A method for calculating three-level utilization rate of bus lane reservation and identifying space-time resources

Abstract

The present application relates to the technical field of urban intelligent traffic control, and discloses a method for calculating the three-level utilization rate of reserved bus lanes and identifying space-time resources, aiming to solve the problems of rough resource quantification, fuzzy identification of residual space-time resources, and lack of theoretical basis for reservation of road rights. The steps of the present application are as follows: S1, based on the virtual occupation space of buses, discretize the bus lane into three levels of space, i.e. grid, road section and route, and construct a digital road network; S2, fuse bus GPS trajectory, signal timing and traffic flow data to analyze the space-time law of vehicle operation; S3, calculate the utilization rate of bus lanes step by step; S4, identify the residual space-time resources according to the utilization rate to provide support for dynamic allocation of reservation road rights. The present application realizes fine quantification and dynamic identification of lane resources, guarantees bus priority while improving lane utilization rate, provides data support for reservation of road rights, and is suitable for intelligent control and road right optimization of urban reserved bus priority lanes.

Description

Technical Field

[0001] This invention relates to the field of urban intelligent traffic control and refined management of bus lanes, specifically to a method for calculating the three-level utilization rate and identifying spatiotemporal resources of reserved bus lanes. Background Technology

[0002] Bus lanes are core infrastructure for ensuring priority passage of urban buses, improving the efficiency of public transportation operations, and increasing the utilization rate of road resources. With the development of vehicle-to-everything (V2X) technology, big data, and intelligent traffic management, reservation-based bus lanes have become the mainstream approach to solving the problems of peak-hour congestion, off-peak idleness, and rigid right-of-way allocation. Accurate calculation of bus lane utilization and identification of remaining time and space resources are the core technical prerequisites for realizing dynamic allocation of reserved right-of-way and transforming lanes from "dedicated" to "priority".

[0003] Existing studies generally use single-section statistics or fixed-time-period estimation to measure the utilization rate of bus lanes, failing to establish a systematic hierarchical and quantitative framework. While some scholars have proposed spatial discretization, this approach remains at the segment level, lacking a hierarchical and multi-scale systematic quantitative framework that ranges from micro-level single-vehicle occupancy to meso-level road segments and macro-level routes. This makes it difficult to accurately characterize the resource occupancy differences of bus lanes at different spatial scales. Furthermore, existing spatiotemporal resource identification methods largely rely on static threshold judgments, failing to integrate multi-source dynamic data such as bus GPS trajectories, intersection signal timing, traffic flow, and bus stops and stops. This makes it difficult to accurately reflect the real-time changes in traffic flow and the coupled impact of signal control, resulting in delayed and biased remaining resource identification results, which cannot support refined right-of-way allocation for reserved vehicles.

[0004] Therefore, developing a method for calculating the three-level utilization rate and identifying spatiotemporal resources of reserved bus lanes can solve the prominent problems of low calculation accuracy, poor dynamic adaptability, lack of coordination among the three levels of space, and insufficient basis for right-of-way allocation in existing technologies, and provide key technical support for the efficient operation and dynamic optimization of right-of-way in urban reserved bus lanes. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a three-level utilization rate calculation and spatiotemporal resource identification method for reserved bus lanes. By constructing a three-level spatial quantification system of "grid level - road segment level - route level", it integrates multi-source dynamic traffic data to achieve accurate calculation of utilization rate at each level, and dynamically identifies remaining spatiotemporal resources based on the calculation results. This method aims to solve the technical problems of coarse quantification of bus lane resources, lagging spatiotemporal resource identification, and insufficient basis for reserved right-of-way allocation in existing bus lanes.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a method for calculating the three-level utilization rate and identifying spatiotemporal resources of reserved bus lanes. This method is based on the virtual space occupied by buses, discretizing the bus lane into three levels: grid level, road segment level, and route level. It integrates multi-source data such as bus GPS trajectories, signal timing, traffic flow, and bus departure frequency to analyze the spatiotemporal patterns of vehicle operation, calculates the utilization rate at each level, identifies remaining spatiotemporal resources, and finally outputs a list of spatiotemporal resources that can be directly used for dynamic allocation of reserved right-of-way. The method includes the following steps:

[0007] S1, based on the virtual occupancy space of buses, the bus lane is discretized into three levels of space: grid level, road segment level and route level, to construct a multi-level digital road network and complete topology coding and attribute mapping;

[0008] S2 integrates bus GPS trajectory, signal timing, traffic flow and bus departure frequency data to analyze the spatial and temporal patterns of vehicle operation in three-level spaces.

[0009] S3 calculates the utilization rate at the grid level, road segment level, and route level based on virtual occupied space, critical density, and cross-sectional space occupancy rate.

[0010] S4 identifies remaining spatiotemporal resources based on the three-level utilization rate and outputs the spatiotemporal resource results of dedicated lanes that can be used for reservation right-of-way allocation.

[0011] Further, in step S1, the specific definition of the three-level space is as follows: a grid-level unit is the space P occupied by a single bus, which is the smallest unit for utilization calculation; a road segment-level unit is the bus lane between two adjacent intersections, which is the smallest unit for reserved right-of-way allocation; a route-level unit consists of... Each road segment level unit and It consists of a combination of signal-controlled intersections, covering the entire dedicated bus route.

[0012] The construction of a multi-level digital road network system includes mapping bus lane attributes to grid cells through GIS spatial connections, splitting and generating independently coded road segment-level units with intersections as nodes, integrating road segment-level units with intersection spaces to generate route-level units, and using PostGIS and GeoPandas to realize three-level spatial digital representation, forming a multi-level spatially integrated digital road network system for bus lanes.

[0013] Furthermore, in step S2, analyzing the spatiotemporal patterns of vehicle operation includes matching multi-source data through a spatiotemporal alignment algorithm, removing abnormal trajectory points and identifying the dwell time of bus stops using the DBSCAN clustering algorithm, identifying the spatiotemporal clustering characteristics of traffic flow using kernel density estimation and spatiotemporal clustering algorithms, and analyzing the three-level spatial topology coupling relationship through a weighted directed multi-layer network model to eliminate the interference of intersection queuing and bus stops on the analysis of spatiotemporal patterns.

[0014] The method for removing abnormal GPS trajectory points from buses using the DBSCAN clustering algorithm is as follows:

[0015] DBSCAN is a classic density-based spatial clustering algorithm that primarily uses neighborhood radius. The two core parameters, MinPts and the minimum number of points, enable the identification and removal of noise points.

[0016] 1. First, construct a bus GPS trajectory point dataset, using latitude and longitude coordinates as the feature dimension, calculate the Euclidean distance between different trajectory points, and construct a trajectory point spatial distance matrix;

[0017] 2. Traverse all unvisited bus trajectory points. If the number of trajectory points in the ε-neighborhood of a trajectory point is greater than or equal to MinPts, mark the point as a core point and recursively merge all density-reachable trajectory points starting from the core point to form an effective driving trajectory cluster. If a trajectory point does not belong to any effective cluster and the number of trajectory points in its ε-neighborhood is less than MinPts, it is determined to be an abnormal noise point caused by GPS signal drift, positioning error, or signal loss, and is directly removed from the trajectory dataset.

[0018] 3. After completing clustering and outlier removal, time interpolation and spatial smoothing are performed on the remaining valid trajectory points to ensure the spatiotemporal continuity and accuracy of bus operation trajectories, providing a high-quality data foundation for subsequent analysis of vehicle operation spatiotemporal patterns in the three-level spatial system.

[0019] Then, by using continuous trajectory point speed threshold determination and dwell time statistics, the dwell time of bus stops is identified, separating the dwell time from the travel time, thereby improving the accuracy of vehicle runtime spatial and temporal pattern analysis.

[0020] A weighted directed multilayer network model used to analyze the topological coupling relationships in a three-level spatial environment is denoted as . ,in For a set of nodes, For the set of connected edges, As a weight set, the core is to abstract the three-level spatial units and their relationships into a network topology structure of "node-edge-weight", quantifying the dynamic transmission and coupling influence of micro-grid resource status to meso-level road segments and macro-level routes.

[0021] Specifically, node set It contains three types of nodes: the bottom layer consists of grid-level nodes. (Single grid-level cell), with road segment-level nodes in the middle. (Single road segment level unit), top level is route level node (The entire bus lane) forms a three-layer node architecture; connecting edge sets It is divided into intra-layer edges and inter-layer edges. Intra-layer edges construct directed connections based on spatial adjacency to represent the direction of traffic flow propagation. Inter-layer edges construct subordinate relationships from grid to road segment and from road segment to route to express hierarchical contribution relationships; weight set Multi-dimensional dynamic assignment: edge weights from grid to road segment The proportion of grid length to the total road segment length reflects the contribution of individual grid utilization to the overall road segment utilization; the weight of the connection between the road segment and the route. The proportion of bus traffic flow in a given route segment to the total traffic flow of the entire route represents the intensity of the impact of the segment's operational status on the overall resource utilization level of the route; the weight of adjacent edges within the layer. By combining spatial distance and vehicle speed with normalized values, the propagation resistance of traffic flow between adjacent spatial units is quantified. Based on the constructed weighted directed multilayer network, by calculating topological characteristic indicators such as network degree, betweenness coefficient, and weighted clustering coefficient, and combining them with network propagation dynamics analysis, the congestion or surplus state of the micro-grid can be accurately quantified to progressively propagate and affect the resource utilization level of the meso-level road segments. This identifies the constraint range and influence intensity of local bottleneck segments on the resource allocation of the entire route, and effectively characterizes the coupling interference effect of key nodes such as intersections and bus stops on the distribution of tertiary spatial resources. This eliminates the adverse effects of non-driving conditions such as intersection queuing and bus stops on the analysis of vehicle runtime spatial patterns, providing rigorous topological theoretical support for the collaborative calculation of tertiary utilization rates and the dynamic identification of remaining spatiotemporal resources.

[0022] Furthermore, in step S3, the step-by-step calculation of the three-level utilization rate includes first using the grid level as the smallest unit, combining the bus lane length, virtual occupied space, and average driving speed to calculate the critical density and cross-sectional space occupancy rate, and constructing a line space density model to determine the grid level utilization rate; then, the segment level utilization rate is obtained by weighted summation of the grid level utilization rate; finally, the route level utilization rate is obtained by comprehensive calculation of the segment level utilization rate.

[0023] The virtual occupied space is composed of the bus body length, minimum safe following distance, and bus clearing distance. The length of the virtual occupied space is dynamically adjusted according to traffic density. During high-density periods, the virtual occupied space is shortened to improve the efficiency of reserved vehicle passage, while during low-density periods, it is increased to ensure bus priority passage. High-density periods are defined as periods with traffic density greater than or equal to 70% of congestion density, low-density periods are defined as periods with traffic density less than or equal to 30% of congestion density, and periods with traffic density in between are defined as medium-density periods. When calculating the average vehicle speed across the three spatial levels, stop time separation and three-level segmented modeling techniques are introduced to eliminate intersection queuing and stop interference. Combining traffic flow, density, and delay variance, a dynamic weight allocation is used to construct a vehicle speed calculation model to improve the accuracy of vehicle speed calculation.

[0024] Specifically, the stop time separation technology refers to separating the time spent by buses at stops, such as picking up and dropping off passengers and opening and closing doors, from the total vehicle operation time, and retaining only the pure driving time of the vehicle within the road segment, thereby avoiding interference from non-driving conditions on vehicle speed calculation.

[0025] The three-level segmented modeling technology refers to: segmenting the bus lane into three spatial units at the grid level, road segment level, and route level, calculating the pure driving speed of each segment without interference, and then aggregating them from bottom to top to obtain the average speed at the road segment level and route level, thereby eliminating the impact of inter-segment interference such as intersection queuing and signal delays on speed calculation.

[0026] Based on this, a three-level spatial average vehicle speed calculation model after eliminating interference is constructed, and its mathematical expression is as follows:

[0027]

[0028] In the formula, To eliminate the average speed of vehicles in the three-level space after stopping and queuing interference; The baseline vehicle speed component is calculated based on the traffic flow of the road segment. For density-corrected vehicle speed components calculated based on traffic density; For the delay-corrected speed component calculated based on the delay variance, , , These are dynamic weighting coefficients, and the sum of the three is 1. The allocation logic is as follows: during peak hours, traffic density is high, and vehicle speed is mainly constrained by density; therefore, the weighting is increased. Weighting emphasizes the impact of density-corrected speed components; during off-peak hours, traffic flow is stable and density is moderate, with vehicle speed primarily influenced by traffic flow and basic capacity, therefore, the weighting is increased. Weighting emphasizes the role of the baseline vehicle speed component; during off-peak hours, traffic flow is low and density is low, and vehicle speed fluctuations mainly stem from the uncertainty of stop delays and signal delays, therefore, weighting is increased. Weighting emphasizes the corrective effect of the delay correction speed component.

[0029] In equation (1), the reference vehicle speed component The mathematical expression for the calculation using the flow-saturation conversion model is as follows:

[0030]

[0031] In equation (2), For free-flow vehicle speed, The traffic saturation level of a road segment is determined by the ratio of actual traffic flow to saturation traffic flow, and the calculation formula is as follows:

[0032]

[0033] In equation (3), This represents the current actual bus traffic volume on the road segment. This represents the saturation flow rate of the road section.

[0034] In equation (1), the density-corrected vehicle speed component The calculation is performed using the Greenshirez linear velocity-density model, and its mathematical expression is as follows:

[0035]

[0036] In equation (4), Free-flow vehicle speed; This represents the actual traffic density of the current grid or road segment. Congestion density refers to the critical density at which traffic flow is completely congested and vehicle speed approaches zero.

[0037] In equation (1), the delay correction speed component The mathematical expression for calculating using the delay variance reduction model is as follows:

[0038]

[0039] In equation (5), For free-flow vehicle speed, The delay reduction factor is determined by the statistical variance of bus stop delays on road sections and traffic delays at intersections. The larger the variance value, the stronger the delay fluctuation and the higher the speed reduction.

[0040] In S3, the line spatial density model is a spatiotemporal occupancy quantification model for bus lane grid cells, and the specific construction process is as follows:

[0041] First, taking a single grid cell as a statistical section, the critical density of bus operation within the grid cell is calculated based on the bus body length, safe following distance, and current average driving speed. This is the maximum number of buses that the grid cell can accommodate without congestion.

[0042] Secondly, by combining the actual road length occupied by buses and the virtual space occupied by buses in the grid unit, the actual cross-sectional space occupancy rate of buses in the grid unit is calculated, that is, the ratio of the space occupied by buses to the available space of the grid unit.

[0043] Subsequently, based on the critical density, the actual cross-sectional space occupancy rate is mapped to the linear density index of the grid cells, which reflects the density of grid cells occupied by the public flow in the time dimension.

[0044] Finally, the line density index is compared with the total available spatiotemporal resources of the grid cell to obtain the grid-level utilization rate, thereby achieving accurate quantification of the spatiotemporal resource occupancy status of the smallest unit of the bus lane.

[0045] The formula for calculating grid-level utilization is:

[0046]

[0047] In equation (6), For grid-level utilization; The spatial density of the grid lines; The critical density of the grid; The average vehicle speed across the grid; Free-flow vehicle speed; This refers to the cross-sectional space occupancy rate.

[0048] Among them, cross-sectional space occupancy rate To calculate the proportion of the total observation time spent by public transport vehicles within a given grid unit during a statistical period, the calculation method is as follows:

[0049]

[0050] In equation (7), Let i be the actual time the i-th bus occupies the grid. This represents the total duration of the statistical period.

[0051] The segment-level utilization rate is obtained by weighting the grid-level utilization rate, and the calculation formula is as follows:

[0052]

[0053] In equation (8), For road segment utilization rate; Let i be the utilization rate of the i-th grid. The length of the i-th grid; This represents the total length of the road segment.

[0054] The route-level utilization rate is obtained by weighting the segment-level utilization rate, and the calculation formula is as follows:

[0055]

[0056] In equation (9), For route-level utilization; Let j be the utilization rate of the j-th road segment; Let be the bus traffic flow for the j-th road segment.

[0057] Furthermore, in step S4, the complete process of identifying remaining spatiotemporal resources based on the three-level utilization rate is as follows:

[0058] 1. First, set the threshold for three levels of utilization rate. The utilization threshold for grid-level cells. The utilization threshold for road segment-level units. The utilization threshold is set for route-level units; the threshold is dynamically adjusted based on road grade and peak / off-peak hours.

[0059] 2. Determine resource status step by step from grid level to road segment level to route level to form spatial resource identification results;

[0060] 3. By combining signal timing and bus travel time calculations to determine available time windows, a final list of reservable time and space resources is obtained.

[0061] The method for identifying remaining time resources is as follows:

[0062] Extract the green light duration for the current phase, using the signal cycle as the unit. Red light duration Cycle duration ; Calculate the theoretical travel time for buses through the road segment. Get the available green light time ;like (Minimum reservation time) indicates that the green light phase is a reservationable time window; at the same time, the number and distribution of consecutive reservationable time windows are counted to form time-segment-level reservationable time resources, avoiding conflicts with bus traffic.

[0063] The rules for identifying remaining space resources at each level are as follows:

[0064] Grid-level determination: If The grid is deemed to have surplus space resources and is marked as an available grid; at the road segment level: if the percentage of surplus grids within a road segment is ≥50% and... The road segment was deemed to have sufficient overall space resources and was marked as a bookable segment; Route-level determination: if the percentage of bookable segments within the route is ≥60% and If a route is deemed eligible for reservations, it is marked as a reservation-available route; otherwise, if resources are scarce, reservation-reserved vehicles are prohibited from entering.

[0065] The threshold of ≥50% surplus grid ratio for road segment-level judgment is set to balance the spatial fluctuation of grid utilization within the road segment, allowing for high utilization areas in some parts of the road segment, while ensuring that reserved vehicles have continuous passage space within the road segment. The higher threshold of ≥60% reservable road segment ratio for route-level judgment is set to strengthen the bottom-line constraint of public transport priority and avoid the interruption of reservation passage and the impact on the overall operation efficiency of public transport due to the low proportion of reservable road segments within the route.

[0066] Finally, the integrated output of reservable spatiotemporal resources is as follows: reservable grids, reservable road segments, reservable routes and corresponding reservable time windows are spatiotemporally matched to generate a standardized resource list containing spatial codes, time intervals, maximum number of vehicles allowed to make reservations and traffic constraints. This list is then directly output to the reservation right-of-way allocation system to achieve precise opening and dynamic management of remaining spatiotemporal resources.

[0067] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0068] 1. Compared to traditional methods that rely solely on segment-level cross-sectional statistics or fixed-time-period estimations of utilization rates, this invention constructs a three-tiered, interconnected calculation system at the grid, segment, and route levels, significantly improving the accuracy and granularity of resource quantification. Traditional methods can only provide macroscopic estimates of segment-level resources, failing to reflect the transmission relationship between microscopic vehicle occupancy and macroscopic road network resources, easily leading to distorted resource identification. This invention uses bus-occupied space as a basis for refined spatial division, combining multi-source dynamic traffic data with weighted calculations at each level. This accurately reflects the full-scale resource utilization status of bus lanes, refining resource identification granularity from the segment level to the grid level, fundamentally solving the problems of coarse resource quantification and large deviations in identification results in existing technologies.

[0069] 2. Compared to methods that rely solely on static thresholds to determine remaining spatiotemporal resources, this invention enables dynamic identification of resources during peak and off-peak hours, significantly improving the utilization efficiency of dedicated bus lanes while ensuring bus priority. Traditional management methods completely close off bus lanes during peak hours and leave a large amount of resources idle during off-peak hours. This invention combines three-level utilization rates with dynamic threshold rules. Verified through implementation examples, it allows for reservation-based passage on non-bottleneck sections during morning and evening peak hours, increasing the average daily resource utilization rate of dedicated bus lanes by over 32% and reducing average bus delays by over 48%. Without affecting bus operation efficiency, it effectively releases redundant road resources, solving the industry problem of "peak-hour congestion and off-peak idleness."

[0070] 3. Compared to management methods that do not integrate multi-source data and multi-level spatial coupling relationships, this invention is more adaptable to complex scenarios such as intersections and bus stops, and the allocation of reserved right-of-way is safer and more reliable. This invention eliminates the interference of intersection queuing and bus stops on resource identification through trajectory cleaning, stopping time separation, spatiotemporal alignment, and three-level topological coupling analysis. Verification by examples shows that the maximum queue length at intersection entrances is reduced by more than 40%, significantly reducing the risk of spatiotemporal conflicts between reserved vehicles and buses, improving the safety and stability of reserved passage, and providing stable and reliable technical support for intelligent management and dynamic allocation of right-of-way for urban reserved bus lanes. Attached Figure Description

[0071] Figure 1 This is a schematic diagram of the method for calculating the three-level utilization rate and identifying spatiotemporal resources of a reservation-only bus lane according to the present invention.

[0072] Figure 2 This is a schematic diagram illustrating the three-level spatial unit division and implementation scenario of the reserved bus lane in this embodiment. Detailed Implementation

[0073] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. However, the embodiments of this invention are not limited thereto. It should be understood that this embodiment is a preferred embodiment of this invention and is used to explain the invention rather than limit its scope of protection. Other embodiments obtained by those skilled in the art based on this embodiment without creative effort are all within the scope of protection of this invention.

[0074] This embodiment selects a continuous bus lane system composed of main and secondary roads in the core urban area as the test platform, covering four typical time periods: morning peak 7:00-8:00, morning off-peak 10:00-11:00, afternoon off-peak 15:00-16:00, and evening peak 18:00-19:00. It fully covers the traffic operation characteristics of urban roads throughout the day, and can systematically demonstrate the technical adaptability and resource identification capability of the present invention under various traffic conditions such as high saturation congestion, medium saturation stability, and low saturation smooth flow. At the same time, it takes into account the differences in road grades between main and secondary roads, and realizes the refined management and dynamic allocation of reservation bus lane resources throughout the day.

[0075] Figure 2The invention is illustrated in the applicable implementation scenarios. The test section consists of one route-level unit, three segment-level units, and two adjacent signal-controlled intersections. The route-level unit is a complete dedicated bus route, formed by the spatial fusion of three consecutive segment-level units and two signal-controlled intersections. The segment-level unit is the dedicated bus lane between two adjacent intersections, which is the smallest operational unit for reserved right-of-way allocation. The grid-level unit uses the space P occupied by a single bus as the length benchmark and is the smallest spatial unit for utilization calculation.

[0076] The specific road conditions are set as follows:

[0077] This route-level unit comprises three segment-level units: R1, R2, and R3. R1 is an urban arterial road located between intersections J1 and J2, with an effective length of 450m, a two-way four-lane cross-section, and a 3.5m wide bus lane on one side, with straight bus stops along the route. R2 is an extension of the arterial road, located in the exit lane area of ​​intersection J2, with an effective length of 300m. R3 is an urban secondary arterial road with an effective length of 350m, featuring a continuous bus lane along its entire length.

[0078] The standard bus length is 12m, the minimum safe following distance is 5m, the free-flow speed on main roads is 40km / h, and the free-flow speed on secondary roads is 30km / h. The congestion density of bus-only lanes is 180veh / km. High-density periods are defined as traffic density greater than or equal to 70% of the congestion density, i.e., 126veh / km; low-density periods are defined as traffic density less than or equal to 30% of the congestion density, i.e., 54veh / km; and periods with traffic density in between are considered medium-density periods. Intersections J1 and J2 along the route use a unified signal timing scheme with a signal cycle length of 120s. The green light duration for the bus-only phase is 35s, the yellow light duration is 3s, and the red light duration is 82s. The bus phase prioritizes the passage of buses without delays; reserved private vehicles can only apply for passage during the red light period of the bus phase.

[0079] The traffic demand parameters for each time period throughout the day are set as follows: during the morning peak, the bus flow rate is 100veh / h (approximately 200 pcu / h, with buses converted to the equivalent of a standard passenger car at 2.0 pcu / veh), and the reserved private vehicle flow rate is 320 pcu / h; during the evening peak, the bus flow rate is 90veh / h (approximately 180 pcu / h), and the reserved private vehicle flow rate is 300 pcu / h; during the morning and afternoon off-peak periods, the bus flow rate is approximately 40veh / h (approximately 80 pcu / h), and the reserved private vehicle flow rates are 180 pcu / h and 190 pcu / h, respectively. The theoretical maximum capacity of the bus lane on the test section is 600 pcu / h. Assuming a uniform distribution of bus traffic during the morning rush hour, the bus traffic volumes for routes R1, R2, and R3 are 35veh / h, 35veh / h, and 30veh / h, respectively, with a total volume of 100veh / h. Similarly, during the evening rush hour, the bus traffic volumes for routes R1, R2, and R3 are 30veh / h, 30veh / h, and 30veh / h, respectively, with a total volume of 90veh / h. For off-peak hours, assuming the off-peak bus traffic volumes for routes R1, R2, and R3 are 16veh / h, 12veh / h, and 12veh / h, respectively, with a total volume of 40veh / h. During the morning rush hour, the measured traffic density was 92 veh / km for section R1, 75 veh / km for section R2, and 60 veh / km for section R3. During the evening rush hour, the measured traffic density was 98 veh / km for section R1, 80 veh / km for section R2, and 65 veh / km for section R3. During off-peak hours, the measured traffic density for all three sections was 35 veh / km.

[0080] S1: Multi-source dynamic traffic data collection and refined preprocessing;

[0081] Relying on roadside microwave vehicle detectors, bus GPS positioning terminals, intersection signal controllers, high-definition video checkpoints, and reservation-based passage management platforms, the system collects real-time data on bus trajectory, signal timing, traffic flow and density, vehicle speed, intersection queue length, bus stop time, and reserved vehicle operation characteristics.

[0082] For bus GPS trajectory data, the DBSCAN density clustering algorithm was used to remove outliers, with a neighborhood radius ε of 8m and a minimum number of points MinPts of 3. All trajectory points were traversed and calculated. Based on the density clustering results, isolated trajectory points and those not meeting the density threshold were identified as abnormal noise points such as GPS drift or signal loss and were removed. Temporal interpolation and spatial smoothing were performed on the remaining valid trajectory points to ensure trajectory continuity and location accuracy. Simultaneously, a spatiotemporal alignment algorithm was used to map multi-source data such as bus trajectories, signal timing, and traffic flow to the same spatiotemporal grid system, eliminating timestamp and spatial coordinate discrepancies.

[0083] The space occupied by a single bus consists of three parts: the bus body length, the minimum safe following distance, and the clearing distance. The clearing distance comprehensively considers vehicle reaction distance, braking distance, lane-changing distance, intersection queue dissipation distance, and bus stop buffer distance. In this embodiment, the bus clearing distance during morning and evening peak hours is taken as 28m, corresponding to a bus-occupied space length of 45m in the grid-level unit; the bus clearing distance during morning and afternoon off-peak hours is taken as 35m, corresponding to a grid-level unit length of 52m.

[0084] S2: Three-level spatial division and construction of digital road network system;

[0085] The dedicated lane is divided into three independently coded road segment units, R1, R2, and R3, with intersections as nodes. Each road segment is further divided into continuously coded grid units according to the grid length benchmark for different time periods. During the morning and evening peak hours, R1 is divided into 10 grids, R2 into 7 grids, and R3 into 8 grids. During the morning and afternoon off-peak hours, R1 is divided into 9 grids, R2 into 6 grids, and R3 into 7 grids.

[0086] By using GIS spatial connectivity, attribute information such as dedicated lane length, speed limit, station location, and signal control node is mapped to each grid-level unit. The grid-level data is integrated using road segment-level units as intermediate carriers. Then, multiple road segment-level units are merged with intersection space to form route-level units. PostGIS and GeoPandas tools are used to realize the three-level spatial digital representation and construct a multi-level spatially integrated bus lane topology network system.

[0087] S3: Analysis of the spatial and temporal patterns of vehicles in three-dimensional space;

[0088] Based on the completion of multi-source data spatiotemporal alignment and GPS trajectory cleaning, the first step was to separate travel time from stop and dwell time. Using the characteristics of continuous speed drops and sustained dwell times of buses, the entire process of bus entry, stopping, and starting was identified, eliminating interference from stopping operations on the calculation of road segment travel speed and flow, thus obtaining spatiotemporal samples under pure travel conditions. Subsequently, kernel density estimation and spatiotemporal clustering algorithms were used to fit the spatiotemporal distribution of flow, density, and occupancy in grid-level units, identifying the spatiotemporal distribution patterns of high-saturation and unobstructed traffic flow areas in different time periods and road segments, and locating congestion-prone grids and bottleneck sections.

[0089] Based on this, the real-time speed sequence of buses during the green light phase at signalized intersections is extracted. High-frequency time series decomposition is used to separate acceleration, constant speed, and deceleration features, obtaining typical driving trajectories and speed change patterns of buses passing through intersections and bus stops. Simultaneously, a complex network model is used to construct a three-level topological coupling relationship of "grid-segment-route," quantifying the transmission and impact of micro-grid resource status on meso-level segment utilization and macro-level route resource supply.

[0090] Based on the above analysis of spatiotemporal patterns, it can be seen that during the morning and evening peak hours, the saturation of bus stops and intersection entrances on the R1 route is relatively high. The bus traffic flow on the R2 and R3 routes is moderate and the operation is stable. The average speeds on each route during the morning peak are 25 km / h, 27 km / h, and 22 km / h, respectively; the average speeds on each route during the evening peak are 24 km / h, 27 km / h, and 22 km / h, respectively. During the off-peak hours in the morning and afternoon, the traffic flow on all routes is smooth and stable. The average speeds on R1 and R2 are 35 km / h, and the average speed on R3 is 27 km / h, with no obvious congestion or queuing.

[0091] S4: Calculation of average vehicle speed in three-level space and step-by-step solution of utilization rate in three-level space;

[0092] Based on the results of three-level spatial division and spatiotemporal pattern analysis, the technology of separating stop time and three-level segmentation modeling is introduced to eliminate interference from intersection queuing and bus stops. Combined with the dynamic weight allocation of traffic density and delay variance, the average vehicle speed at the grid level, road segment level and route level is accurately calculated.

[0093] During peak hours, density has the primary influence, with weighting. Take 0.3, Take 0.6, The value is set to 0.1, with traffic flow being the primary factor during off-peak hours, and the weighting is adjusted accordingly. Take 0.6, Take 0.3, Taking 0.1, during off-peak hours, delay fluctuations are the main impact, with weighting... Take 0.2, Take 0.2, Take 0.6. Where the base vehicle speed is... The road segment saturation is first calculated using saturated and actual flow rates, and then converted using free-flow vehicle speeds. Taking the morning rush hour on R1 as an example, the road segment's saturated flow rate is 600 pcu / h, and the actual bus flow rate is 35veh / h (approximately 70 pcu / h). Based on the 40 km / h free-flow vehicle speed of the main road in this segment, the conversion is performed. Value taken as 35 km / h; density-corrected speed Based on the Greenhills linear velocity-density relationship, and substituting the free-flow vehicle speed, measured traffic density, and road congestion density, the following conversions were performed: The free-flow vehicle speed on the R1 arterial road is 40 km / h, the measured traffic density during the morning peak is 92 veh / km, and the congestion density is 180 veh / km. The corresponding conversions are as follows: Value taken as 20km / h; speed correction for delay Corrected for delay variance at stops and intersections, during morning and evening peak hours, the overall speed reduction factor corresponding to the delay variance of R1 and R2 segments is 0.65, and that of R3 segment is 0.70; during off-peak hours, the reduction factor for R1 and R2 segments is 0.75, and that of R3 segment is 0.80. This is combined with free-flow speed conversion. The speed was initially set to 26 km / h. Then, the three component speeds were weighted according to the corresponding peak-hour weighting, resulting in an average speed of 25 km / h for R1 (uninterrupted). Following the same step-by-step calculation logic, the morning peak-hour speeds (R2 and R3) were calculated as 27 km / h and 22 km / h, respectively. During the evening peak, each road segment was calculated step-by-step based on measured data of traffic flow, density, and delays for the corresponding time periods. The final average speeds for R1, R2, and R3 were 24 km / h, 27 km / h, and 22 km / h, respectively. During the morning and afternoon off-peak hours, traffic flow was stable, and a uniform off-peak weighting method was used for calculation. After calculating the speeds for all road segments, the average speeds for R1 and R2 were 35 km / h, and for R3, it was 27 km / h. After completing the speed calculations, based on the line space density model, and combining the actual line density, critical density, average speed, free-flow speed, and cross-sectional space occupancy rate of the grid, the three-level utilization rate was calculated step-by-step.

[0094] After the average vehicle speed is calculated, the utilization rate of the bus lane is accurately calculated at three levels, from grid level to road segment level and then to route level, following a progressive logic. Table 1 shows the basic calculation parameters for each grid of road segments R1, R2, and R3 during the morning peak hours; Table 2 shows the basic calculation parameters for each grid of road segments R1, R2, and R3 during the evening peak hours; and Table 3 shows the basic calculation parameters for each grid of road segments R1, R2, and R3 during the morning (afternoon) off-peak hours. Taking the third grid of the R1 road section during the morning peak as an example, the grid line space density is 92 veh / km, the critical density of the bus lane is 120 veh / km, the average vehicle speed of the grid is 27 km / h, the free-flow vehicle speed is 40 km / h, and the cross-sectional space occupancy rate is 0.82. Substituting these values ​​into the calculation, the utilization rate of this grid is approximately 0.42. The calculation parameters for the grid utilization rate during the evening peak are similar to those during the morning peak, and the results are basically consistent. During the off-peak hours in the morning and afternoon, the traffic density is low and the space occupancy rate is small. The utilization rate of the grid at the same location is only about 0.07, indicating a significant characteristic of idle resources.

[0095] Table 1. Basic Calculation Parameters for Each Grid of Road Sections R1, R2, and R3 During Morning Peak Hours

[0096]

[0097] Table 2 Basic Calculation Parameters for Each Grid of Road Sections R1, R2, and R3 During Evening Peak Hours

[0098]

[0099] Table 3 Basic Calculation Parameters for Each Grid of Road Sections R1, R2, and R3 during Off-Peak Hours in the Morning and Afternoon

[0100]

[0101] After weighted calculation, the utilization rates of road sections R1, R2, and R3 during the morning peak were 0.42, 0.31, and 0.27, respectively, while those during the evening peak were 0.41, 0.31, and 0.28, respectively. During the morning off-peak, the utilization rates of the three road sections were 0.07, 0.06, and 0.06, respectively, and during the afternoon off-peak, they were also 0.07, 0.06, and 0.06, respectively. This indicates that the utilization rate during off-peak hours is relatively low, and the resource idleness is significant.

[0102] To calculate route-level utilization, this embodiment uses the proportion of bus traffic to total route traffic in each segment as the weight for weighted averaging. Assuming a uniform distribution of bus traffic during the morning peak hours, with bus traffic volumes of 35veh / h, 35veh / h, and 30veh / h for routes R1, R2, and R3 respectively, and a total traffic volume of 100veh / h, then the route utilization rate during the morning peak hours is 0.34. Similarly, during the evening peak hours, with bus traffic volumes of 30veh / h, 30veh / h, and 30veh / h for routes R1, R2, and R3 respectively, and a total traffic volume of 90veh / h, then the route utilization rate during the evening peak hours is 0.33. For off-peak hours, assuming the morning off-peak bus traffic volumes of 16veh / h, 12veh / h, and 12veh / h for routes R1, R2, and R3 respectively, and a total traffic volume of 40veh / h, then the route utilization rate during the morning off-peak hours is 0.06. The bus traffic distribution during the afternoon off-peak hours is the same as that during the morning off-peak hours, so the route utilization rate is also 0.06.

[0103] In summary, the route utilization rates for the morning peak, evening peak, morning off-peak, and afternoon off-peak were 0.34, 0.33, 0.06, and 0.06, respectively. The overall utilization rate exhibits a typical "double peak and double valley" distribution pattern, which is highly consistent with the urban road traffic operation pattern.

[0104] S5: Identification and inventory output of remaining spatiotemporal resources based on three-level utilization;

[0105] Based on the three-level utilization rate calculation results for four time periods throughout the day, the remaining spatiotemporal resources are identified step by step according to the graded thresholds. The thresholds are dynamically adjusted according to road grade and peak and off-peak traffic conditions. Here, a grid-level threshold is set for peak hours. Road segment level threshold Route-level threshold Off-peak hours grid-level threshold Road segment level threshold Route-level threshold .

[0106] Spatial resource identification follows a hierarchical judgment rule from grid level to road segment level and then to route level. A grid with a utilization rate below a corresponding threshold is considered a resource-rich grid. If the proportion of surplus grids within a road segment is not less than 50% and the road segment utilization rate is below a corresponding threshold, it is considered a reservable road segment. If the proportion of reservable road segments within a route is not less than 60% and the route utilization rate is below a corresponding threshold, it is considered a reservable route. Temporal resource identification uses signal cycles as units to calculate the theoretical travel time for buses through road segments, obtaining the available green light time. When the available green light time is greater than the minimum reservation travel time of 6 seconds, that phase is considered the effective reservable time window.

[0107] Based on the assessment, during the morning peak hours, the utilization rate of section R1 was slightly higher than the section threshold, with only some grids available for reservation. The utilization rates of sections R2 and R3 were 0.31 and 0.27 respectively, both lower than the section threshold of 0.40, thus meeting the criteria for reservationable sections. During the evening peak hours, the utilization rate of section R1 was also slightly higher than the section threshold, while the utilization rates of sections R2 and R3 were 0.31 and 0.28 respectively, both lower than the section threshold of 0.40, also meeting the criteria for reservationable sections. The proportion of reservationable sections within the routes during both morning and evening peak hours exceeded 60%, the route-level utilization rate was lower than the route threshold, and the available green light time was greater than 6 seconds, indicating an effective reservation time window. Therefore, sections R2 and R3 can be opened for reserved vehicles during morning and evening peak hours. During the morning and afternoon off-peak hours, the proportion of surplus grids across the entire route exceeded 80%, and all three sections met the criteria for reservationable sections, allowing full reservation access.

[0108] Finally, by integrating spatial coding, time intervals, maximum number of vehicles allowed to be reserved, and traffic constraints, a standardized list of reservable spatiotemporal resources for the whole day is generated. During the morning and evening peak hours, the reservable road segments are R2 and R3, while during off-peak hours, the reservable road segments are R1, R2, and R3. The reservable time window is the entire red light phase for buses. The maximum number of vehicles that can be reserved is 6 pcu for R1, 4 pcu for R2, and 5 pcu for R3. The constraint is that the dedicated lane is automatically cleared 3 seconds before the bus arrives to ensure that buses can pass without interference.

[0109] Verification has shown that, by adopting the method of this invention, the average daily resource utilization rate of bus lanes has increased by more than 32%, and the average bus delay has decreased by more than 48%. Compared with the traditional control method that completely prohibits reservation-based passage during peak hours, this invention, while rigidly guaranteeing the priority passage of buses, achieves accurate identification and reservation of the remaining time and space resources of bus lanes during peak hours, significantly increasing the peak-hour passage time for reserved vehicles, and reducing the maximum queue length at intersection entrances by more than 40%. This fully demonstrates that the method of this invention is scientific, feasible, practical, and efficient, and can provide reliable technical support for the intelligent management and right-of-way optimization of reserved bus lanes in cities.

Claims

1. A method for calculating the three-level utilization rate and identifying spatiotemporal resources of a reserved bus lane, characterized in that, Includes the following steps: S1, based on the virtual occupancy space of buses, the bus lane is discretized into three levels of space: grid level, road segment level and route level, to construct a multi-level digital road network and complete topology coding and attribute mapping; S2 integrates bus GPS trajectory, signal timing, traffic flow and bus departure frequency data to analyze the spatial and temporal patterns of vehicle operation in three-level spaces. S3 calculates grid-level utilization based on virtual occupied space, critical density, and cross-sectional space occupancy rate; The segment-level utilization rate is obtained by weighting the grid length. The route-level utilization rate is obtained by weighting the bus traffic flow of each section. S4 identifies remaining spatiotemporal resources based on the three-level utilization rate and outputs the spatiotemporal resource results of dedicated lanes that can be used for reservation right-of-way allocation.

2. The method for calculating the three-level utilization rate and identifying spatiotemporal resources of a reserved bus lane according to claim 1, characterized in that, The virtual occupied space in S1 is composed of the bus body length, the minimum safe following distance, and the bus clearing distance. The length of the virtual occupied space is dynamically adjusted with traffic density. During high-density periods, the virtual occupied space is shortened to improve the passage efficiency of reserved vehicles, while during low-density periods, the virtual occupied space is increased to ensure priority passage for buses. The high-density period is the period when the traffic density is greater than or equal to 70% of the congestion density, the low-density period is the period when the traffic density is less than or equal to 30% of the congestion density, and the period when the traffic density is between the two is the medium-density period.

3. The method for calculating the three-level utilization rate and identifying spatiotemporal resources of a reserved bus lane according to claim 1, characterized in that, In S1, the grid-level unit is the virtual space occupied by a single bus, the road segment-level unit is the bus lane between two adjacent intersections and is the smallest reservation unit, and the route-level unit is composed of multiple road segment-level units and signal-controlled intersections.

4. The method for calculating the three-level utilization rate and identifying spatiotemporal resources of a reserved bus lane according to claim 1, characterized in that, The construction of a multi-level digital road network system in S1 includes mapping bus lane attributes to grid cells through GIS spatial connections, splitting and generating independently coded road segment-level units with intersections as nodes, integrating road segment-level units with intersection spaces to generate route-level units, and using PostGIS and GeoPandas to achieve three-level spatial digital representation.

5. The method for calculating the three-level utilization rate and identifying spatiotemporal resources of a reserved bus lane according to claim 1, characterized in that, The analysis of vehicle runtime spatial-temporal patterns in S2 includes matching multi-source data through a spatio-temporal alignment algorithm, removing abnormal trajectory points and identifying bus stop dwell times using the DBSCAN clustering algorithm, identifying traffic flow spatio-temporal clustering characteristics using kernel density estimation and spatio-temporal clustering algorithms, and analyzing the three-level spatial topology coupling relationship through a weighted directed multi-layer network model with a complex network model.

6. The method for calculating the three-level utilization rate and identifying spatiotemporal resources of a reserved bus lane according to claim 1, characterized in that, The step-by-step calculation of the three-level utilization rate in S3 includes first using the grid level as the smallest unit, and then calculating the critical density and cross-sectional space occupancy rate by combining the bus lane length, virtual occupied space, and average driving speed, to construct a line space density model to determine the grid-level utilization rate; the specific construction process of the line space density model is as follows: First, taking a single grid cell as a statistical section, the critical density of bus operation within the grid cell is calculated based on the bus body length, safe following distance, and current average driving speed. This is the maximum number of buses that the grid cell can accommodate without congestion. Secondly, by combining the actual road length occupied by buses and the virtual space occupied by buses in the grid unit, the actual cross-sectional space occupancy rate of buses in the grid unit is calculated, that is, the ratio of the space occupied by buses to the available space of the grid unit. Subsequently, based on the critical density, the actual cross-sectional space occupancy rate was mapped to the linear density index of the grid cells. Finally, the line density index is compared with the total available spatiotemporal resources of the grid cells to obtain the grid-level utilization rate; The segment-level utilization rate is then obtained by weighted summation of the grid-level utilization rates; finally, the route-level utilization rate is calculated by comprehensively calculating the segment-level utilization rates; the formula is: In the formula, For grid-level utilization, The spatial density of the grid lines; The critical density of the grid; The average vehicle speed across the grid; Free-flow vehicle speed; This refers to the cross-sectional space occupancy rate. For road segment utilization, Let i be the utilization rate of the i-th grid. The length of the i-th grid; This refers to the total length of the road segment; For route-level utilization, Let j be the utilization rate of the j-th road segment; Let be the bus traffic flow for the j-th road segment.

7. The method for calculating the three-level utilization rate and identifying spatiotemporal resources of a reserved bus lane according to claim 6, characterized in that, When calculating the average vehicle speed across three levels of space, a stop time separation and three-level segmented modeling technique is introduced to eliminate intersection queuing and stop interference. A vehicle speed calculation model is constructed by combining traffic density and delay variance with dynamic weight allocation. The mathematical expression of the vehicle speed calculation model is: In the formula, To eliminate the average speed of vehicles in the three-level space after stopping and queuing interference; The baseline vehicle speed component is calculated based on the traffic flow of the road segment. For density-corrected vehicle speed components calculated based on traffic density; For the delay-corrected speed component calculated based on the delay variance, , , These are dynamic weighting coefficients, and the sum of the three is 1.

8. The method for calculating the three-level utilization rate and identifying spatiotemporal resources of a reserved bus lane according to claim 1, characterized in that, The remaining spatiotemporal resource identification method based on three-level utilization in S4 includes: first, setting utilization level thresholds at the grid level, road segment level, and route level, with the thresholds dynamically adjusted according to road grade and peak / off-peak hours; determining resource status step by step from grid level to road segment level to route level to form spatial resource identification results; combining signal timing and bus travel time to calculate available time windows, and finally integrating them to obtain a list of reservable spatiotemporal resources.

9. The method for calculating the three-level utilization rate and identifying spatiotemporal resources of a reserved bus lane according to claim 8, characterized in that, The remaining time resource identification method is as follows: extract the current phase green light duration, red light duration, and cycle duration in units of signal cycle; Calculate the theoretical travel time for buses to pass through the road segment to obtain the available green light time; If the available green light time is greater than the minimum reservation time, then the green light phase is a reservationable time window; at the same time, the number and distribution of consecutive reservationable time windows are counted to form time-segment-level reservationable time resources to avoid conflicts with bus traffic.

10. The method for calculating the three-level utilization rate and identifying spatiotemporal resources of a reserved bus lane according to claim 8, characterized in that, The rule for identifying remaining space resources at each level is as follows: if the grid-level utilization rate is less than the corresponding grid-level classification threshold, the grid space resources are determined to be abundant and marked as an occupiable grid. If the proportion of surplus grids in a road segment is ≥50% and the road segment utilization rate is less than the corresponding road segment classification threshold, the overall space resources of the road segment are deemed sufficient and it is marked as a reservable road segment. If the percentage of bookable sections within a route is ≥60% and the route-level utilization rate is less than the corresponding route-level classification threshold, then the route is deemed to meet the conditions for opening reservations and is marked as a bookable route. Conversely, if resources are deemed scarce, reserved vehicles will be prohibited from entering.

Images