Dynamic berth distribution method and system applied to unmanned bus digital station

Abstract

The invention relates to a dynamic parking space distribution method and system applied to an unmanned bus digital station, and the method comprises the steps: W1, obtaining the data information of the driving state parameters of a vehicle in real time in the operation process of the unmanned bus digital station, and collecting the data information of the parking space use state parameters in a fixed time; and W2, based on the data information of the parking space use state parameters, constructing an ARIMA time sequence prediction model optimized and improved by a bat algorithm, predicting the demand state of the parking space, and obtaining the predicted data information of the demand state of the parking space. According to the invention, berths are dynamically and reasonably distributed, the berth utilization rate and the bus operation efficiency are improved, unmanned buses are accurately parked, potential safety hazards are eliminated, and the travel experience of passengers is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of driverless buses, and in particular to a dynamic berth allocation method and system for a digital platform of a driverless bus. Background Art

[0002] With the rapid development of driverless technology, driverless buses have gradually become an important development direction of urban public transportation. However, during the operation of driverless buses, there are many problems in berth allocation and docking guidance at bus stops. Currently, the berth planning of traditional bus stops is mostly in a static fixed mode and cannot be reasonably adjusted according to dynamic factors such as real-time traffic conditions and the running status of bus vehicles. When buses arrive at the station concentratedly during peak hours, there is likely to be a shortage of berths, resulting in vehicles queuing up and waiting, or even the phenomenon of vehicles parking randomly, seriously affecting the bus operation efficiency and the traffic order around the station; while during off-peak hours, there may be a waste of idle berths. At the same time, for driverless buses, there is a lack of a precise docking guidance mechanism, making it difficult to ensure that vehicles accurately dock at the designated berths, affecting the convenience and safety of passengers getting on and off the bus.

[0003] In the prior art, Chinese Patent (Application No.: 202311735848.4, Publication No.: CN 117649756A) discloses a real-time allocation method, system, device and medium for adaptive berths at a station, which generates and guides the incoming vehicles to dock through a guidance information instruction based on bus information, berth occupancy status, and berth allocation principles. Facing complex and changeable traffic conditions, such as traffic congestion, sudden road construction, etc., this solution may not be able to quickly and accurately adjust the berth allocation strategy to meet the actual needs. The implementation highly depends on the normal operation of devices such as the acquisition module and the allocation and guidance module. If problems such as device failures and signal transmission interruptions occur, it may lead to deviations or even failures in berth allocation and guidance, and there is a lack of interaction with passengers and other traffic systems: in terms of interaction with passengers, it does not consider providing passengers with real-time berth allocation information to guide passengers waiting for the bus; there are deficiencies in interaction with other traffic systems (such as urban traffic management systems, intelligent bus dispatching systems, etc.), making it difficult to achieve traffic coordination and optimization on a larger scale.

[0004] In the prior art, a Chinese patent (application number: 201710635814.6, publication number: CN 107452215B) discloses a method for allocating berths for buses to stop at stations and guiding passengers, as well as an intelligent bus platform. 1) Use marking lines to delimit the berths at the stop, use infrared obstacle avoidance sensors to detect the vacancy of each berth in the stop, and use infrared remote control devices to obtain information that the bus is about to arrive at the station; based on the above information, dynamically allocate the vacant berths at the stop by applying the berth allocation principle. 2) Comprehensively consider the time required for passengers to reach the corresponding waiting position after learning the vehicle stop information, and determine the advance amount of the predicted time of the vehicle arrival information; according to the dynamic allocation result, use digital tubes and voice broadcast devices to guide the bus and passengers in real time, so that they arrive at the corresponding waiting area of the berth where the bus is about to arrive in advance and prepare to board. This solution highly depends on hardware devices such as infrared obstacle avoidance sensors and infrared remote control devices. If these devices malfunction, such as sensor failure or abnormal signal reception of the remote control device, it may lead to inaccurate detection of the vacancy of the berth and failure to obtain the vehicle arrival information, and further cause the entire berth allocation and guidance system to fail to operate properly. Infrared devices are easily interfered by environmental factors. For example, strong light, bad weather (such as rain, snow, fog, etc.) may affect the signal transmission and detection accuracy of infrared obstacle avoidance sensors and infrared remote control devices, reducing the reliability and stability of the system. Summary of the Invention

[0005] In view of the above deficiencies of the prior art, the present invention provides a dynamic berth allocation method and system for an unmanned bus digital platform, which not only realizes the dynamic and reasonable allocation of berths, improves the berth utilization rate and bus operation efficiency, but also realizes the precise docking of unmanned buses, eliminates potential safety hazards, and improves the passenger travel experience.

[0006] In order to achieve the above object and other related objects, the technical solution provided by the present invention is as follows: A dynamic berth allocation method for an unmanned bus digital platform, the method comprising: W1. During the operation of the unmanned bus digital platform, real-time obtain the data information of the driving state parameters of the vehicle, and collect the data information of the berth usage state parameters within a fixed time. W2. Based on the data information of the berth usage state parameters, construct an ARIMA time series prediction model optimized and improved by the bat algorithm to predict the demand state of the berths, and obtain the data information of the predicted demand state of the berths. W3. Based on the data information of the driving state parameters of the vehicle and the data information of the berth usage state parameters, construct an improved queuing theory M / M / C model to calculate the average queue length and average waiting time of each berth at present, and obtain the data information of the average queue length and average waiting time of each berth at present. W4. Based on the data information of the current average queue length and average waiting time of each berth, and the data information of the predicted demand status of the berths, an improved sparrow optimization algorithm based on the attenuation factor and dynamic learning is used to optimize the berth allocation, and the data information of the dynamic allocation of the platform berths is obtained.

[0007] Further, in step W2, constructing an ARIMA time series prediction model optimized by the bat algorithm to predict the demand status of the berths includes: W21. Based on the data information of the berth usage status parameters, an improved ARIMA model for the berth demand status is constructed, and the model formula is: , where, Δ d is the difference operator, c is a constant parameter, y t is the berth demand status at time t, y t-i is the berth demand status at time t - i, m and n are positive integers, α i is the autoregressive mean, β j is the moving average, σ t is the white noise sequence at time t, σ t-j is the white noise sequence at time t - j. Initialize the autoregressive mean and the moving average to obtain the initialized autoregressive mean and moving average; W22. Based on the initialized autoregressive mean and moving average, initialize each parameter of the bat population, including the population size, the maximum number of iterations, the upper bound of the parameters, and the lower bound of the parameters, to obtain the data information of the initialized bat population; W23. Based on the data information of the initialized bat population, establish an optimization iteration function Q for the population, , where, x is the data information of the initialized bat population, η1, η2, and η3 are the error factors of the population iteration. Optimize the initialized autoregressive mean and moving average to obtain the optimized autoregressive mean and moving average; W24. Based on the optimized autoregressive mean and moving average, obtain an optimized improved ARIMA time series prediction model. Input the data information of the berth usage status parameters to predict the demand status of the berths, and obtain the data information of the predicted demand status of the berths.

[0008] Further, the error factors η1, η2, and η3 of the population iteration are , , , Among them, x is the data information of the bat population after initialization.

[0009] Furthermore, in step W3, the calculation of the current average queue length and average waiting time of each berth by constructing the improved queuing theory MMC model includes: W31. Based on the data information of the driving state parameters of the vehicle and the data information of the usage state parameters of the parking space, perform normalization processing to obtain the data information of the driving state parameters of the vehicle and the data information of the usage state parameters of the parking space after normalization processing; W32. Based on the data information of the driving state parameters of the vehicle and the data information of the usage state parameters of the parking space after normalization processing, establish the current average queue length function L p and the average waiting time function U, , , where z1 is the data information of the driving state parameters of the vehicle after normalization processing, z2 is the data information of the usage state parameters of the parking space, and λ1, λ2, and λ3 are weight coefficients; W33. Based on the current average queue length function L p and the average waiting time function U of each berth, calculate the current average queue length and average waiting time of each berth to obtain the data information of the current average queue length and average waiting time of each berth.

[0010] Furthermore, the constraint conditions for the weight coefficients λ1, λ2, and λ3 are .

[0011] Furthermore, in step W4, the optimization of berth allocation by using the improved sparrow optimization algorithm based on the attenuation factor and dynamic learning includes: W41. Based on the data information of the current average queue length and average waiting time of each berth and the data information of the predicted demand state of the parking space, initialize the sparrow population, determine the population parameters and the maximum number of iterations, and obtain the data information of the initialized sparrow population; W42. Based on the data information of the initialized sparrow population, establish the fitness function S of the population individuals based on the attenuation factor, , Among them, h is the data information of the initialized sparrow population, μ1, μ2 and μ3 are attenuation factors, and the fitness values ​​of the individuals in the sparrow population are calculated to obtain the data information of the fitness values ​​of the individuals in the sparrow population; W43. Based on the data information of the fitness value of the sparrow population individuals, establish the objective function G, , Among them, g is the data information of the fitness value of the sparrow population individuals, ρ1, ρ2 and ρ3 are dynamic learning factors, and the parking space allocation is optimized to obtain the data information of the dynamic allocation of platform parking spaces.

[0012] Furthermore, the dynamic learning factors ρ1, ρ2 and ρ3 are, , , , Among them, g is the data information of the fitness value of the individual sparrow population.

[0013] Furthermore, the constraint function f of the attenuation factors μ1, μ2 and μ3 is, , , The constraint function f has a value range of (0, 1).

[0014] In order to achieve the above-mentioned purpose and other related purposes, the present invention also provides a dynamic berth allocation system applied to an unmanned bus digital station, including a computer device, which is programmed or configured to execute any one of the steps of the dynamic berth allocation method applied to an unmanned bus digital station.

[0015] In order to achieve the above-mentioned purpose and other related purposes, the present invention also provides a computer-readable storage medium, which stores a computer program programmed or configured to execute any one of the methods for dynamic berth allocation applied to unmanned bus digital stations.

[0016] The present invention has the following positive effects: 1. The present invention constructs a dynamic berth allocation mechanism, uses the perception layer to collect data in real time, analyzes it through the data processing layer, and combines the decision control layer with the algorithm to achieve dynamic and reasonable allocation of berths, thereby improving berth utilization and public transportation operation efficiency.

[0017] 2. For driverless buses, the existing technology lacks a precise docking guidance mechanism, making it difficult to ensure that the vehicle accurately docks at the designated berth, which affects the convenience and safety of passengers getting on and off the bus. This solution relies on the decision-making control layer and the execution layer of the system. Based on the dynamic berth allocation results, it uses path planning algorithms to generate precise docking guidance instructions, and through berth indication devices and the vehicle's autonomous driving control system, realizes the precise docking of driverless buses, eliminates potential safety hazards, and improves the passenger travel experience.

[0018] 3. This invention eliminates the obstacles in multi-source data processing and collaborative decision-making. The operation of buses involves multi-source data such as platform berth status, traffic flow, and vehicle operation. Traditional methods are difficult to effectively integrate and process these data and cannot provide comprehensive support for decision-making. This solution introduces a cloud platform as the data storage and operation center, realizes centralized data management and efficient operation, supports the operation of complex algorithms, and interacts with external systems to obtain more information, ensuring the scientificity and accuracy of system decision-making and optimizing the bus operation decision-making process. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a schematic flow chart of the method of the present invention; Figure 2 is a schematic flow chart of constructing an improved ARIMA time series prediction model optimized by the bat algorithm of the present invention; Figure 3 is a schematic flow chart of constructing an improved queuing theory MMC model of the present invention; Figure 4 is a schematic flow chart of an improved sparrow optimization algorithm based on attenuation factor and dynamic learning of the present invention; Figure 5 is a schematic diagram of the scenario of berth allocation of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] The following describes exemplary embodiments of the present disclosure, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for clarity and conciseness, the description below omits the description of well-known functions and structures.

[0021] Embodiment 1: As Figure 1 or Figure 5 shown, a method for dynamic berth allocation of a digital platform for driverless buses, the method includes: W1. During the operation of the digital platform of the driverless bus, real-time data information of the driving state parameters of the vehicle is obtained, and data information of the berth usage state parameters within a fixed time is collected; W2. Based on the data information of the parking space usage status parameters, construct an ARIMA time series prediction model optimized and improved by the bat algorithm to predict the demand status of parking spaces, and obtain the data information of the predicted demand status of parking spaces; W3. Based on the data information of the driving status parameters of the vehicle and the data information of the parking space usage status parameters, construct an improved queuing theory MMC model to calculate the current average queue length and average waiting time of each berth, and obtain the data information of the current average queue length and average waiting time of each berth; W4. Based on the data information of the current average queue length and average waiting time of each berth and the data information of the predicted demand status of the parking spaces, use an improved sparrow optimization algorithm based on the attenuation factor and dynamic learning to optimize the berth allocation, and obtain the data information of the dynamic allocation of the platform parking spaces.

[0022] In this embodiment, as Figure 2 shown, in step W2, the construction of the ARIMA time series prediction model optimized and improved by the bat algorithm to predict the demand status of parking spaces includes: W21. Based on the data information of the parking space usage status parameters, construct an improved ARIMA model for the demand status of parking spaces. The model formula is: , where Δ d is the difference operator, c is a constant parameter, y t is the demand status of the parking space at time t, y t-i is the demand status of the parking space at time t - i, m and n are positive integers, α i is the autoregressive mean, β j is the moving average, σ t is the white noise sequence at time t, σ t-j is the white noise sequence at time t - j. Initialize the autoregressive mean and the moving average to obtain the initialized autoregressive mean and moving average; W22. Based on the initialized autoregressive mean and moving average, initialize each parameter of the bat population, including the population size, the maximum number of iterations, the upper bound and the lower bound of the parameters, and obtain the data information of the initialized bat population; W23. Based on the data information of the initialized bat population, establish an optimization iteration function Q for the population, , Among them, x is the data information of the bat population after initialization, and η1, η2, and η3 are the error factors of population iteration, which optimize the initialized autoregressive average and moving average to obtain the optimized autoregressive average and moving average. W24. Based on the optimized autoregressive average and moving average, an optimized improved ARIMA time series prediction model is obtained. The data information of the parking space usage status parameter is input to predict the demand status of the parking space, and the data information of the predicted demand status of the parking space is obtained.

[0023] In this embodiment, the error factors η1, η2, and η3 of population iteration are , , , where x is the data information of the bat population after initialization.

[0024] In this embodiment, as Figure 3 shown, in step W3, the construction of the improved queuing theory MMC model to calculate the current average queue length and average waiting time of each berth includes: W31. Based on the data information of the vehicle driving state parameter and the data information of the parking space usage status parameter, normalization processing is performed to obtain the data information of the normalized vehicle driving state parameter and the data information of the parking space usage status parameter; W32. Based on the data information of the normalized vehicle driving state parameter and the data information of the parking space usage status parameter, the current average queue length function L p and average waiting time function U of each berth are established. , , where z1 is the data information of the normalized vehicle driving state parameter, z2 is the data information of the parking space usage status parameter, and λ1, λ2, and λ3 are weight coefficients; W33. Based on the current average queue length function L p and average waiting time function U of each berth, the current average queue length and average waiting time of each berth are calculated to obtain the data information of the current average queue length and average waiting time of each berth.

[0025] In this embodiment, the constraint conditions of the weight coefficients λ1, λ2, and λ3 are .

[0026] In this embodiment, the station-vehicle collaborative management platform, according to the received berth allocation information, scheduling instructions and docking instruction information, displays the guiding information of the driverless vehicle to the platform passengers and station staff through the platform information display screen and the mobile software (APP or WeChat mini-program) (such as bus A will dock at berth C soon). This enables passengers and station staff to understand the docking situation of the driverless bus after it enters the station, facilitating boarding and organizing related vehicle operations.

[0027] The V2X OBU on the driverless bus transmits the received scheduling instructions and precise docking instructions to the autonomous driving controller and the intelligent cockpit controller through the in-vehicle CAN network; The intelligent cockpit controller publishes the information that the driverless bus is about to park and the berth information to the passengers in the vehicle through the in-vehicle display screen and voice broadcast, so that the passengers can get off the bus in an orderly manner after the bus stops; The autonomous driving controller adjusts information such as vehicle speed and steering angle according to the parameters in the above instructions, and precisely docks at the allocated berth. After docking is completed, the "docking completed" status information is uploaded to the station-vehicle collaborative management platform through the V2X OBU; and then forwarded to the information display screen and the mobile software (APP or WeChat mini-program) for display to passengers and staff.

[0028] Embodiment 2: On the basis of an application of the dynamic berth allocation method for a driverless bus digital platform in Embodiment 1, the present invention will be further described and explained below.

[0029] As Figure 1 or Figure 5 shown, an application of the dynamic berth allocation method for a driverless bus digital platform, the method includes: W1. During the operation of the driverless bus digital platform, the data information of the vehicle's driving state parameters is obtained in real time, and the data information of the berth usage state parameters within a fixed time is collected; W2. Based on the data information of the berth usage state parameters, a bat algorithm-optimized and improved ARIMA time series prediction model is constructed to predict the demand state of the berths, and the data information of the predicted demand state of the berths is obtained; W3. Based on the data information of the vehicle's driving state parameters and the data information of the berth usage state parameters, an improved queuing theory M / M / C model is constructed to calculate the current average queue length and average waiting time of each berth, and the data information of the current average queue length and average waiting time of each berth is obtained; W4. Based on the data information of the current average queue length and average waiting time of each berth, and the data information of the predicted demand status of the berths, an improved sparrow optimization algorithm based on the attenuation factor and dynamic learning is used to optimize the berth allocation, and the data information of the dynamic allocation of the platform berths is obtained.

[0030] In this embodiment, as Figure 4 shown, in step W4, the use of the improved sparrow optimization algorithm based on the attenuation factor and dynamic learning to optimize the berth allocation includes: W41. Based on the data information of the current average queue length and average waiting time of each berth, and the data information of the predicted demand status of the berths, initialize the sparrow population, determine the population parameters and the maximum number of iterations, and obtain the data information of the initialized sparrow population; W42. Based on the data information of the initialized sparrow population, establish a fitness function S of the population individuals based on the attenuation factor , where h is the data information of the initialized sparrow population, and μ1, μ2, and μ3 are attenuation factors, and calculate the fitness values of the sparrow population individuals to obtain the data information of the fitness values of the sparrow population individuals; W43. Based on the data information of the fitness values of the sparrow population individuals, establish an objective function G , where g is the data information of the fitness values of the sparrow population individuals, and ρ1, ρ2, and ρ3 are dynamic learning factors, and optimize the berth allocation to obtain the data information of the dynamic allocation of the platform berths.

[0031] In this embodiment, the dynamic learning factors ρ1, ρ2, and ρ3 are , , , where g is the data information of the fitness values of the sparrow population individuals.

[0032] In this embodiment, the constraint function f of the attenuation factors μ1, μ2, and μ3 is , , where the value range of the constraint function f is (0, 1).

[0033] In this embodiment, a multi-objective decision-making optimization algorithm is used for berth allocation. The specific formula is as follows: maxZ = w1×f1 + w2×f2 + w3×f3 + w4×f4, wherein, is the weight for minimizing the average waiting time of vehicles; is the weight for maximizing the berth utilization rate; is the weight for minimizing the total transfer time of passengers; is the weight for minimizing the energy consumption; in this solution, the values are as follows: w1 = 0.3, w2 = 0.2, w3 = 0.3, w4 = 0.2; where: f1 is a measurement function for the waiting time of each bus vehicle waiting for an inbound berth; such as the predicted waiting time t of bus A A1 , the predicted waiting time t of bus B B1 , and the predicted waiting time t of bus C C1 , then , where: f2 is a measurement function based on the proportion of berth usage duration; such as for berths A, B, and C, the usage durations in the past period (the value in this solution is 15 min) are = T A , T B , T C , then for berth A: , And so on to calculate the values of f2 for other berths and conduct a comprehensive evaluation; where: f3 is an evaluation function combining the number of passengers and the transfer distance; such as the number of passengers on driverless bus A is n A ; the transfer distance is d A , then for driverless bus A: , And so on to calculate the values of f3 for other vehicles; where: f4 is an evaluation function based on the remaining battery power and remaining driving mileage of the vehicle; such as the remaining battery power e of driverless bus A A , and the remaining driving distance is l A ; then for driverless bus A: , Similarly, calculate the values of f4 for other vehicles; Use the genetic algorithm to solve this multi-objective function; the fitness function is: ; After the MEC runs the genetic algorithm, perform selection, crossover, and mutation operations on the chromosomes. After m rounds of iterative calculation (the value in this solution is 2000), finally obtain a conclusion such as "Assign berth C to bus A". The MEC uploads the berth allocation result to the station-vehicle collaborative management platform through the V2X RSU.

[0034] Based on the A* algorithm, plan the docking paths for each driverless bus vehicle that has been assigned a parking space; the heuristic function is: h(n)=α×d(n,G)+β×t(n)+γ×e(n), In this solution: the values of ɑ, β, and γ are 0.4, 0.3, and 0.3; Take the example of driverless bus A being assigned to berth C: d(A,C) is the Euclidean distance from driverless bus A to berth C; t(A) is the estimated driving time for driverless bus A to travel to berth C: , e(A) is the remaining power of driverless bus A; After the path is generated, use the B-spline curve for smoothing processing. The control point equation is: , u∈[0,1], In this solution, starting from the vehicle dynamics characteristics of the driverless bus vehicle, select 5 control points P0 - P4 on the path; calculate the path points corresponding to different μ values through the above B-spline curve; combine the real-time state of the driverless bus to generate adjustment instructions such as speed (e.g., reducing to x km / h at a distance of m meters from the berth) and steering angle (e.g., turning left by θ degrees) precise docking instructions, and upload the above scheduling instructions and precise docking instructions to the station-vehicle collaborative control system; at the same time, transmit the above instructions directly to the V2X OBU on the driverless bus vehicle through the V2X RSU.

[0035] In this embodiment, the present invention provides an application for a dynamic berth allocation system of a driverless bus digital platform, including a computer device, which is programmed or configured to execute the steps of any one of the methods for applying the dynamic berth allocation of a driverless bus digital platform.

[0036] In this embodiment, the present invention provides a computer-readable storage medium, on which a computer program is stored that is programmed or configured to execute the method for applying the dynamic berth allocation of a driverless bus digital platform according to any one of the above.

[0037] Any reference to memory, storage, database, or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), among others.

[0038] In this embodiment, taking an unmanned bus stop in a certain city as an example, the stop has 10 parking spaces, and the average daily service vehicles are about 200 times. By deploying the dynamic berth allocation method proposed in the present invention, the following effects are achieved: Data collection and processing: The system collects data on vehicle driving status and parking space usage status once every minute to form a continuous time series. After preprocessing, the data is input into the ARIMA prediction model. Demand prediction: Based on the ARIMA model trained with historical data and optimized by the bat algorithm, the demand change trend of each parking space within the next 1 hour is successfully predicted. The prediction results show that the demand is the largest during the morning peak period (7:00 - 9:00), reaching more than 80%. Queue status analysis: The improved M / M / C model calculates that the average queue length of each berth is 2.5 vehicles, and the average waiting time is 3 minutes. These indicators provide important references for subsequent optimization. Berth allocation optimization: The improved sparrow optimization algorithm dynamically adjusts the berth allocation plan according to the prediction results and the current status. Experiments show that the optimized plan reduces the average waiting time of vehicles by 40% and increases the utilization rate of parking spaces by 25%.

[0039] Comparison ratio and application examples: To verify the effect of the present invention, three comparison schemes are selected for testing: the traditional first-come-first-served strategy, the optimization scheme based on the genetic algorithm, and the dynamic allocation method proposed in the present invention. The test results are shown in the following table: 。

[0040] The application example shows that at a large transportation hub platform, after adopting the method of the present invention, the vehicle queue length during the peak period is shortened from the original 10 vehicles to 4 vehicles, and the passenger satisfaction is significantly improved. In addition, the system also supports multi-platform collaborative scheduling, further improving the overall operation efficiency.

[0041] The present invention proposes an innovative dynamic berth allocation method for digital platforms of driverless buses, which has the following advantages: for the first time, it combines the ARIMA model optimized by the bat algorithm, the improved queuing theory MMC model and the sparrow optimization algorithm based on the attenuation factor to form a complete solution. Through practical application verification, this method significantly reduces the vehicle waiting time, improves the utilization rate of berths, and has good popularization value. The system architecture is flexible, can adapt to driverless bus platforms of different scales and types, and supports function expansion and performance optimization. Objective evaluation shows that the present invention performs excellently in terms of technical solution design, algorithm innovation and actual application effects, providing strong support for the intelligent management of driverless bus platforms.

[0042] In summary, the present invention not only realizes the dynamic and reasonable allocation of berths, improves the berth utilization rate and bus operation efficiency, but also realizes the precise docking of driverless buses, eliminates potential safety hazards, and improves the passenger travel experience.

[0043] The above specific embodiments do not constitute a limitation to the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. A dynamic berth allocation method for digital platforms of driverless buses, characterized in that, The method includes: W1. During the operation of the digital platform of driverless buses, the data information of the driving state parameters of the vehicle is obtained in real time, and the data information of the usage state parameters of the parking spaces within a fixed time is collected; W2. Based on the data information of the usage state parameters of the parking spaces, an ARIMA time series prediction model optimized and improved by the bat algorithm is constructed to predict the demand state of the parking spaces, and the data information of the predicted demand state of the parking spaces is obtained; W3. Based on the data information of the driving state parameters of the vehicle and the data information of the usage state parameters of the parking spaces, an improved queuing theory MMC model is constructed to calculate the current average queue length and average waiting time of each berth, and the data information of the current average queue length and average waiting time of each berth is obtained; W4. Based on the data information of the current average queue length and average waiting time of each berth and the data information of the predicted demand state of the parking spaces, an improved sparrow optimization algorithm based on the attenuation factor and dynamic learning is used to optimize the berth allocation, and the data information of the dynamic allocation of the platform parking spaces is obtained.

2. The dynamic berth allocation method for the digital platform of driverless buses according to claim 1, characterized in that, In step W2, the construction of the ARIMA time series prediction model optimized and improved by the bat algorithm to predict the demand state of the parking spaces includes: W21. Based on the data information of the usage state parameters of the parking spaces, an improved ARIMA model for the demand state of the parking spaces is constructed, and the model formula is: ， where, Δ d is the difference operator, c is a constant parameter, y t is the parking space demand status at time t, y t-i is the parking space demand status at time t - i, m and n are positive integers, α i is the autoregressive average, β j is the moving average, σ t is the white noise sequence at time t, σ t-j is the white noise sequence at time t - j. Initialize the autoregressive average and the moving average to obtain the initialized autoregressive average and moving average; W22. Based on the initialized autoregressive mean and moving average, each parameter of the bat population is initialized, including the population size, the maximum number of iterations, the upper bound and the lower bound of the parameters, and the data information of the initialized bat population is obtained; W23. Based on the data information of the initialized bat population, an optimization iteration function Q of the population is established, ， where x is the data information of the initialized bat population, and η1, η2, and η3 are the error factors of population iteration, and the initialized autoregressive mean and moving average are optimized to obtain the optimized autoregressive mean and moving average; W24. Based on the optimized autoregressive mean and moving average, an optimized improved ARIMA time series prediction model is obtained, the data information of the usage state parameters of the parking spaces is input, and the demand state of the parking spaces is predicted to obtain the data information of the predicted demand state of the parking spaces.

3. The dynamic berth allocation method for the digital platform of driverless buses according to claim 2, characterized in that: The error factors η1, η2, and η3 of the population iteration are ， ， ， where x is the data information of the initialized bat population.

4. The dynamic berth allocation method for an unmanned bus digital platform according to claim 1, characterized in that, In step W3, the construction of the improved queuing theory MMC model to calculate the current average queue length and average waiting time of each berth includes: W31. Based on the data information of the driving state parameters of the vehicle and the data information of the usage state parameters of the parking spaces, normalization processing is performed to obtain the data information of the normalized driving state parameters of the vehicle and the data information of the usage state parameters of the parking spaces; W32. Based on the data information of the vehicle driving state parameters and the data information of the parking space usage state parameters after the normalization process, establish the current average queue length function L of each berth p and the average waiting time function U ， ， where z1 is the data information of the normalized driving state parameters of the vehicle, z2 is the data information of the usage state parameters of the parking spaces, and λ1, λ2, and λ3 are weight coefficients; W33. Based on the current average queue length function L of each berth p and the average waiting time function U, the current average queue length and average waiting time of each berth are estimated to obtain the data information of the current average queue length and average waiting time of each berth.

5. The dynamic berth allocation method for digital platforms of driverless buses according to claim 4, wherein: The constraint conditions for the weight coefficients λ1, λ2, and λ3 are as follows: 。 6. The dynamic berth allocation method for an unmanned bus digital platform according to claim 1, wherein, In step W4, the use of an improved sparrow optimization algorithm based on an attenuation factor and dynamic learning to optimize berth allocation includes: W41. Based on the data information of the current average queue length and average waiting time of each berth, and the data information of the predicted demand status of the berth positions, initialize the sparrow population, determine the population parameters and the maximum number of iterations, and obtain the data information of the initialized sparrow population. W42. Based on the data information of the initialized sparrow population, establish a fitness function S for the population individuals based on the attenuation factor. ， where h is the data information of the initialized sparrow population, and μ1, μ2, and μ3 are attenuation factors, and calculate the fitness values of the sparrow population individuals to obtain the data information of the fitness values of the sparrow population individuals. W43. Based on the data information of the fitness values of the sparrow population individuals, establish an objective function G. ， where g is the data information of the fitness values of the sparrow population individuals, and ρ1, ρ2, and ρ3 are dynamic learning factors, and optimize the berth allocation to obtain the data information of the dynamic allocation of the platform berth positions.

7. The dynamic berth allocation method for the digital platform of driverless buses according to claim 6, wherein: The dynamic learning factors ρ1, ρ2, and ρ3 are as follows: ， ， ， where g is the data information of the fitness values of the sparrow population individuals.

8. The dynamic berth allocation method for the digital platform of driverless buses according to claim 6, characterized in that: The constraint function f for the attenuation factors μ1, μ2, and μ3 is as follows: ， ， where the value range of the constraint function f is (0, 1).

9. A dynamic berth allocation system for a digital platform of driverless buses, including computer equipment, characterized in that, The computer device is programmed or configured to execute the steps of the method for dynamic berth allocation of an unmanned bus digital platform according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program programmed or configured to execute the steps of the method for dynamic berth allocation of an unmanned bus digital platform according to any one of claims 1 to 8.

Images