational demand, determine the adjustable sub-control strategies in the initial operational control strategy and the priority of the control strategies; adjust the adjustable sub-control strategies according to the priority to obtain the adjusted sub-control strategies and obtain the trend of overall operational pressure; use the trend to determine the execution triggering conditions and execution time of the adjusted sub-control strategies to obtain the target operational control strategy.

[0081] In this embodiment of the application, the device further includes: a feedback module, used to obtain corresponding control feedback after controlling the ride-hailing vehicle in accordance with the target operation control strategy; if the control feedback indicates that the operation risk of the ride-hailing vehicle has been reduced, then other ride-hailing vehicles with the same risk in the area where the ride-hailing vehicle is located are simultaneously controlled in accordance with the target operation control strategy.

[0082] Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 4As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system).

[0083] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0084] The memory 20 stores instructions executable by at least one processor 10 to cause the at least one processor 10 to perform the method shown in the above embodiments.

[0085] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device as shown by a landing page for an app. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0086] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0087] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0088] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0089] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A risk detection method for ride-hailing operations, characterized in that, The method includes: Obtain real-time order data corresponding to the current ride-hailing order and the trip perception data associated with the current order; By analyzing the real-time order data and the trip perception data, the current operational risk status of the ride-hailing vehicle can be obtained; Obtain an initial operational control strategy that matches the aforementioned operational risk situation; The initial operation and control strategy is adjusted based on the overall operational pressure in the area where the ride-hailing vehicle is located to obtain the target operation and control strategy, and the ride-hailing vehicle is managed in accordance with the target operation and control strategy.

2. The method according to claim 1, characterized in that, The acquisition of real-time order data corresponding to the real-time ride-hailing order and the trip awareness data associated with the real-time order includes: Retrieve the current order of the ride-hailing vehicle, and obtain the real-time price data and operating time data of the ride-hailing vehicle in the current order, and use the real-time price data and the operating time data as the real-time data of the order; Obtain the real-time audio recording data and vehicle driving data generated by the ride-hailing vehicle in the current order; Obtain real-time traffic data associated with the ride-hailing vehicle's trip in the current order; The trip perception data is based on the real-time audio recording data, the vehicle driving data, and the real-time traffic data.

3. The method according to claim 2, characterized in that, The analysis of the real-time order data and the trip perception data yields the current operational risk status of the ride-hailing service, including: Detect whether there are any abnormal data in the real-time price data and the operating time data; If there are corresponding abnormal data in the real-time price data and the operating time data, then a first risk quantification value is calculated based on the abnormal data in the real-time price data and the operating time data; Detect whether there are any abnormal data in the real-time audio recording data, the vehicle driving data, and the real-time traffic data; If there are any abnormal data in the real-time price data and the operating time data, a second risk quantification value is calculated based on the abnormal data in the real-time recording data, the vehicle driving data and the real-time traffic data. The first risk quantification value and the second risk quantification value are obtained to determine the current operational risk status of the ride-hailing vehicle.

4. The method according to claim 3, characterized in that, The step of obtaining the first risk quantification value and the second risk quantification value to determine the current operational risk status of the ride-hailing vehicle includes: The target risk quantification value is obtained by weighted summation of the first risk quantification value and the second risk quantification value. Obtain the numerical range into which the target risk quantification value falls, and or region the risk level corresponding to the numerical range; Based on the second quantitative risk value, it is determined that the driver of the ride-hailing vehicle is currently engaging in risky driving behavior; The risk level and the driving risk behavior are considered as the operational risk situation.

5. The method according to claim 1, characterized in that, The acquisition of an initial operational control strategy that matches the operational risk situation includes: Extract risk levels and driving risk behaviors from the described operational risk conditions; If the risk level is higher than the preset level, then obtain the order intervention strategy corresponding to the risk level; Based on the aforementioned driving risk behaviors, determine the corresponding driving intervention strategies; The initial operation control strategy is generated based on the order intervention strategy and the driving intervention strategy.

6. The method according to claim 1, characterized in that, The adjustment of the initial operation control strategy based on the overall operational pressure in the area where the ride-hailing service is located to obtain the target operation control strategy includes: Analyze the overall operational pressure to determine the current operational demand in the area where the ride-hailing service is located; Based on the operational needs, determine the adjustable sub-control strategies in the initial operational control strategy, as well as the priorities corresponding to the control strategies; The adjustable sub-control strategies are adjusted according to the priority to obtain the adjusted sub-control strategies, and the changing trend of the overall operational pressure is obtained. By utilizing the aforementioned trend, the execution trigger conditions and execution time of the adjusted sub-control strategy are determined, thus obtaining the target operation control strategy.

7. The method according to claim 1, characterized in that, The method further includes: After managing the ride-hailing vehicles in accordance with the target operation and management strategy, obtain the corresponding management feedback. If the feedback from the control measures indicates that the operational risk of the ride-hailing vehicle has decreased, then other ride-hailing vehicles with the same risk in the same area as the ride-hailing vehicle will be subject to synchronized operational control in accordance with the target operational control strategy.

8. A risk detection device for ride-hailing operations, characterized in that, The device includes: The data acquisition module is used to acquire real-time order data corresponding to the real-time orders of ride-hailing vehicles and trip perception data associated with the real-time orders; The analysis module is used to analyze the real-time order data and the trip perception data to obtain the current operational risk status of the ride-hailing vehicle; The acquisition module is used to acquire an initial operational control strategy that matches the operational risk situation. The adjustment module is used to adjust the initial operation control strategy based on the overall operational pressure of the area where the ride-hailing vehicle is located, to obtain the target operation control strategy, and to control the ride-hailing vehicle in accordance with the target operation control strategy.

9. A computer device, characterized in that, include: A memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, the processor executing the computer instructions to perform the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the method of any one of claims 1 to 7.

Images