A method for dynamic berth guidance in medium-capacity bus corridors

Abstract

The present invention discloses a method for dynamic berth guidance in a medium-capacity bus corridor, which is characterized by comprising the following steps: guiding vehicle judgment; predicting the travel time t of the guiding vehicle j based on a trained SVM model; aj ; Predict the berth occupancy of platform s; Guide berth allocation; Publish berth allocation information on the vehicle side and the platform side. The dynamic berth guidance proposed in the present invention is aimed at the operation scenario of multi-line buses in medium-capacity corridors. Based on the parking characteristics of vehicles in multi-berth bus stations, an arrival time prediction model and a bus stop time calculation model for buses in the corridor are constructed. Based on parameters such as the estimated vehicle stop time and the estimated berth occupancy, a dynamic guidance method for multi-berth platforms is proposed, which focuses on bus dynamic guidance, orderly vehicle entry, and real-time fixed-point queuing of passengers, thereby improving the traffic organization order in bus stations, shortening vehicle delays in stations, improving platform traffic capacity, and ensuring the operation efficiency of mainline buses.

Description

Technical Field

[0001] The invention relates to a dynamic berth guidance method for a medium-capacity bus corridor Background Art

[0002] The urban medium-capacity corridor is based on medium-capacity bus lines, coordinated with bus lines with a large repetition coefficient, and optimized and merged ground bus lines to form a large and medium-capacity passenger transport system. It uses public transportation as the main carrying mode to attract intensive travel of passengers along the medium-capacity corridor, thereby effectively diverting the main traffic flow in the city, promoting urban public transportation priority, and attracting travel to shift to a sustainable transportation mode.

[0003] Medium-capacity corridors can further improve the utilization efficiency of bus lanes. However, due to the limited number of parking spaces at the platform and the large number of stopping lines, vehicles from multiple lines arrive at the same time during peak hours. This may cause vehicle overflow and reduce the platform service level. On the other hand, passengers are not sure which parking space to wait for the bus, which may cause passengers to chase the bus after it enters the station, making the platform waiting order chaotic and the waiting time for vehicles long, reducing the bus service level and operating efficiency. Summary of the invention

[0004] The purpose of the present invention is to realize dynamic berth guidance of the main line and other lines in a medium-capacity bus corridor.

[0005] In order to achieve the above object, the technical solution of the present invention is to provide a method for dynamic berth guidance in a medium-capacity bus corridor, wherein the berth numbers in the platform are berth 1, berth 2, ..., berth i in the direction of vehicle travel, and the method is characterized in that the method comprises the following steps:

[0006] Step 1: guiding the vehicle to make a judgment, including the following steps:

[0007] Step 101, determining the main line vehicle guide area and the branch line vehicle guide area corresponding to each station in the medium-capacity corridor;

[0008] Step 102: The backend management system obtains the vehicle's location information through the vehicle terminal, and determines whether the current vehicle has entered the berth guidance section of the current platform. If yes, proceed to step 2; if no, continue to execute this step;

[0009] Step 2: Predict the travel time t of the guided vehicle j based on the trained SVM model aj ;

[0010] Step 3: predicting the berth occupancy of platform s, including the following steps:

[0011] Step 301: The backend management system obtains the current time t in real time s berth occupancy at platform s;

[0012] Step 302: Use the platform stop time BP neural network model to predict the platform stop time t of the vehicle m that has entered the station at each berth pm ;

[0013] Step 303: Calculate the estimated departure time t of each berth vehicle m that has entered the station lm :

[0014] t lm =t s +t pm -t sm

[0015] Among them, t sm is the stop time of vehicle m that has entered the station;

[0016] Step 4: Guide berth allocation

[0017] According to actual needs, select fixed guidance berth allocation rule, semi-fixed guidance berth allocation rule or dynamic guidance berth allocation rule for berth allocation, where:

[0018] The implementation of fixed guidance berth allocation rules includes the following steps:

[0019] Step 4011: Calculate the ratio r of the main line departure frequency to the departure frequency of all stop lines at platform s m :

[0020]

[0021] Among them, f m is the peak departure frequency of the main line, f bi is the peak departure frequency of branch line i stopping at platform s;

[0022] Step 4012: Calculate the number of main line berths n at platform s m :

[0023] n m =ceil(r m *i)

[0024] Where i is the number of berths at platform s, and ceil() is the upward rounding function;

[0025] Step 4013: Calculate the number of berths n for the branch line at platform s b :

[0026] n b =in m

[0027] Step 4014: berth No. 1 of platform s, ..., n mBerth No. is used as a fixed berth for mainline vehicles, n m + berth 1, ..., berth i are fixed berths for feeder vehicles;

[0028] Step 4015: If the vehicle j to be guided is a main line vehicle, determine t aj At time s, berth 1 of platform s, ..., n m Is there any berth available at berth number:

[0029] If yes, assign the largest parking space to vehicle j;

[0030] If there is no parking space, the vehicle's departure time is earlier than t aj , which means that vehicle j needs to wait in line outside the station. At this time, vehicle j is guided for speed or station time, and the travel time t of guided vehicle j is updated again. aj Then, re-execute step 4015;

[0031] Step 4016: If the guided vehicle j is a branch line vehicle, determine the travel time t of the guided vehicle j. aj , platform s n m + berth 1, ..., whether berth i has vacant berths available:

[0032] If yes, assign the largest parking space to vehicle j;

[0033] If there is no parking space, the vehicle's departure time is earlier than t aj , which means that vehicle j needs to wait in line outside the station. At this time, the vehicle speed guidance or station time guidance is carried out, and the travel time t of guided vehicle j is updated again. aj Then, re-execute step 4016;

[0034] The implementation of the semi-fixed guidance allocation rule includes the following steps:

[0035] Step 4021: Berth No. 1 of platform s is used as a fixed berth for mainline vehicles, berth No. i is used as a fixed berth for branchline vehicles, and other berths are used as mobile berths;

[0036] Step 4022: Determine the type of the guided vehicle:

[0037] If the leading vehicle j is the main line vehicle, determine t aj Whether berths 1, ..., i-1 of platform s are available at the moment, if berths 1, ..., k are available, then berth k is allocated to guide vehicle j, otherwise, proceed to step 4025;

[0038] If the leading vehicle j is a branch line vehicle, determine t ajIs berth i of platform s available at the moment? If available, allocate berth i to guided vehicle j; if not available, determine t aj Whether there are available berths at berths 1, ..., i-1 of platform s at the time, if there are available berths, execute step 4023, if not, execute step 4025;

[0039] Step 4023: If berth 1, ..., berth 1, ..., berth k among berth i-1 are available, and k≥1, then berth k is pre-allocated to the guide vehicle j, and the process goes to step 4024; otherwise, the process goes to step 4025;

[0040] Step 4024: Determine whether aj ,t aj +t pj ], whether the number of mainline vehicles arriving is greater than k, where t pj is the platform stop time of the guide vehicle j predicted by the platform stop time BP neural network model: if it is less than k, it is determined to allocate berth k to the guide vehicle j; if it is greater than k, execute step 4025;

[0041] Step 4025: Provide vehicle speed guidance or station time guidance for vehicle j, and re-update t aj , return to step 4022;

[0042] The implementation of dynamic guidance allocation rules includes the following steps:

[0043] Step 4031: Determine t aj Whether there are available berths at platform s at the moment;

[0044] Step 4032: If there is an available berth, allocate the largest berth to vehicle j;

[0045] Step 4033: If there is no available parking space, then the vehicle j is guided in terms of speed or station time, and t is updated again. aj , execute step 4031;

[0046] Step 5: Publish berth allocation information on the vehicle and platform.

[0047] Preferably, step 2 comprises the following steps:

[0048] Step 201, using historical data to establish a training data set of travel time between bus stops, including time period, weather, and vehicle speed;

[0049] Step 202: Select appropriate kernel functions and parameters, and train the SVM model using the training data set;

[0050] Step 203: Use the trained SVM model to predict the travel time t of vehicle jaj .

[0051] Preferably, in step 302, the establishment of the platform stop time BP neural network model includes the following steps:

[0052] Step 3021: Establish a training data set based on historical data, including time period, route name, and headway time with the oncoming vehicles;

[0053] Step 3022: Determine the appropriate number of input variables and hidden nodes through testing, and use the training data set to train the platform stop time BP neural network model;

[0054] Step 3023: Use the trained platform stop time BP neural network model to predict the platform stop time of vehicle m that has entered the station.

[0055] Preferably, in step 5, when the vehicle-mounted terminal publishes the berth allocation information:

[0056] After guiding vehicle j to leave the previous platform, the background management system sends the estimated arrival time of the vehicle and the allocated parking space information of platform s to the vehicle-mounted terminal, and the driver stops at the station according to the allocated parking space information.

[0057] Preferably, in step 5, when the berth allocation information is released at the platform end:

[0058] A passenger information display screen is set up in the platform. The background management system sends the names of all lines that stop at the platform, the estimated arrival time, the estimated berth information, and the car load rate to the passenger information display screen to guide passengers; if the vehicle of the line has not yet been allocated a berth, the estimated berth will be displayed as pending allocation.

[0059] Preferably, in step 5, when the berth allocation information is released at the platform end:

[0060] Each berth in the platform is equipped with a berth information release screen. If the berth has been allocated to a vehicle waiting to enter the station, the background management system will send the name of the route that will stop at the berth and the estimated arrival time to the berth information release screen to guide passenger flow.

[0061] The dynamic berth guidance proposed in the present invention is aimed at the operation scenario of multi-line buses in medium-capacity corridors. Based on the parking characteristics of vehicles in multi-berth bus stations, an arrival time prediction model and a bus stop time calculation model for buses in the corridor are constructed. Based on parameters such as the estimated vehicle stop time and the estimated berth occupancy, a dynamic guidance method for multi-berth platforms with dynamic bus guidance, orderly vehicle entry, and real-time fixed-point queuing of passengers as the core is proposed, thereby improving the traffic organization order in bus stations, shortening vehicle delays in stations, improving platform traffic capacity, and ensuring the operation efficiency of mainline buses. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It shows the contents displayed on the passenger information display screen;

[0063] Figure 2 It shows the contents displayed on the berth information release screen. DETAILED DESCRIPTION

[0064] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall within the scope limited by the appended claims of the application equally.

[0065] Stations in medium-capacity corridors can be divided into the following three types:

[0066] Type A station: the first station that a branch line stops at when it enters a medium-capacity corridor

[0067] Type B station: Mainline and feeder vehicles have formed a fleet in the medium-capacity corridor

[0068] Type C station: no transfer station, stops in sequence

[0069] The present invention takes type A station as the research object and conducts berth guidance research, which specifically includes the following contents:

[0070] 1) Basic assumptions of the problem

[0071] 1. There are dedicated bus lanes in the medium-capacity corridors. Buses are not allowed to overtake in the corridors and must queue up to enter the station;

[0072] 2. The main line is a bus line that runs along the medium-capacity corridor;

[0073] 3. The branch line is a bus line that does not have its first and last stops in the corridor and only runs a certain distance in the corridor;

[0074] 4. The berth numbers in the platform along the direction of vehicle travel are berth 1, berth 2... berth i.

[0075] 2) Model establishment

[0076] Step 1: Guide the vehicle to judge:

[0077] Step 101, determining the main line vehicle guide area and the branch line vehicle guide area corresponding to each station in the medium-capacity corridor;

[0078] Step 102, the backend management system obtains the vehicle's positioning information through the vehicle terminal, and determines whether the current vehicle enters the berth guidance section of the current platform. If yes, proceed to step 2, if not, continue to execute this step.

[0079] Step 2: Guide vehicle j travel time t aj The calculation includes the following steps:

[0080] Step 201, using historical data to establish a training data set of travel time between bus stops, including variables such as time period, weather, and vehicle speed;

[0081] Step 202: Select appropriate kernel functions and parameters, and train the SVM model using the training data set;

[0082] Step 203: Use the trained SVM model to predict the travel time t of vehicle j aj .

[0083] Step 3: Prediction of berth occupancy at platform s, including the following steps:

[0084] Step 301, calculating the estimated vehicle stop time, includes the following steps:

[0085] Step 3011, establishing a training data set based on historical data, including variables such as time period, route name, and headway time with the vehicles on the same day;

[0086] Step 3012: Determine the appropriate number of input variables and hidden nodes through testing, and use the training data set to train the platform stop time BP neural network model;

[0087] Step 3013: Use the trained platform stop time BP neural network model to predict the platform stop time of vehicle m that has entered the station;

[0088] Step 3014: Calculate the estimated departure time of vehicle m that has entered the station based on the platform stop time of vehicle m that has entered the station.

[0089] Step 302, calculating the berth status of platform s, includes the following steps:

[0090] Step 3021: The backend management system obtains the current time t in real time based on the bus positioning information and the video information of the platform video camera. s berth occupancy at platform s;

[0091] Step 3022: Use the platform stop time BP neural network model obtained in step 301 to predict the platform stop time t of vehicle m that has entered berth i. pm ;

[0092] Step 3023: Calculate the estimated departure time t of vehicle m that has entered berth i lm :

[0093] t lm =t s +t pm -t sm

[0094] Among them, t sm is the stop time of vehicle m that has entered the station;

[0095] Step 3024, updating the estimated occupancy status of the platform berth;

[0096] Step 303: Estimated departure time t of guiding vehicle j lj The calculation includes the following steps:

[0097] Step 3031: predict the platform stop time t of vehicle j according to the platform stop time BP neural network model using step 301 pj ;

[0098] Step 3032: Calculate the estimated departure time t of the leading vehicle j lj , t lj =t aj +t pj .

[0099] Step 3033, updating the estimated occupancy status of the platform berth;

[0100] Step 4: Guide berth allocation

[0101] According to actual needs, you can flexibly select any of the following berth allocation rules for berth allocation:

[0102] First) Fixed guide

[0103] This allocation rule fixes the waiting position for passengers and there is no need to monitor the berth allocation in real time. The probability of mainline vehicles waiting outside the station is low, which can ensure the operating efficiency of mainline vehicles. However, the utilization rate of berths on the platform is relatively low.

[0104] Fixed guidance berth allocation scheme is divided into peak allocation scheme and flat peak allocation scheme. The calculation method is the same, taking the calculation of platform z peak berth allocation scheme as an example, including the following steps:

[0105] Step 4011: Calculate the ratio r of the main line departure frequency to the departure frequency of all stop lines at platform s m :

[0106]

[0107] Among them, f m is the peak departure frequency of the main line; fbi is the peak departure frequency of branch line i stopping at platform z;

[0108] Step 4012: Calculate the number of main line berths n at platform s m

[0109] n m =ceil(r m *i)

[0110] Where i is the number of berths at platform s, and ceil() is the upward rounding function;

[0111] Step 4013: Calculate the number of berths n for the branch line at platform s b

[0112] n b =in m

[0113] Step 4014: berth No. 1 of platform s, ..., n m Berth No. is used as a fixed berth for mainline vehicles, n m + berth 1, ..., berth i are fixed berths for feeder vehicles;

[0114] Step 4015: If the vehicle j to be guided is a main line vehicle, determine t aj At time s, berth 1 of platform s, ..., n m Is there any berth available at berth number:

[0115] If yes, assign the largest parking space to vehicle j;

[0116] If there is no parking space, the vehicle's departure time is earlier than t aj , which means that vehicle j needs to wait in line outside the station. At this time, vehicle j is guided for speed or station time, and the travel time t of guided vehicle j is updated again. aj Then, re-execute step 4015;

[0117] Step 4016: If the guided vehicle j is a branch line vehicle, determine the travel time t of the guided vehicle j. aj , berth 1 of platform s, …, n m Are all berths available? m + berth 1, ..., whether berth i has vacant berths available:

[0118] If yes, assign the largest parking space to vehicle j;

[0119] If there is no parking space, the vehicle's departure time is earlier than t aj, which means that vehicle j needs to wait in line outside the station. At this time, the vehicle speed guidance or station time guidance is carried out, and the travel time t of guided vehicle j is updated again. aj Then, re-execute step 4016.

[0120] 2) Semi-fixed guidance

[0121] The allocation rule has a relatively fixed waiting position for passengers, a low probability of mainline vehicles waiting outside the station, and a higher utilization rate of berths in the platform than the fixed guidance rule, including the following steps:

[0122] Step 4021: Berth No. 1 of platform s is used as a fixed berth for mainline vehicles, berth No. i is used as a fixed berth for branchline vehicles, and other berths are used as mobile berths;

[0123] Step 4022: Determine the type of the guided vehicle:

[0124] If the leading vehicle j is the main line vehicle, determine t aj Whether berths 1, ..., i-1 of platform s are available at the moment, if berths 1, ..., k are available, berth k is allocated to guide vehicle j, otherwise, go to step 4025;

[0125] If the leading vehicle j is a branch line vehicle, determine t aj Is berth i of platform s available at the moment? If available, allocate berth i to guided vehicle j; if not available, determine t aj Whether there are available berths at berths 1, ..., i-1 of platform z at the moment, if there are available berths, execute step 4023, if not, execute step 4025;

[0126] Step 4023: If berths 1, ..., k among berths 1, ..., i-1 are available and k ≥ 1, then berth k is pre-allocated to the guide vehicle j and the process goes to step 4024; otherwise, the process goes to step 4025;

[0127] Step 4024: Determine whether aj ,t aj +t pj ] Whether the number of mainline vehicles arriving during the time period is greater than k: if less than k, determine to allocate berth k to the guide vehicle j; if greater than k, execute step 4025;

[0128] Step 4025: Provide vehicle speed guidance or station time guidance for vehicle j, and re-update t aj , return to step 4022.

[0129] 3) Dynamic Guidance

[0130] The allocation rule randomly arranges passengers' waiting positions. Passengers need to pay attention to the berth allocation information in real time to determine their waiting positions. Mainline vehicles have a certain probability of waiting outside the station, but the overall utilization rate of platform berths is relatively high. The following steps are included:

[0131] Step 4031: Determine t aj Whether there are available berths at platform s at the moment;

[0132] Step 4032: If there is an available berth i, and berths 1, ..., i are all available, then berth i is allocated to vehicle j;

[0133] Step 4033: If there is no available parking space, then the vehicle j is guided in terms of speed or station time, and t is updated again. aj , execute step 4031.

[0134] Step 5: Berth allocation information release includes the following steps:

[0135] Step 501: Vehicle-mounted terminal information release:

[0136] After guiding vehicle i to leave the previous platform, the background management system sends the estimated arrival time of the vehicle and the allocated parking space information of platform s to the vehicle terminal, and the driver stops at the station according to the allocated parking space information.

[0137] Step 502: Platform information release

[0138] The platform is equipped with a passenger information display screen. The backend management system sends the names of all routes that stop at the platform, the estimated arrival time, the estimated parking space information, and the car load rate to the passenger information display screen for passenger guidance. If the vehicle on the route has not yet been assigned a parking space, the estimated parking space will be displayed as pending allocation.

[0139] Each berth in the platform is equipped with a berth information release screen. If the berth has been allocated to a vehicle waiting to enter the station, the background management system will send the name of the route that will stop at the berth and the estimated arrival time to the berth information release screen to guide passenger flow.

Claims

1. A method for dynamic berth guidance in a medium-capacity bus corridor, wherein the berths in the platform are numbered in the order of berth 1, berth 2, ..., berth i along the direction of vehicle travel, characterized in that: The following steps are involved: Step 1: guiding the vehicle to make a judgment, including the following steps: Step 101, determining the main line vehicle guide area and the branch line vehicle guide area corresponding to each station in the medium-capacity corridor; Step 102: The backend management system obtains the vehicle's location information through the vehicle terminal, and determines whether the current vehicle has entered the berth guidance section of the current platform. If yes, proceed to step 2; if no, continue to execute this step; Step 2: Predict the travel time t of the guided vehicle j based on the trained SVM model aj ; Step 3: predicting the berth occupancy of platform s, including the following steps: Step 301: The backend management system obtains the current time t in real time s berth occupancy at platform s; Step 302: Use the platform stop time BP neural network model to predict the platform stop time t of the vehicle m that has entered the station at each berth pm ; Step 303: Calculate the estimated departure time t of each berth vehicle m that has entered the station lm : t lm =t s +t pm -t sm Among them, t sm is the stop time of vehicle m that has entered the station; Step 4: Guide berth allocation According to actual needs, select fixed guidance berth allocation rule, semi-fixed guidance berth allocation rule or dynamic guidance berth allocation rule for berth allocation, where: The implementation of fixed guidance berth allocation rules includes the following steps: Step 4011: Calculate the ratio r of the main line departure frequency to the departure frequency of all stop lines at platform s m : Among them, f m is the peak departure frequency of the main line, f bi is the peak departure frequency of branch line i stopping at platform s; Step 4012: Calculate the number of main line berths n at platform s m : n m =ceil(r m *i) Where i is the number of berths at platform s, and ceil() is the upward rounding function; Step 4013: Calculate the number of berths n for the branch line at platform s b : n b =i-n m Step 4014: berth No. 1 of platform s, ..., n m Berth No. is used as a fixed berth for mainline vehicles, n m + berth 1, ..., berth i are fixed berths for feeder vehicles; Step 4015: If the vehicle j to be guided is a main line vehicle, determine t aj At time s, berth 1 of platform s, ..., n m Is there any berth available at berth number: If yes, assign the largest parking space to vehicle j; If there is no parking space, the vehicle's departure time is earlier than t aj , which means that vehicle j needs to wait in line outside the station. At this time, vehicle j is guided for speed or station time, and the travel time t of guided vehicle j is updated again. aj Then, re-execute step 4015; Step 4016: If the guided vehicle j is a branch line vehicle, determine the travel time t of the guided vehicle j. aj , platform s n m + berth 1, ..., whether berth i has vacant berths available: If yes, assign the largest parking space to vehicle j; If there is no parking space, the vehicle's departure time is earlier than t aj , which means that vehicle j needs to wait in line outside the station. At this time, the vehicle is guided for speed or station time, and the travel time t of guided vehicle j is updated again. aj Then, re-execute step 4016; The implementation of the semi-fixed guidance allocation rule includes the following steps: Step 4021: Berth No. 1 of platform s is used as a fixed berth for mainline vehicles, berth No. i is used as a fixed berth for branchline vehicles, and other berths are used as mobile berths; Step 4022: Determine the type of the guided vehicle: If the leading vehicle j is the main line vehicle, determine t aj Whether berths 1, ..., i-1 of platform s are available at the moment, if berths 1, ..., k are available, then berth k is allocated to guide vehicle j, otherwise, proceed to step 4025; If the leading vehicle j is a branch line vehicle, determine t aj Is berth i of platform s available at the moment? If available, allocate berth i to guided vehicle j; if not available, determine t aj Whether there are available berths at berths 1, ..., i-1 of platform s at the time, if there are available berths, execute step 4023, if not, execute step 4025; Step 4023: If berth 1, ..., berth 1, ..., berth k among berth i-1 are available, and k≥1, then berth k is pre-allocated to the guide vehicle j, and the process goes to step 4024; otherwise, the process goes to step 4025; Step 4024: Determine whether aj , t aj +t pj ], whether the number of mainline vehicles arriving is greater than k, where t pj is the platform stop time of the guide vehicle j predicted by the platform stop time BP neural network model: if it is less than k, it is determined to allocate berth k to the guide vehicle j; if it is greater than k, execute step 4025; Step 4025: Provide vehicle speed guidance or station time guidance for vehicle j, and re-update t aj , return to step 4022; The implementation of dynamic guidance allocation rules includes the following steps: Step 4031: Determine t aj Whether there are available berths at platform s at the moment; Step 4032: If there is an available berth, allocate the largest berth to vehicle j; Step 4033: If there is no available parking space, then the vehicle j is guided in terms of speed or station time, and t is updated again. aj , execute step 4031; Step 5: Publish berth allocation information on the vehicle and platform.

2. A method for dynamic berth guidance in a medium-capacity bus corridor as claimed in claim 1, characterized in that: The step 2 comprises the following steps: Step 201, using historical data to establish a training data set of travel time between bus stops, including time period, weather, and vehicle speed; Step 202: Select appropriate kernel functions and parameters, and use the training data set to train the SVM model; Step 203: Use the trained SVM model to predict the travel time t of vehicle j aj .

3. The method for dynamic berth guidance in a medium-capacity bus corridor according to claim 1, characterized in that: In step 302, the establishment of the platform stop time BP neural network model includes the following steps: Step 3021: Establish a training data set based on historical data, including time period, route name, and headway time with the vehicles on the same shift; Step 3022: Determine the appropriate number of input variables and hidden nodes through testing, and use the training data set to train the platform stop time BP neural network model; Step 3023: Use the trained platform stop time BP neural network model to predict the platform stop time of vehicle m that has entered the station.

4. The method for dynamic berth guidance in a medium-capacity bus corridor according to claim 1, characterized in that: In step 5, when the parking space allocation information is released on the vehicle terminal: After guiding vehicle j to leave the previous platform, the background management system sends the estimated arrival time of the vehicle and the allocated parking space information of platform s to the vehicle-mounted terminal, and the driver stops at the station according to the allocated parking space information.

5. The method for dynamic berth guidance in a medium-capacity bus corridor according to claim 1, characterized in that: In step 5, when the berth allocation information is released on the platform side: A passenger information display screen is set up in the platform. The background management system sends the names of all lines that stop at the platform, the estimated arrival time, the estimated berth information, and the car load rate to the passenger information display screen to guide passengers; if the vehicle of the line has not yet been allocated a berth, the estimated berth will be displayed as pending allocation.

6. A method for dynamic berth guidance in a medium-capacity bus corridor as claimed in claim 5, characterized in that: In step 5, when the berth allocation information is released on the platform side: Each berth in the platform is equipped with a berth information release screen. If the berth has been allocated to a vehicle waiting to enter the station, the background management system will send the name of the route that will stop at the berth and the estimated arrival time to the berth information release screen to guide passenger flow.

Images