756 results about "Transit bus" patented technology

The present invention discloses a bus route optimization design method and system based on resident travel big data, the method comprising the steps of: obtaining mobile phone signaling data of residents in a target area; obtaining the stop points of residents' travel trajectories; dividing the target area into a number of traffic areas; obtaining an OD matrix based on the stop points of residents' travel trajectories and traffic areas; constructing a bus route planning model; generating a number of bus alternative routes based on a shortest path algorithm; obtaining a target bus route from the alternative routes based on a genetic algorithm and in combination with an objective function and an OD matrix. Compared with the prior art, the present invention not only innovatively combines spatiotemporal big data and genetic algorithms in technical means, improving the accuracy and efficiency of bus route planning, but also demonstrates strong advantages in system architecture and data processing capabilities, and can provide a more scientific, dynamic and sophisticated optimization solution for the urban public transportation system, contributing to the efficient operation of urban transportation and the construction of smart cities.

The invention relates to the technical field of information and communication suitable for logistics management, and aims to provide a passenger and freight co-transport bus route planning method based on an adaptive large neighborhood search algorithm. Comprising the steps of receiving passenger and freight transportation information through a reservation mode, and determining a passenger and freight co-transportation strategy based on different demand characteristics; a path optimization model is established, and on the premise of meeting all requirements, factors of vehicle capacity and time window limitation are emphasized, and a path scheme with the lowest cost is sought; solving the path optimization model based on a self-adaptive large neighborhood search algorithm, breaking a current solution by using a damage operation from an initial solution, and exploring a new solution space by using a repair operation; and finding a path scheme meeting a cost minimization target through continuous iteration. According to the invention, supply and demand balance in a passenger and freight co-transport mode can be realized, preferential meeting of passenger demands is taken as a core, a freight transport service is taken into consideration, and a new view angle and a solution are provided for efficient operation of a reservation type public transportation system.

The invention discloses an urban cross-domain multi-mode network space modeling method, and belongs to the technical field of traffic simulation. In order to solve the problem that urban traffic cross-domain spatial modeling is complex, the method comprises the steps that road, hub and track BIM models are classified, traffic semantics in different BIM objects in the BIM models are recognized, and traffic elements are extracted; geometric element abstraction is carried out, spatial topology connection relations and attribute information of geometric elements are defined, and single-mode traffic network model modeling is carried out; coordinate conversion is carried out, and the relative position coordinates of the single-mode traffic network model are converted into 1984 world geodetic surveying geographic coordinates; a walking traffic network is used as an intermediate network, a walking traffic mode and transfer rules of roads, rails, buses and hubs are used as strategies, network fusion is achieved through a map matching algorithm, and a multi-mode network space is modeled; and carrying out topological structure verification on the obtained multi-mode network space. According to the invention, full-link calculation from design to simulation is realized.

The invention discloses a bus and station interactive scheduling method and system based on V2X and deep reinforcement learning decision, a medium and equipment, and the method comprises the steps: collecting bus operation dynamic data, station passenger flow and environment data in real time, and combining V2X network interaction information to construct a full-dimension perception system; a deep reinforcement learning model is adopted for dynamic decision making, intelligent closed-loop control of vehicle scheduling, station service and traffic signal cooperation is achieved, the bus punctuality rate is remarkably increased, the waiting time of passengers is shortened, operation safety is guaranteed through active early warning of abnormal conditions, and intelligent upgrading of a bus system from passive response to active prevention is achieved.

The invention relates to a method for measuring the toughness of an urban public transport network in extreme rainstorm weather. The method comprises the following steps: acquiring station and line data of the urban public transport network, rainstorm disaster information and passenger flow data; a bus-subway double-layer traffic network is constructed; based on the constructed bus-subway double-layer traffic network, establishing a linear relation between rainstorm rainfall intensity and passenger flow change, and calculating disturbance values of the rainstorm to stations and connecting edges in the network; constructing a bus-subway double-layer traffic network cascade failure model considering passenger flow, and representing a dynamic evolution process of an urban public traffic network in extreme rainstorm weather; and determining a network performance measurement index by using the constructed bus-subway double-layer traffic network cascade failure model, and measuring the toughness level of the urban public traffic network in extreme rainstorm weather. According to the method, the urban public transport network toughness can be accurately measured, and the coping capacity of an urban public transport system in extreme rainstorm weather is improved.

The invention discloses an STM32F407-based intelligent walking stick system and a control method, which are used for solving the problems of single function, high false alarm rate and insufficient data security of the traditional walking stick. The system integrates a posture detection module, a heart rate and blood oxygen monitoring module and an environment sensing module, constructs a user health risk pre-judgment model through multi-modal data fusion, realizes a three-level adaptive alarm mechanism in combination with a dynamic decision algorithm, and triggers acousto-optic warning, cloud communication and emergency contact linkage according to risk levels. The payment and data transmission reliability is guaranteed by adopting a physically isolated security architecture and a national secret encryption technology; the dynamic power management system optimizes energy consumption distribution according to scene requirements, and the endurance of equipment is remarkably prolonged. According to the system, the monitoring accuracy is improved through a closed-loop protection system (risk identification-adaptive response-energy optimization), bus payment and smart home control are supported, the system has the characteristics of low false alarm rate, high environmental adaptability and low power consumption, and a safe and convenient life auxiliary solution is provided for old users.

Disclosed herein are methods, devices, and systems for detecting bus lane moving violations. One aspect of the disclosure concerns a method comprising capturing a video showing a vehicle located in a bus lane, inputting video frames from the video to an object detection deep learning model to detect the vehicle and bound the vehicle in a vehicle bounding polygon, determining a trajectory of the vehicle in an image space of the video frames, transforming the trajectory of the vehicle in the image space into a trajectory of the vehicle in a GPS space, inputting the trajectory of the vehicle in the GPS space to a vehicle movement classifier to yield a movement class prediction and a class confidence score, and evaluating the class confidence score against a predetermined threshold based on the movement class prediction to determine whether the vehicle was moving when located in the bus lane.

The invention relates to a bus passenger flow prediction method based on multi-scale space-time deep learning, which belongs to the field of traffic information big data, is realized by a bus data collection and processing system, and comprises the following steps: S1, generating a bus arrival and departure time table and a passenger getting-on and getting-off clock-in data table; s2, generating a spatial-temporal feature vector matrix; s3, introducing external factor information corresponding to the site, and coding by adopting one-hot coding to form time-space sequence data; s4, a bus passenger flow prediction model based on multi-scale space-time deep learning is constructed and trained; and S5, performing prediction according to a real-time request of a user or a bus dispatching center, and outputting a passenger flow prediction value of each station in a future time period. According to the method, complex passenger flow changes can be accurately captured, the prediction accuracy is improved, the adaptability and generalization ability of the model are enhanced, scientific data support is provided for bus dispatching optimization, route planning and transport capacity configuration, and the operation efficiency and service level of a bus system are improved.

The invention discloses an automatic driving bus collaborative formation dynamic scheduling method, which comprises the following steps of: after introducing a vehicle formation operation, constructing a vehicle energy consumption model based on vehicle specific power, and measuring the vehicle operation energy consumption of an automatic driving bus formation in any formation mode; in an intelligent network connection environment, based on a rolling time domain optimization framework, a Markov decision process model is constructed, and an automatic driving bus collaborative formation dynamic scheduling problem is explained; defining a state variable and a decision variable of the Markov decision process model, and determining constraint conditions of related variables and a target function of the Markov decision process model; the decision space of the Markov decision process model is reduced, and an approximate dynamic programming algorithm is used to solve the automatic driving bus collaborative formation scheduling problem; a multi-step look-ahead strategy based on dynamic planning is provided, the convergence efficiency of an approximate dynamic planning algorithm is improved, and an automatic driving bus collaborative formation optimization scheduling scheme is obtained. And the automatic driving bus capacity utilization rate is improved.

The invention discloses a dynamic estimation method and device for the carbon emission reduction effect of a bus lane and a medium, and the method comprises the steps: estimating the mileage of a bus line running on the bus lane and the mileage of a bus line running on a common lane; the types of the bus routes are divided, so that the bus routes are divided into a first bus route mainly running on the bus lane and a second bus route mainly running on the common lane; and according to the types of the bus lines, taking bus station level, bus link level and bus line level influence factors as covariables, taking each bus line as a sample, and adopting a dynamic emission reduction effect estimation model to estimate a condition emission reduction index and an average emission reduction index of the bus lane so as to quantify global and local emission reduction effects of carbon emission factors of unit passengers. In combination with cross-city bus data, a regional hierarchy random effect is introduced, and a multi-level dynamic emission reduction effect estimation model is constructed to reveal the difference of the emission reduction effects of bus lanes between different cities.

The embodiment of the invention provides a training method of a signal lamp control model, a signal lamp control method, a device, equipment, a medium and a product, and relates to the technical field of urban road traffic. The method comprises the following steps: acquiring a plurality of first traffic data, constructing a state characterization matrix and a reward function according to the plurality of first traffic data, and performing model training according to the plurality of first traffic data, the state characterization matrix and the reward function to obtain a signal lamp control model. According to the method, in a model training process, a state representation matrix is utilized to dynamically quantify passing priorities of buses in all directions, and meanwhile, a reward function is utilized to guide the model training process, so that a signal lamp control model obtained through training can recognize and preferentially respond to bus passing demands, and signal lamp actions are dynamically adjusted; therefore, the collaborative optimization of multi-direction bus priority requests in a complex scene is realized, and the coordination capability and response efficiency of a bus priority system in a complex urban traffic environment are improved.

The invention provides a demand response bus intelligent scheduling method and system driven by real-time data. A demand response bus intelligent scheduling method driven by real-time data comprises the following steps: S1, collecting and fusing multi-source data in real time, and constructing a dynamic data set; s2, carrying out short-time demand prediction and clustering analysis based on LSTM and DBSCAN; s3, dynamic service area division and greedy algorithm line generation; s4, elastic vehicle scheduling based on combination of mixed integer programming and a simulated annealing algorithm; s5, dynamically adjusting the real-time ticket price driven by the supply and demand game model; s6, passenger personalized service matching is achieved through a collaborative filtering algorithm; s7, performing multi-target simulation verification on the digital twin platform; and S8, reinforcing a learning-driven dynamic feedback optimization mechanism. According to the real-time data-driven demand response bus intelligent scheduling method and system provided by the invention, the operation efficiency, the passenger satisfaction degree and the enterprise benefit of the demand response type customized bus can be remarkably improved, and the method and the system are suitable for construction of an urban intelligent bus system and have remarkable advantages.

The invention belongs to the field of road vehicle control methods, and relates to a bus departure control method based on a master-slave game in an intelligent network connection environment, which comprises the following steps of: (1) setting a control area CZ by taking a harbor type stop station as a center; (2) after getting on and off the bus, triggering a control decision, and sharing information such as real-time position, speed and acceleration of each bus in the bus B and the social vehicle group FV; (3) obtaining a bus driving track when the income of the bus B and the total income of the social vehicle group FV are maximum through an upper layer planning model; obtaining a social vehicle driving track under the condition that the social vehicle group FV selects the maximum individual income based on the decision of the public transport vehicle B through a lower-layer planning model; (4) solving the bilevel planning model through NSGA-II to obtain an optimal bus and social vehicle driving track; and (5) executing the vehicle track. According to the method, the negative influence caused by bus departure can be reduced to the minimum.

The invention relates to an intelligent bus combined scheduling method and device and a storage medium, and relates to the field of scheduling optimization. Bus card swiping records are called to establish a card swiping station feature set, navigation station triggering is established to carry out navigation record matching analysis, and a space-time popularity prediction map is established; constructing a multi-source traffic flow data set to perform road congestion state prediction and construct a region-level road congestion state diagram; establishing a station passenger flow demand, route travel time and a congestion state according to the space-time popularity prediction map and the region-level road congestion state map; a multi-target scheduling channel is configured to take station passenger flow demands, route travel time and congestion states as input data, and route bus scheduling optimization is carried out through a passenger waiting time target, a scheduling cost target, a vehicle utilization rate target and a congestion influence target. The technical problems of insufficient data utilization, limited prediction precision and lack of flexibility of a scheduling strategy in bus scheduling are solved.

An electrical busway system including one or more busbar assemblies and one or more connector assemblies. Each busbar assembly includes a plurality of busbars. Each connector assembly includes a plurality of connector subassemblies fitted together within a clamp assembly. Each connector subassembly includes a conducting portion, an insulating portion surrounding an exterior of the conducting portion, and a channel proximate an interior of the conducting portion, the channel being configured to receive one of the plurality of busbars.

The invention relates to the technical field of urban traffic management and carbon emission reduction accounting, and provides a bus travel carbon emission reduction accounting method fused with Beidou spatio-temporal data, which comprises the following steps: acquiring spatio-temporal trajectory data based on a Beidou satellite navigation system, and constructing a passenger-trajectory-station three-dimensional correlation model for passenger flow analysis; constructing a passenger transfer analysis system based on the passenger flow analysis result; constructing a dynamic'perception-prediction-optimization 'dynamic scheduling optimization system for passenger flow detail analysis on the basis of space-time trajectory data and a reinforcement learning algorithm; and constructing a mileage-frequency distribution model and a station efficiency evaluation model based on the spatio-temporal trajectory data, respectively generating a trajectory matching optimization strategy and a differentiated station optimization scheme, and realizing adaptive optimization of the public transportation system under dynamic demand change. Compared with a traditional static data dependence method, the method achieves the dynamic and fine-grained monitoring of the operation process of the public transportation system, and effectively improves the timeliness and integrity of data.

The invention discloses a bus carbon efficiency key causal chain identification method, electronic equipment and a medium, and the method comprises the steps: constructing a line-level bus carbon efficiency variable set which comprises a control variable, a covariant and unit passenger mileage carbon emission; the control variable is a bus operation strategy; the covariable is a bus carbon efficiency influence factor; assuming that the line-level bus carbon efficiency variable set obeys a directed acyclic graph structure formed by linear equations, and solving a causal effect matrix; establishing a bus carbon efficiency causal directed graph by taking a line-level bus carbon efficiency variable as a node and a causal effect as an edge weight; and performing post-nonlinear causal relationship test and conditional independence causal test on the bus carbon efficiency causal digraph to obtain an optimized bus carbon efficiency causal digraph, thereby obtaining a bus carbon efficiency key causal link.

The present invention discloses a bus intelligent dispatching system, which relates to the technical field of bus dispatching, and includes: an information collection module, a dispatching management module, a dispatching evaluation module and an early warning dispatching module, wherein; the information collection module is used to collect environmental characteristic information and vehicle characteristic information in real time, and send the environmental characteristic information and vehicle characteristic information to the dispatching evaluation module, wherein the environmental characteristic information includes: road condition information and planned route information; the vehicle characteristic information includes: vehicle condition information, current vehicle position information and passenger flow. The present invention makes accurate predictions for the actual conditions of route nodes, thereby formulating reasonable dispatching plans, optimizing dispatching and departure times, making flexible and efficient dispatching decisions, meeting the needs of various dispatching strategies, and timely responding to emergency dispatching, and being able to adjust routes and trips in a timely manner, thereby improving the utilization rate of bus resources.

The invention discloses an urban passenger transport resource allocation method and system based on Internet big data, and belongs to the field of intelligent transportation. A bus station service matching model is constructed, service areas are divided through bus station space distribution, service radius and travel data, travel demand intensity and station matching degree are evaluated, and the service efficiency is improved. The method comprises the following steps: establishing an accessibility evaluation model, evaluating the accessibility level of each station through destination information around the station, establishing a bus and subway collaborative evaluation model, evaluating the substitution effect of the subway on the bus and the collaborative degree of the bus and the subway through transfer data, and establishing a station residence time optimization model. The matching degree, the accessibility, the substitution degree and the cooperation degree are imported into a station residence time optimization model to adjust the residence time of the bus at each station, a bus shift dynamic adjustment model is constructed, weather and event external factors are imported into the bus shift dynamic adjustment model to predict passenger flow changes and optimize shift setting, and the bus shift dynamic adjustment model is constructed. And the bus dispatching efficiency and the operation stability are improved.

The invention belongs to the field of bus dispatching optimization, and particularly relates to a modular bus dispatching optimization method for a bus corridor, in particular to a modular bus dispatching optimization method for the bus corridor. The objective of the invention is to solve the problems of low operation efficiency, long passenger transfer waiting time, long passenger in-bus time and high operation cost of a bus company of an existing bus system. The invention discloses a modularized bus scheduling optimization method for a bus corridor. The method comprises the following steps: 1, collecting basic information of bus stations, configuring modularized vehicles, configuring modularized bus shifts, and collecting passenger travel request data; 2, defining optimization variables; 3, judging a transfer strategy, and calculating the number of transferred passengers based on the transfer strategy; 4, splitting and combining the vehicle to complete passenger transfer, and updating shift information based on a passenger transfer result; 5, establishing a modularized bus scheduling optimization model; and 6, solving the modularized bus scheduling optimization model, and outputting an optimal scheduling scheme.

The invention provides a bus stop layout and stop control strategy integration optimization method and system based on an agent model, and belongs to the technical field of urban traffic operation services. A bus stop layout model and a bus stop control model are established respectively; determining an objective function of a bus stop layout optimal stop control integration stochastic optimization problem, and constructing and iteratively updating an agent model to estimate and predict an objective function response value corresponding to an unsimulated decision variable; and fitting a simulated solution and a target function of the solution by using the agent model of Gaussian process regression to give a posteriori prediction of a target function response value of a solution which is not simulated yet, and guiding bus stop layout and a stop control strategy. According to the invention, passenger travel costs of different levels are integrated, and the ability of a bus stop layout scheme design result of a planning level to maintain bus service regularity in an operation level is enhanced.

The invention discloses a method and equipment for constructing a bus carbon emission factor library of unit passengers, and a medium. The method comprises the steps: obtaining a bus network, and estimating bus attribute parameters including the bus passenger capacity, the bus speed and the bus size; calculating an energy consumption factor, a fuel consumption factor, a fuel combustion carbon emission factor, a cold start carbon emission factor and / or a lubricant carbon emission factor of the bus so as to estimate direct carbon emission of the bus; calculating the total carbon emission of the corresponding oil well to oil tank stage of the bus, the carbon emission factor of the corresponding oil well to oil tank stage of the electric bus, the carbon emission factor of the whole life cycle of the bus and / or the infrastructure carbon emission factor of the bus so as to estimate the indirect carbon emission of the bus; and deriving individual-level, grid-level, region-level, link-level and line-level bus carbon emission factors according to the estimated direct and indirect carbon emission of the bus.

The invention discloses an intelligent bus shelter-based operation management and control method and system, and the method comprises the steps: generating operation prompt information which comprises the predicted arrival time of a bus, unoccupied seats and passenger congestion degree information; when the first bus enters the queuing time, the passenger flow density of passengers in the first queuing area is calculated according to the queuing condition, and whether the passenger flow density is smaller than a preset density value is judged; if yes, whether the first queuing area needs to be scheduled or not is judged according to the passenger crowding degree of the first bus, the passenger flow density of the next stop and the getting-on and getting-off behavior mode, and the first target passenger is drained to the first queuing area within the schedulable range of the first queuing area when scheduling is needed; and if not, traversing the operation prompt information to find a second bus associated with the first bus, and scheduling the second target passenger to a second queuing area of the second bus. Through intelligent display and passenger flow scheduling, the bus taking order of the bus shelter is optimized, and the utilization efficiency of vehicle resources is improved.

The invention discloses a trunk bus dynamic speed optimization method based on time headway equalization, and a dynamic optimization model is constructed by taking the minimum total deviation of the time headway on a bus line as a target. According to the method, factors such as departure time, driving speed and signalized intersection delay of intelligent networked buses are comprehensively considered, and the departure time of the buses at stations and the driving speed of the buses when the signalized intersection delay is not considered are optimized by establishing a calculation method for the bus delay at the signalized intersection. Meanwhile, according to a time headway balance principle, a calculation method based on the waiting time of passengers is provided so as to coordinate the operation efficiency of the motorcade and the experience of the passengers. According to the method, the fluctuation of the time headway is effectively reduced, the bunching effect is avoided, the overall operation efficiency and the service level of the bus route are remarkably improved, and theoretical support and technical guarantee are provided for design and implementation of an intelligent network connection bus system.

The invention relates to the technical field of intelligent monitoring, in particular to an intelligent bus kiosk monitoring method and system with an abnormal behavior intelligent monitoring function. The method comprises the following steps: acquiring multi-source monitoring data of a target bus kiosk, and based on the multi-source monitoring data, performing character and article linkage behavior analysis on the multi-source monitoring data to generate an entity interaction feature set; extracting a timestamp sequence and a space coordinate sequence based on the entity interaction feature set, and generating a space-time parameter set based on the timestamp sequence and the space coordinate sequence; based on the entity interaction feature set, according to the space-time parameter set, analyzing whether the current entity interaction feature has an abnormal behavior in the corresponding bus kiosk space-time scene, and if yes, extracting abnormal behavior information; and according to the abnormal behavior information, determining and outputting an abnormal early warning signal. According to the method, the abnormal characteristics of the behaviors under the bus kiosk are analyzed in a multi-dimensional mode, safety early warning is achieved, and the public traffic order is guaranteed.

The invention provides a method and device for predicting bus arrival time, electronic equipment, a medium and a program product, and relates to the field of data processing, in particular to the fields of intelligent transportation, smart cities, map big data and the like. According to the implementation scheme, predetermined arrival time information and reference arrival time information of a target bus for a to-be-predicted shift are obtained, and the predetermined arrival time information comprises predetermined time when the target bus arrives at all stations in the to-be-predicted shift; the reference arrival time information comprises the real time when the target bus arrives at at least part of stations in all stations in a historical shift corresponding to the to-be-predicted shift; correcting the predetermined arrival time information based on the reference arrival time information to obtain predicted arrival time information for the to-be-predicted shift; and determining the time when the target bus arrives at the target station based on the predicted arrival time information and the information of the target station provided by the user.

The invention discloses a multi-unmanned aerial vehicle distribution and bus charging combined path optimization method, which aims at minimizing task total time, constructs a mixed integer programming model based on a space-time network, and comprehensively considers unmanned aerial vehicle electric quantity constraint, demand point full coverage, bus time window and charging pile number limitation. An original model is decoupled to limit a main problem and a sub-problem by adopting branch pricing and a Dantzigzag-Wolfe decomposition theory, the main problem deals with demand coverage and charging resource allocation, and modeling is a set coverage problem; for single-machine path generation, the latter is modeled as a resource-constrained and replenishable shortest path problem with a time window dependent feature. A multi-unmanned aerial vehicle initial solution is constructed through a random generation method, a dual variable is iteratively solved after a main problem is initialized, a sub-problem is solved based on a multi-label algorithm, and global optimal solution search is realized in combination with a column generation mechanism and a branch strategy. According to the invention, through joint optimization of the departure time and path planning of the unmanned aerial vehicle, the collaborative optimization problem of distribution and charging is accurately solved.

The invention relates to a modular bus jump and relocation scheduling method considering stop time, and belongs to the field of urban public transport operation scheduling. The method comprises the following steps: acquiring geographical location information of a certain conventional bus route and a station and travel data information of bus passengers, and determining the number of passengers getting on and off the bus at the station; the method comprises the following steps: defining a jump station and relocation coupling operation as a decision variable, and establishing a relational expression between the stop time of a modular bus at a stop and the number of passengers getting on and off according to a modular bus single-door design characteristic and a bus combination form; taking the minimum sum of the passenger travel time cost and the vehicle operation cost as an objective function, setting a modular bus flow constraint and an operation time constraint, and constructing a modular bus scheduling optimization model; and solving to obtain an optimal modularized bus jump station and relocation scheduling scheme under the passenger demand of the corresponding line. According to the invention, for a bus route scene where part of bus stations are relatively high in load, modular bus operation scheduling which efficiently utilizes bus transport capacity resources is realized.

The invention discloses an intelligent bus shelter data processing method and system based on digital analysis, and the method comprises the steps: collecting the use information of seats in a target vehicle and the use state of an air conditioner, calculating the temperature distribution information and congestion degree information of different seats, and transmitting the information to all bus shelters for data sharing; receiving a request instruction of a target passenger, identifying a historical riding record and current riding demand information of the target passenger, and predicting a riding preference behavior and an expected distance from a demand seat to a get-off gate; and building a seat recommendation model according to the temperature distribution information, the congestion degree information, the riding preference behavior and the expected distance, generating a target seat for recommendation, and when the target passenger determines to use the target seat, updating the data, synchronizing the data to all bus shelter, and continuing to perform seat recommendation for the next passenger. Personalized requirements of different passengers are considered, seats can be scientifically and reasonably recommended for the passengers, the problem that the passengers frequently replace the seats after selecting the seats blindly is avoided, and the riding experience of the passengers is improved.

The invention relates to the technical field of image or video recognition or understanding, and discloses a bus driver driving state evaluation method, which comprises the following steps: acquiring a face image, an electrocardiosignal and voice audio data of a bus driver, performing timestamp alignment and preprocessing, and constructing a multi-mode driving state data set; extracting multi-modal features of the collected data, and mapping the multi-modal features to a unified feature space; fusing multi-modal features through a cross-modal attention mechanism, dynamically adjusting attention weight based on an emotional change difficulty index, and extracting emotional change key features; and predicting a two-dimensional continuous emotion value of the driver by using the fusion features, calculating a long-time-sequence emotion driving risk score based on an emotion stimulation dynamic model, and performing evaluation and early warning of a driving state. The problems of single-mode analysis, lack of long-period early warning and driving state static recognition in the prior art are solved, and the purposes of accurate evaluation, high safety, multi-mode fusion and long-time-sequence prediction are achieved.