A mixed traffic emergency vehicle space-time passing corridor cooperative control method and system

Abstract

The application discloses a kind of mixed traffic under emergency vehicle space-time passing corridor coordination control method and system, belong to intelligent traffic and vehicle-road coordination technical field.The application is classified to motor vehicle according to control authority and communication ability, and space-time occupation envelope of kinematic reachable set and allowable activity area constraint is generated in combination with communication time delay, motion prediction error;Without presetting fixed priority lane, passing path is divided into space-time passage unit and candidate corridor is generated, target release vehicle set is determined by incremental solution, and execution window is allocated based on action envelope space-time conflict;When complete sequence is unavailable, passageway is released in stages by safety waiting unit.The scheme is output control instruction and give way prompt after safety check, and passing corridor is rolling maintained, adapts mixed traffic and communication limited scene, and gives consideration to passing reliability and traffic disturbance control.

Description

Technical Field

[0001] This invention relates to the fields of autonomous driving, vehicle-road cooperation and intelligent traffic control technology, specifically to a method and system for coordinated control of the spatiotemporal passage corridor for emergency vehicles under mixed traffic conditions. Background Technology

[0002] With the continuous growth of urban road traffic flow, the problem of emergency vehicles such as ambulances, fire trucks, police cars, and emergency rescue vehicles encountering obstructions in congested road sections, intersections, tunnels, ramps, and highways is becoming increasingly prominent. Traditional yielding mainly relies on drivers observing police lights and sirens and then taking measures such as slowing down, changing lanes, or moving to the side of the road. However, the inconsistent response time and direction of vehicle actions can easily cause intermittent traffic flow, localized congestion, and secondary traffic risks.

[0003] Existing technologies can determine whether a single autonomous vehicle is obstructing traffic based on the predicted trajectory of emergency vehicles, and control the vehicle to change lanes, pull over, or stop. Other solutions (such as CN118411838B) are geared towards hybrid connected traffic environments, estimating or reconstructing the trajectory of manually driven vehicles based on the state of connected vehicles, combining this with traffic signal phase settings to establish time priority strategies, and organizing lane changes, speed guidance, and traffic restoration through preset priority lanes or dynamically cleared areas. While these solutions can address emergency priority passage issues in specific scenarios, their control targets are typically concentrated on signal phases, preset priority lanes, and their cleared areas.

[0004] In real-world driving, directly controllable vehicles, communicable but not directly controllable vehicles, and non-communicable vehicles may coexist. For non-communicable or unpredictable vehicles, if a single precise trajectory is reconstructed based on information from adjacent connected vehicles, communication latency, packet loss, positioning errors, and changes in human driving behavior may cause the reconstructed trajectory to deviate from the actual range of motion, thus misjudging the actually occupied space as free space; at the same time, the system cannot issue cooperative control commands to such vehicles.

[0005] Furthermore, control methods centered on fixed priority lanes or dynamically cleared areas may require vehicles that do not affect the actual passage of emergency vehicles to also perform lane changes or speed adjustments. When multiple vehicles simultaneously perform lateral offsets, lane changes, or forward releases, their target space and execution time may also compete with each other; relying solely on periodic rolling optimization is also insufficient to respond promptly to sudden changes in vehicle communication or control capabilities, decreases in lane width, and new spatiotemporal conflicts.

[0006] On the other hand, in congested intersections, if the prediction errors of non-communicating vehicles and vulnerable road users are directly expanded over a long time domain, multiple occupancy envelopes may quickly cover most of the drivable area, rendering all candidate corridors infeasible. Simultaneously, the combination of candidate corridor search, target release vehicle set optimization, and action conflict scheduling has high computational complexity. If a globally accurate solution is performed for each state change, it is difficult to meet the real-time control requirements of vehicle-side or roadside edge nodes. When controllable vehicles are surrounded by non-communicating vehicles, the complete spatiotemporal conflict relationship graph may not have a one-time executable sequence, requiring phased lane release under the condition that emergency vehicles can brake safely at any time.

[0007] Therefore, a collaborative control method is needed that does not rely on a single accurate trajectory reconstruction and fixed priority lane for non-communicating vehicles, can limit envelope expansion by road boundaries and dynamic reachability sets, adaptively shorten the prediction time domain according to uncertainties, reduce computation by using candidate vehicle screening, incremental solution and time-budgeted feasible solution output, and continuously advance corridor construction by using safe waiting units and phased release when the complete corridor or complete action sequence is not yet available. Summary of the Invention

[0008] To address the aforementioned technical problems, this invention provides a method for coordinated control of spatiotemporal passage corridors for emergency vehicles under mixed traffic conditions, comprising: Obtain status information of emergency vehicles and traffic participants within the target road area, and classify motor vehicles according to their control authority and communication capabilities; The spatiotemporal occupancy envelope of each traffic participant in the prediction time domain is generated by combining state uncertainty. The spatiotemporal occupancy envelope is expanded by uncertainty from the predicted occupancy area and restricted to the intersection of the dynamic reachability set and the allowed activity area of ​​the corresponding traffic participant type. When the uncertainty of the state exceeds a preset threshold, the effective prediction time domain is shortened and the subsequent spatiotemporal occupancy envelope is generated in a segmented rolling manner. Instead of pre-setting fixed priority lanes, the road space is divided along the priority passage path of emergency vehicles and corresponding estimated arrival time windows are set to form multiple spatiotemporal channel units. Based on the overlap relationship between the spatiotemporal occupancy envelope and the spatiotemporal channel unit, candidate spatiotemporal passage corridors are generated from different passable spatial zones; when there are no candidate spatiotemporal passage corridors that meet the conditions in the complete prediction time domain, segmented spatiotemporal passage corridors covering the current continuous confirmable area and ending at the safe waiting unit are generated. Based on the effective width, spatiotemporal continuity, and executable actions of the candidate spatiotemporal passageways, a target set of vehicles for releasing the passageway space is determined. Based on the spatiotemporal conflict relationship between the target spatiotemporal occupancy envelopes corresponding to the candidate actions of vehicles in the target release vehicle set, the execution order or execution window of the candidate actions is determined to form a collaborative control scheme. The cooperative control scheme is subjected to safety constraint verification, cooperative control commands are output to vehicles with control authority, yield prompts are output to vehicles with communication capabilities but no control authority, and the spatiotemporal occupancy envelope and spatiotemporal passage corridor are updated according to the updated traffic participant status.

[0009] Preferably, the generation of the spatiotemporal occupancy envelope specifically includes: The predicted occupied area is determined based on the predicted location, heading, and outline dimensions of traffic participants at the predicted time. The uncertainty expansion is determined based on data staleness, communication delay, positioning error, speed error, acceleration error, traffic participant type, and prediction duration, and the dynamic reachability set is determined based on the movement capabilities of traffic participants. The predicted occupancy area is geometrically expanded according to the uncertainty expansion amount, and the expansion result is restricted to the intersection of the dynamic reachability set and the allowed activity area corresponding to the traffic participant type to obtain the spatiotemporal occupancy envelope. The permitted activity areas for motor vehicles include motor vehicle lanes and authorized shoulder areas; the permitted activity areas for non-motor vehicles include non-motor vehicle lanes and mixed traffic areas; and the permitted activity areas for pedestrians include sidewalks and crosswalks.

[0010] Preferably, the generation of the spatiotemporal channel unit and the candidate spatiotemporal passage corridor specifically includes: The road space area is divided along the priority passage path according to the preset spatial length or road topology nodes, and the corresponding time window is set according to the estimated time of emergency vehicles to arrive at each road space area to form the spatiotemporal channel unit. The effective width of any spatiotemporal passage unit in the candidate spatiotemporal passage corridor shall not be less than the sum of the outer width of the emergency vehicle and the safety buffer distance on both sides; Adjacent spatiotemporal channel units have overlapping areas in space or meet the dynamic reachability conditions of emergency vehicles, and continuously cover the expected arrival range of emergency vehicles in time; The safety waiting unit is a spatiotemporal channel unit that accommodates emergency vehicles to stop or maintain low speed, meets safe braking distance and road boundary constraints, and maintains a preset safe distance from downstream unconfirmed conflict areas.

[0011] Preferably, determining the target set of released vehicles specifically includes: Candidate vehicles whose spatiotemporal occupancy envelopes overlap with candidate spatiotemporal passage corridors or their preset neighborhoods are selected, and the occupancy relationship between candidate vehicles and spatiotemporal passage units is established. The channel release contribution is calculated based on the target spatiotemporal occupancy envelope after the candidate vehicle completes its action. The channel release contribution is determined by at least one of the following: the number of newly added spatiotemporal channel units that meet the effective width condition, the increase in the effective width of the corridor, the increase in the continuous passable length, or the increase in the available time margin of the corridor. Under the conditions of satisfying the effective width, spatiotemporal continuity and safety constraints of the spatiotemporal channel unit, the first optimization objective is to minimize the number of actively participating vehicles, and the second optimization objective is to minimize traffic disturbance and action execution cost, and the target release vehicle set is determined. Within a preset computation time budget, the feasible solutions from the previous update cycle are reused in an incremental solution manner to generate an initial feasible set, and the initial feasible set is updated through heuristic search or local improvement. When the computation time budget expires, output the feasible set that has the minimum known cost and passes the security constraint check.

[0012] Preferably, the candidate actions include at least one of maintaining driving, deceleration, stopping, lateral movement within the lane, changing lanes to the left or right, pulling over to give way, and forward release; Based on whether the target spatiotemporal occupancy envelopes of candidate actions intersect, whether the candidate action trajectories cross, whether they compete for the same release space, and whether their execution time windows overlap, spatiotemporal conflict relationships are established between corresponding candidate actions. For candidate actions that have spatiotemporal conflicts, different start times, completion deadlines, or mutually exclusive execution windows are assigned, and execution priorities are determined according to the expected arrival order of emergency vehicles and the release order of spatiotemporal channel units. If there is no executable sequence covering all target actions within the preset calculation time budget, select a subset of conflict-free actions that are executable within the current time window and whose channel release contribution meets the preset conditions. First, release the preceding spatiotemporal channel units that emergency vehicles can continuously reach, and then redetermine the execution order or execution window for the remaining candidate actions.

[0013] Preferably, based on the vehicle's control authority and communication capabilities, motor vehicles are classified into directly controllable vehicles, vehicles that can provide communication prompts, and non-communication vehicles. Send at least one of the following cooperative control commands to the directly controllable vehicle: target speed, target acceleration, target lane, lateral offset, target trajectory, execution window, and failure time; Send a yielding prompt message to the communicable prompting vehicle, consisting of at least one of the following: suggested yielding direction, suggested speed, suggested execution time, and risk warning; Instead of sending cooperative control commands to the non-communicating vehicles, the spatiotemporal occupancy envelope is generated based on their perception state, motion boundary, and prediction error. This envelope is used as an uncertain dynamic constraint object, and the candidate spatiotemporal passage corridors and other vehicles' candidate actions are adjusted accordingly.

[0014] Preferably, the security constraint verification specifically includes: Collision safety verification is performed based on whether the spatiotemporal occupancy envelopes of any two traffic participants intersect or whether the minimum envelope distance is less than a preset safety threshold. Verify whether the vehicle's target spatiotemporal occupancy envelope is within the drivable area and has not entered the restricted area; verify whether the vehicle's longitudinal acceleration, lateral acceleration, yaw rate, and target trajectory curvature are within the allowable range of vehicle dynamics; Safety verification of vulnerable road users is conducted based on the probability of trajectory conflict, expected collision time, or degree of risk exposure. Verify the effective width and spatiotemporal continuity of candidate spatiotemporal passageways when emergency vehicles are expected to arrive at each spatiotemporal passage unit.

[0015] Preferably, the method further includes: Based on regional status coverage, communication latency, packet loss rate, data obsolescence time, and status consistency, the system switches between full regional collaborative mode, local communication collaborative mode, and vehicle-side perception local mode. Local incremental updates are performed using a first update frequency, and global re-solution is performed using a second update frequency lower than the first update frequency, in order to update the spatiotemporal occupancy envelope, related occupancy relationships, and spatiotemporal conflict relationships of traffic participants whose states have changed. When the actual state of traffic participants exceeds the corresponding spatiotemporal occupancy envelope, the effective width of the spatiotemporal channel unit is lower than the preset threshold, the communication capability or control authorization status of vehicles in the target release vehicle set changes, a new spatiotemporal conflict relationship occurs, or the expected arrival time deviation of emergency vehicles exceeds the preset threshold, the spatiotemporal passage corridor is updated. The mode switching adopts a hysteresis control mechanism, and the degradation threshold and recovery threshold are set with different values ​​to avoid frequent switching of control mode caused by communication link fluctuations.

[0016] Preferably, the method further includes a traffic restoration step: After the emergency vehicle has passed, the order or window for restoring vehicle speed and lane status is determined based on the current position of the participating vehicles, the available space in the target lane, and the spatiotemporal conflict between the restoration actions. Output collaborative control commands to vehicles with control authority, and output recovery prompt messages to vehicles with communication capabilities but no control authority; When a recovery action causes an overlap of the spatiotemporal occupancy envelope or a new traffic bottleneck, the recovery time of the corresponding vehicle is delayed or the current safe state is maintained.

[0017] The present invention also provides a spatiotemporal traffic corridor coordination control system for emergency vehicles under mixed traffic conditions, for implementing the method described above, the system comprising: The status and occupancy envelope processing module is used to obtain status information of emergency vehicles and traffic participants, and classify motor vehicles according to their control permissions and communication capabilities. The spatiotemporal occupancy envelope of each traffic participant in the prediction time domain is generated by combining state uncertainty. The spatiotemporal occupancy envelope is constrained to the intersection of the dynamic reachability set and the allowed activity area of ​​the corresponding traffic participant type after uncertainty expansion of the predicted occupancy region. When the uncertainty of the state exceeds a preset threshold, the effective prediction time domain is shortened and the subsequent spatiotemporal occupancy envelope is generated in a segmented rolling manner. The corridor and release vehicle set determination module is used to divide the road space area along the priority passage path of emergency vehicles without pre-setting fixed priority lanes and set corresponding expected arrival time windows to form a spatiotemporal channel unit. Candidate spatiotemporal passage corridors are generated from different passable space zones based on the spatiotemporal occupancy envelope, and a target set of release vehicles for releasing corridor space is determined. When there are no candidate corridors that meet the conditions within the complete prediction time domain, a segmented spatiotemporal passage corridor is generated, covering the current continuous verifiable area and ending at the safe waiting unit. The spatiotemporal conflict scheduling and safety verification module is used to determine the execution order or execution window based on the spatiotemporal conflict relationship between the target spatiotemporal occupancy envelopes of the candidate vehicle actions in the target release vehicle set, form a collaborative control scheme and perform safety constraint verification. The control output and corridor maintenance module is used to output collaborative control commands to vehicles with control authority, output yield prompts to vehicles with communication capabilities but no control authority, and continuously update the spatiotemporal occupancy envelope and spatiotemporal passage corridor based on the updated traffic participant status.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention does not rely on a single precise trajectory reconstruction of non-communicable vehicles. Instead, it converts communication delay, data obsolescence, positioning error, and motion prediction error into a spatiotemporal occupancy envelope that can be used for channel planning. This can reduce misjudgment of free space caused by the vehicle's actual motion deviating from the reconstructed trajectory.

[0019] This invention uses the permitted activity area and the dynamic reachability set of traffic participants corresponding to the type of traffic participant to prune the expanded envelope, and adjusts the prediction time domain according to the state confidence, so that the envelope covers the physically reachable area, while avoiding unnecessary closure of the entire congested area due to the accumulation of long time domain errors.

[0020] This invention does not require the complete clearing of fixed priority lanes. Instead, it generates multiple candidate corridors from different passable space zones and determines the target release vehicle set required to form and maintain the corridors by the occupancy relationship between vehicles and spatiotemporal channel units, thereby reducing the lane changing, deceleration and lateral deviation actions of irrelevant vehicles.

[0021] This invention continuously outputs safe and verified feasible solutions within a limited time by using candidate vehicle space screening, reuse of feasible solutions from the previous cycle, local incremental updates, and heuristic solutions with time budgets, thereby reducing the dependence of global combinatorial optimization on on-board or roadside computing resources in dense traffic scenarios.

[0022] This invention establishes a spatiotemporal conflict relationship based on the spatiotemporal overlap of the candidate action target occupancy envelope, and assigns mutually exclusive execution windows to conflicting actions, so that multiple vehicles execute actions in the order of channel unit release, reducing the risk of multiple vehicles entering the same space at the same time or lane change trajectories intersecting.

[0023] This invention outputs the largest conflict-free subset of actions when the complete action sequence is unavailable, and adopts a phased corridor mechanism of "preceding channel release - safety waiting unit - downstream asynchronous solution" to ensure that emergency vehicles retain verifiable braking and exit conditions in each stage, thereby improving degradation robustness under extreme congestion boundary scenarios.

[0024] This invention employs different processing methods for directly controllable vehicles, vehicles that can provide communication prompts, and vehicles that cannot communicate. Even in non-fully connected autonomous driving environments and communication-restricted scenarios, it can still maintain the spatiotemporal passage corridor for emergency vehicles on a rolling basis.

[0025] This invention combines fixed-period updates with event-triggered updates, promptly recalculating the corridor when there is insufficient corridor width, changes in communication capabilities or control authorization status, the emergence of new spatiotemporal conflicts, or when the expected arrival time deviation of emergency vehicles exceeds the limit; at the same time, it sets degradation thresholds and recovery thresholds to form hysteresis conditions, avoiding frequent switching of control modes caused by link fluctuations. Attached Figure Description

[0026] Figure 1 This is an overall flowchart of the method of the present invention.

[0027] Figure 2 This is a schematic diagram of the spatiotemporal channel unit and the spatiotemporal occupancy envelope.

[0028] Figure 3A diagram illustrating the relationship between the release of a vehicle set for a target and spatiotemporal conflicts. Detailed Implementation

[0029] The present invention will be further described below with reference to the accompanying drawings and embodiments. It should be understood that the embodiments described are used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention. The thresholds, weights, prediction time domain, and spatial division lengths can be configured according to road type, emergency vehicle size, vehicle performance, and communication conditions.

[0030] This invention applies to urban roads, signalized intersections, expressways, highways, tunnels, ramps, park roads, and closed test roads. Emergency vehicles include ambulances, fire trucks, police cars, emergency rescue vehicles, and other mission vehicles that are legally entitled to priority passage.

[0031] refer to Figure 1 The specific steps of the emergency vehicle spatiotemporal passage corridor collaborative control method under mixed traffic conditions provided by the present invention are as follows: S1. Status Information Acquisition and Vehicle Control Permission Classification: The system acquires status information of the target road area through at least one of the following: vehicle-mounted sensors, vehicle-to-vehicle communication, vehicle-to-infrastructure communication, roadside sensing, traffic signal status interface, edge computing nodes, and cloud platform. The status information includes vehicle ID, location, speed, acceleration, heading, lane, vehicle dimensions, control mode, communication status, road boundaries, lane lines, shoulders, speed limits, signal phase, obstacles, and pedestrian and non-motorized vehicle status. This invention uses traffic signal status as a constraint for time windows and traffic feasibility, without requiring proactive adjustment of traffic signal phase as a necessary step.

[0032] For each external status message, the system records the message generation time. Reception time Current calculation time The parameters include: last valid update time, communication latency, packet loss rate, signal quality, message integrity, and time synchronization error. Among these, communication latency... for: ; Status stale time after reception for: ; Communication delay and post-received state staleness time characterize transmission delay and post-message state lag, respectively, to avoid repeatedly including the same time in the uncertainty expansion.

[0033] The system classifies motor vehicles into directly controllable vehicles, vehicles that can provide communication prompts, and non-communication vehicles based on whether they have vertical or lateral control authority, whether they can receive structured instructions, and whether they can provide feedback on execution status. Vehicle classification can be dynamically updated according to communication link and control authorization status.

[0034] Directly controllable vehicles include those controlled by the same autonomous driving system, vehicle-to-infrastructure (V2I) control platform, or authorized vehicle-mounted controller; vehicles capable of receiving yield prompts can receive yield prompts, but the final action is determined by the vehicle's own system or a human driver; non-communicable vehicles include those not connected to the collaborative system, those whose messages have timed out, and those whose identity or control authorization status cannot be confirmed.

[0035] For pedestrians, cyclists and other non-motorized vehicles, the system does not output cooperative control commands. Instead, it generates a spatiotemporal occupancy envelope based on their current location, speed, orientation, activity area and behavior prediction, and sets higher risk weights or safety buffers in corridor generation and safety verification.

[0036] S2. Emergency vehicle identification and priority route prediction: The system identifies emergency vehicles based on at least one of the following: a certified V2X priority passage request, a traffic management platform task, roadside perception results, vehicle type, or a warning light or siren signal. To reduce the risk of false triggering, the system may require the authentication result and the perception result to be consistent within a preset time window.

[0037] When an emergency vehicle is less than a preset distance threshold from the target area, its estimated arrival time is less than a preset time threshold, or a valid priority passage request is received, the spatiotemporal passage corridor construction is initiated. This triggering process uses preliminary road occupancy information and does not depend on subsequent refined target release vehicle set results.

[0038] The system predicts the priority travel path in the future prediction time domain based on the emergency vehicle's current location, speed, heading, target direction, road topology, and mission route, and uses the current traffic signal status as a constraint condition for the estimated arrival time and time window. The priority travel path can be represented as a lane sequence, road centerline, trajectory point sequence, or drivable area.

[0039] When the relative speed of an emergency vehicle to the vehicle ahead along the path exceeds a preset threshold, the estimated approach time is... It can be represented as: ; The longitudinal distance measured along the priority passage path. and These are the speed components along the path for the emergency vehicle and the vehicle ahead, respectively. When the relative speed is not greater than a threshold, the estimated approach time is set to infinity or triggered only by distance and mission request conditions.

[0040] S3. Generation of spatiotemporal occupancy envelope for communication uncertainties: The state uncertainty includes data staleness, communication delay, positioning error, speed error, and acceleration error. Let traffic participants... The predicted area at prediction time τ is The predicted occupancy area is determined by its predicted motion state, heading, and vehicle outline or human occupancy area. For non-communicating vehicles, the system does not reconstruct a single precise trajectory based on the trajectories of adjacent connected vehicles, but instead forms one or more possible occupancy areas based on local perception state, permitted activity area, and motion boundary. Simultaneously, the system generates a dynamic reachable set based on the maximum achievable speed, acceleration, deceleration, turning rate, or walking motion boundary of the traffic participant, then geometrically expands the predicted occupancy area and intersects it with the dynamic reachable set and the permitted activity area corresponding to the traffic participant type to obtain the spatiotemporal occupancy envelope. ; For the spatiotemporal occupancy envelope, For the radius or anisotropy scale Uncertainty expansion region Let be the dynamic reachable set of traffic participants at the predicted time τ. For the permitted activity area corresponding to the type of traffic participant, ⊕ represents geometric expansion or Minkowski expansion. The permitted activity area for motor vehicles includes motor vehicle lanes, authorized shoulders, or other motor vehicle-accessible areas; the permitted activity area for non-motorized vehicles includes non-motorized vehicle lanes, mixed-traffic areas, or permitted crossing areas; the permitted activity area for pedestrians includes sidewalks, crosswalks, safety islands, and other permitted activity areas. For vehicles, different expansion amounts can be set along both the longitudinal and lateral directions.

[0041] In one implementation, the uncertainty expansion is calculated according to the following relationship: ; The base buffer size is related to the category of traffic participant. The stale time for the status after message reception. For communication delay, , and These are the error limits for positioning, velocity, and acceleration, respectively. The weighting coefficients for the corresponding state aging time; This is the weighting coefficient corresponding to the communication delay; These are the weighting coefficients corresponding to the positioning error bound; These are the weighting coefficients corresponding to the velocity error boundary; These are the weighting coefficients corresponding to the acceleration error boundary. , , , , It has corresponding units that convert all terms into length quantities, and each error quantity can be determined using an upper bound, confidence interval, or historical statistical value.

[0042] When messages are continuously lost, external information is inconsistent with the vehicle's perceived information, or the intentions of traffic participants are unclear, the system increases the longitudinal or lateral expansion, but the spatiotemporal occupancy envelope is still limited by the dynamic reachable set and the allowed activity area corresponding to the type of traffic participant. When stable and consistent information is obtained in multiple consecutive update cycles, the system gradually reduces the expansion to avoid abrupt envelope changes. For objects whose uncertainty exceeds a preset threshold, the system shortens its effective prediction time domain, uses a high-confidence envelope only for the current safe braking distance of emergency vehicles and several adjacent spatiotemporal channel units, and continuously supplements the downstream envelope in subsequent update cycles.

[0043] S4. Spatiotemporal channel unit division: Combination Figure 2 , Figure 2 This is a schematic diagram of the spatiotemporal channel unit and the spatiotemporal occupancy envelope. Here, t1 and t2 represent different update times; , , These represent the 1st, 2nd, and 3rd spatiotemporal channel units in candidate spatiotemporal travel k, respectively. , , These are the 1st, 2nd, and 3rd ordinary vehicles, respectively. The diagram illustrates the process from t1 to t2, where ordinary vehicles free up passage space by lateral maneuvering or moving forward, allowing emergency vehicles to enter subsequent spatiotemporal passage units along the priority direction indicated by the arrows. The system divides the priority path into multiple road regions according to preset spatial lengths or road topology nodes, and sets corresponding time windows based on the estimated arrival time of emergency vehicles in each road region. The m-th spatiotemporal passage unit is shown in the k-th candidate path. It can be represented as: ; This refers to the road space area. The time window for emergency vehicles to occupy the area is defined, and this time window can be expanded based on fluctuations in emergency vehicle speed and control delays.

[0044] Adjacent spatiotemporal channel units should have overlapping areas in the road space or meet the dynamic reachability conditions for emergency vehicles. The interval between the end of the time window of the preceding spatiotemporal channel unit and the beginning of the time window of the following spatiotemporal channel unit should not exceed a preset time threshold to continuously cover the estimated time for emergency vehicles to enter the next unit from the previous unit. For intersections, ramps, and forked roads, one spatiotemporal channel unit can correspond to multiple candidate successor units.

[0045] S5. Generation and continuity determination of candidate spatiotemporal passageways: The system searches for available space for emergency vehicles within each spatiotemporal channel unit based on road boundaries, lane width, shoulders, traffic dividers, traffic signal status, restricted areas, and the spatiotemporal occupancy envelopes of each traffic participant. Candidate corridors are not pre-defined as fixed priority lanes; they can be located in the center, left, right, authorized shoulders, or feasible passage areas at intersections, or formed by connecting segments of space corridors at different locations. The system prioritizes generating two or more candidate corridors with different spatial locations or connection methods before determining their continuity.

[0046] The effective width of the candidate corridor at any prediction time It should meet the following requirements: ; For the outer width of emergency vehicles, This is the single-sided safety buffer distance for emergency vehicles. The effective width is calculated based on the drivable area of ​​the road after deducting the spatiotemporal occupancy envelope expanded by each traffic participant.

[0047] The continuity of candidate corridors requires that adjacent corridor units have spatial overlap or a connecting area that is dynamically reachable by emergency vehicles, and that there are no uncoverable discontinuities in the time window. If a unit can only be released briefly before or after the expected arrival of an emergency vehicle, then that unit does not satisfy the spatiotemporal continuity requirement.

[0048] The system performs a preliminary ranking of multiple candidate corridors based on the number of vehicles that need to move, the expected delay, the corridor width margin, and the uncertainty margin. However, traffic signal phase adjustment or clearing of the entire preset priority lane is not a necessary condition for the formation of a corridor. The preliminary score does not replace the subsequent hard constraint verification, and all indicators are normalized to the same range before participating in the weighting.

[0049] If none of the candidate corridors within the complete prediction time domain meet the constraints, the system does not immediately classify the entire target area as permanently infeasible. Instead, it searches for continuously verifiable preceding passage units within the safe braking distance of emergency vehicles to generate segmented spatiotemporal passage corridors. The terminals of these segmented corridors are selected as safe waiting units. These safe waiting units can accommodate emergency vehicles stopping or maintaining low speeds, satisfying safe braking distance and road boundary constraints, and maintaining a preset safe distance from downstream unconfirmed conflict areas. The system continues to update the downstream envelope and solve for subsequent corridors while the emergency vehicle passes through the preceding passage unit.

[0050] S6. Target release vehicle set determined: Combination Figure 3 , Figure 3 This diagram illustrates the relationship between the target vehicle set and spatiotemporal conflicts. The system first identifies potential obstructing vehicles based on the occupancy relationship between candidate vehicles and spatiotemporal channel units in candidate corridor k, where black dots represent 1. Then, by combining action executability, channel release contribution, and safety constraints, the target vehicle set for release is determined. Next, candidate actions are generated for the target vehicle, and a spatiotemporal conflict relationship graph of the candidate actions is established. In the graph, nodes represent candidate actions, and lines represent spatiotemporal conflicts between actions. Finally, mutually exclusive execution windows are assigned to conflicting actions, and the execution order is determined. , These represent the 4th and 5th spatiotemporal channel units in candidate spatiotemporal travel k, respectively. , , These are the 4th, 5th, and 6th regular buses, respectively. , These represent the first and second candidate actions for releasing vehicle 1, respectively. This indicates the first candidate action for releasing vehicle 3. , These represent the first and second candidate actions for releasing vehicle 4, respectively. , , , , These are the 1st, 2nd, 3rd, 4th, and 5th mutually exclusive execution windows assigned to conflicting actions. For candidate corridor k, the system first uses spatial indexing or occupancy grids to filter vehicles whose envelopes intersect with the candidate corridor and its safe neighborhood, and only processes the filtered candidate vehicles. Establish an occupancy relationship with the spatiotemporal channel unit m. If the vehicle... Spatiotemporal occupancy envelope and channel unit If they overlap, they occupy matrix elements. Select 1 if the vehicle is far from the candidate corridor and cannot reach its neighborhood within the prediction time domain. Vehicles that are far from the candidate corridor and cannot reach their neighborhood within the prediction time domain will not be included in the current cycle's combinatorial optimization.

[0051] The system generates candidate actions that reduce lane occupancy for directly controllable vehicles and vehicles with communicable prompts, and calculates the release contribution of the corresponding lane unit after the action is completed. The lane release contribution can be determined by at least one of the following: the number of newly added spatiotemporal lane units that meet the effective width condition, the increase in the effective width of the corridor, the increase in the continuous passable length, or the increase in the available time margin of the corridor. Non-communicable vehicles cannot be assumed to perform the actions specified by the system. The system does not reconstruct a single precise trajectory for them, but treats the spatiotemporal occupancy envelope formed by local perception, motion boundaries, and prediction errors as an uncertain dynamic constraint object.

[0052] Under the conditions of satisfying the effective width, spatiotemporal continuity, road boundary, and safety constraints of each spatiotemporal channel unit of the candidate corridor, the target release vehicle set can be determined by the following hierarchical optimization: First layer: ; Second layer: Under the constraint of the minimum number of actively participating vehicles obtained in the first layer. ; For the first A binary variable indicating whether a vehicle is selected into the active participation set. For the selected vehicle's action combination, The cost of traffic disruption The cost of executing the action, These are the weighting coefficients corresponding to the costs of traffic disruption; These are the weighting coefficients for the execution costs of the corresponding actions. Through hierarchical optimization, the number of vehicles actively participating is first determined, and then traffic disturbances and action execution costs are compared among feasible sets with the same number of vehicles.

[0053] The real-time solution for the target vehicle release set adopts a hierarchical approach: First, the vehicle set and actions that still satisfy the safety constraints from the previous update cycle are reused as a hot start, and a heuristic feasible initial solution is generated based on the vehicle channel release contribution, action executability, and conflict degree; then, the initial solution is improved using local search, restricted branch search, integer programming, or equivalent methods within the remaining computation time. The system sets a computation time budget for each solution; when the budget expires, the search immediately stops and outputs the feasible solution with the lowest known cost that passes the safety constraint check; the search for the globally optimal solution continues only when computational resources are sufficient or during non-real-time planning stages.

[0054] When a candidate corridor cannot be formed in one go due to non-communicable vehicles occupying it, the system identifies the bottleneck vehicle and bottleneck lane unit causing the continuity interruption. Instead of outputting cooperative control commands to the non-communicable vehicles, the system executes the following sequentially: selecting other spatial zones that bypass the non-communicable vehicle's spatiotemporal occupancy envelope; selecting a phased vehicle set that can release the most consecutive preceding lane units; sending directional yielding prompts or area clearing prompts to communicable vehicles and road management equipment; setting the segmented corridor terminal as a safe waiting unit; and reconfirming the downstream corridor before the emergency vehicle enters the next phase. Emergency vehicles are only triggered to stop at minimum risk if no preceding lane meets the safety braking and road boundary constraints.

[0055] S7. Generation of candidate yield actions and target spatiotemporal occupancy envelopes: For directly controllable vehicles in the target release vehicle set, the system can generate candidate control actions such as maintaining traffic flow, deceleration, stopping, lateral lane departure, changing lanes to the left or right, yielding, and forward release. Forward release refers to a yielding action where the vehicle in front accelerates or passes through the intersection ahead of time to make way for emergency vehicles behind. In intersection scenarios, forward release is only performed when the current traffic signal allows passage, traffic management authorization is obtained, or the vehicle can safely complete the action before the stop line. If these conditions are not met, deceleration, stopping, or lateral space release actions are used. Emergency vehicles themselves can generate candidate actions such as maintaining traffic flow, decelerating, or deviating from their lane within the corridor.

[0056] For vehicles that can communicate with the system, the system generates a yielding prompt message consisting of a suggested yielding direction, suggested speed, suggested execution time, and risk warning. Since the system does not have direct control authority, the execution probability of the action can be set during planning based on the historical response rate of this type of vehicle, confirmation messages, or the vehicle's trajectory feedback, and the corresponding spatiotemporal occupancy envelope is retained for cases where execution is not confirmed.

[0057] The system predicts the target spatiotemporal occupancy envelope for each candidate action during execution and after completion. When two candidate actions encounter at least one of the following situations within an overlapping time window: target envelope intersection, lane change trajectory intersection, competition for the same lane gap, or the same release space, a spatiotemporal conflict relationship is established between the corresponding actions. This spatiotemporal conflict relationship is represented by a spatiotemporal conflict relationship diagram, as shown below: ; This represents a spatiotemporal conflict graph, where A is the set of candidate actions and E is the set of spatiotemporal conflict edges. and Actions and The execution time window and Actions and The target spatiotemporal occupancy envelope at time τ. In addition to the intersection of target envelopes at the same time, traffic rule conflicts, execution order dependencies, and insufficient target gaps can also form spatiotemporal conflict relationship edges.

[0058] The system assigns different start times, execution durations, and completion deadlines to conflicting actions based on the expected arrival order of emergency vehicles and the release order of lane units. Actions that are closer to the emergency vehicle and contribute more to the release of the current lane unit are given priority; subsequent vehicle actions are initiated after the preceding release space is formed and safety conditions are met.

[0059] When the deviation between the actual execution state and the planned state of a vehicle exceeds a threshold, the system deletes or updates the relevant action nodes and spatiotemporal conflict relationship edges, and recalculates the execution window for subsequent actions. This update does not require vehicles that have entered the safe execution phase to immediately reverse their actions, in order to avoid frequent switching.

[0060] When there are conflict clusters or circular dependencies in the spatiotemporal conflict graph that prevent all target actions from being completed within the specified window, the system does not force all actions to be satisfied simultaneously. Instead, it selects the largest executable subset of conflict-free actions based on the channel unit release contribution, action completion time, and emergency vehicle arrival order. After completing the current subset, the system deletes the resolved spatiotemporal conflict edges with the actually formed release space as the new state and reschedules the remaining actions. If the current subset can only form a partial corridor, its terminal is aligned with the aforementioned safety waiting unit.

[0061] S8. Security constraint verification based on spatiotemporal occupancy envelope: The system performs safety constraint verification on the cooperative control scheme, using the spatiotemporal occupancy envelope instead of just the vehicle center point trajectory for collision verification. For the first... Traffic participants and the For each traffic participant, calculate the minimum distance between their envelopes within the prediction time domain. : ; and They represent the first Traffic participants and the Each traffic participant at time The spatiotemporal occupancy envelope is used. When two envelopes intersect or the minimum envelope distance is less than a safety threshold associated with the traffic participant category, the corresponding scheme is eliminated. Different safety thresholds can be set for vehicle-to-vehicle, vehicle-to-pedestrian, and vehicle-to-non-motorized vehicle interactions.

[0062] Road boundary verification requires that the target occupancy envelope of vehicles at all predicted times be within the drivable area and that they do not enter construction zones, median zones, oncoming lanes, or other prohibited areas. Implementation methods that allow temporary use of the shoulder should be explicitly authorized by traffic rules, road management instructions, or scenario configuration.

[0063] Vehicle dynamics verification includes longitudinal acceleration, lateral acceleration, yaw rate, steering angle, target trajectory curvature, and braking distance. Any action exceeding the reachability boundaries of the corresponding vehicle is prohibited.

[0064] Risk for vulnerable road users is calculated on a case-by-case basis, rather than simply summed by category. For vulnerable road users, the risk value is determined based on their category weight, trajectory conflict probability, estimated collision time, time difference after passing, and exposure level. If the risk exceeds a threshold, the system expands the corresponding envelope, stops lateral movement of adjacent vehicles, or selects an alternative corridor.

[0065] Emergency vehicle delay This can be expressed as the actual estimated transit time. Reference time with unobstructed state The non-negative value of the difference: Road boundaries, dynamic reachability, and motion conflicts are considered hard constraints; the system calculates the normalized comprehensive cost only after all hard constraints are satisfied. : superscript This indicates that the corresponding indicator has been normalized. This is the normalized emergency vehicle delay index. To mitigate collision risk, Risks for vulnerable road users The cost of traffic disruption To actively participate in the number of vehicles, Cost of switching actions Indicators for the spatiotemporal continuity and width margin of the corridor; For emergency vehicle delay weighting, For collision risk weighting, Risk weights for vulnerable traffic participants As a weight for traffic disturbance costs, To actively participate in the weighting of vehicle quantity, Weight the cost of switching actions. The weights for the corridor's spatiotemporal continuity and width margin indicators are determined. The system selects the cooperative control scheme that minimizes J(P).

[0066] S9. Hierarchical instruction output and execution feedback: For directly controllable vehicles, the system outputs at least one of the following: target speed, target acceleration, target lane, lateral offset, target trajectory, execution window, priority, and failure time. Cooperative control commands should include an event identifier and action version number so that the vehicle rejects expired or duplicate commands.

[0067] For vehicles that can communicate with the system, the system outputs a yielding prompt message consisting of a suggested yielding direction, suggested speed, suggested start time, and risk level. Whether the vehicle executes the yielding prompt is determined by the vehicle's autonomous driving system or the driver; the system judges the actual response based on confirmation messages, trajectory changes, or visual perception, and updates the spatiotemporal occupancy envelope accordingly.

[0068] For non-communicating vehicles, the system only outputs evasive actions for other directly controllable vehicles or warnings for road management equipment, without assuming to send cooperative control commands to non-communicating vehicles, and treats them as uncertain dynamic constraint objects.

[0069] The system periodically provides feedback on the vehicle's current location, speed, movement stage, and estimated completion time. It calculates the deviation between the actual and planned states; if the deviation exceeds a threshold, a communication message times out, or a new traffic participant enters the corridor's influence area, a rolling replanning is triggered.

[0070] S10, Collaboration Mode Switching and Scroll Corridor Maintenance: The system switches between full-area collaborative mode, local communication collaborative mode, and vehicle-end perception local mode based on regional status coverage, effective communication vehicle ratio, average latency, packet loss rate, status obsolescence time, and consistency with external information.

[0071] The full-area collaborative mode uses area information provided by the vehicle, roadside, edge, or cloud to generate corridors across multiple road units; the local communication collaborative mode uses one-hop communication or local communication cluster information to limit the corridor range to areas with verifiable status; the vehicle-aware local mode only controls the vehicle itself or directly authorized vehicles and treats other traffic participants as uncertain dynamic constraints.

[0072] When any quality indicator falls below the degradation threshold, the system expands the occupancy envelope of relevant traffic participants, shortens the planning range, reduces the target speed and lateral movement amplitude, and re-solves for the target release vehicle set.

[0073] After communication is restored, the entire safety margin is not immediately reduced. Instead, the spatiotemporal occupancy envelope is gradually reduced and the vehicle's perception state is consistent with the recovery threshold after N consecutive update cycles, and the vehicle is restored to the previous level of cooperative mode. Different values ​​are used for the degradation threshold and the recovery threshold to form a hysteresis interval, avoiding frequent switching caused by link fluctuations.

[0074] Rolling maintenance employs a multi-rate and incremental computation mechanism: The spatiotemporal occupancy envelopes of traffic participants whose states have changed, the occupancy relationships of relevant spatiotemporal channel units, and the edges of local spatiotemporal conflict relationships are updated at a first update frequency. Global re-solution of candidate corridors and target release vehicle sets is then performed at a second update frequency lower than the first update frequency. Unaffected envelopes, occupancy relationships, action nodes, and execution windows that have passed safety constraint checks are reused. Local incremental updates are triggered immediately when the actual state of any traffic participant exceeds the corresponding spatiotemporal occupancy envelope, the effective width of any spatiotemporal channel unit is lower than a threshold, the communication capability or control authorization status of vehicles in the target release vehicle set changes, new spatiotemporal conflict relationship edges are generated between candidate actions, the estimated arrival time deviation of emergency vehicles exceeds a threshold, or a new traffic participant enters the corridor's influence range. A computation time budget is set for each global re-solution; if the budget expires before an improved solution is completed, the most recently feasible solution that has passed safety constraint checks and still satisfies the current state is executed, or the process switches to a segmented corridor solution ending at a safety waiting unit. When the next channel unit cannot be released on time, the system adjusts the subsequent vehicle actions in advance or controls the emergency vehicle to slow down in order to avoid entering a local space without continuous exits.

[0075] S11. Emergency vehicle passage assessment and traffic restoration: When emergency vehicles have passed the target road area, priority passage requests have failed, mission routes have changed, or corridor maintenance conditions have been lifted, the system enters the traffic recovery phase.

[0076] The recovery phase does not simply involve returning all vehicles to their original lanes simultaneously. Instead, it involves rearranging the recovery sequence based on the vehicles' current positions, the clearance in front of and behind the target lane, the target envelope of the recovery action, and the spatiotemporal conflict graph. Recovery actions with spatiotemporal conflict edges are completed within different execution windows.

[0077] Directly controllable vehicles receive coordinated control commands to restore speed, return to the correct lane, or maintain the current safe lane; vehicles with communication capabilities receive restoration prompts; non-communication vehicles continue to act as objects of uncertain dynamic constraints. If returning to the original lane would cause new conflicts or bottlenecks, the vehicle can maintain its current lane and continue driving.

[0078] The method of this invention is implemented through a spatiotemporal traffic corridor collaborative control system for emergency vehicles under hybrid traffic conditions. The system can be deployed on vehicle-side controllers, roadside units, edge computing nodes, cloud-based traffic management platforms, or combinations thereof. The system includes a status and occupancy envelope processing module, a corridor and release vehicle set determination module, a spatiotemporal conflict scheduling and safety verification module, and a control output and corridor maintenance module. Specifically, the status and occupancy envelope processing module acquires status information, classifies vehicle control permissions, predicts priority passage paths, and generates spatiotemporal occupancy envelopes; the corridor and release vehicle set determination module divides spatiotemporal channel units, generates candidate spatiotemporal passage corridors, and determines the target release vehicle set; the spatiotemporal conflict scheduling and safety verification module establishes spatiotemporal conflict relationships, allocates execution order or execution windows, and performs safety constraint verification; the control output and corridor maintenance module outputs collaborative control commands or yielding prompts in a tiered manner, monitors communication status, switches collaborative modes, continuously maintains corridors, and restores traffic.

[0079] The status and occupancy envelope processing module uniformly manages data sources, timestamps, communication status, and control permissions, and dynamically adjusts the expansion amount based on the credibility and aging of each data source; the corridor and release vehicle set determination module only assigns active actions to the target release vehicles required to form the corridor; the spatiotemporal conflict scheduling and safety verification module ensures that the actions of multiple vehicles do not compete with each other in time and space and completes safety constraint verification; the control output and corridor maintenance module realizes hierarchical output of vehicles with different control permissions and communication capabilities and the rolling maintenance of the corridor.

[0080] To meet real-time requirements, envelope updates, occupancy relationship updates, and local conflict detection can be performed in parallel on the vehicle or roadside. Global ranking of candidate corridors and improvement of feasible solutions can be performed asynchronously by edge nodes. Each node shares the feasible solutions and version numbers of the previous cycle. If an edge node fails to return a new solution within the computation time budget, it does not cover the currently verified safety solutions on the vehicle or roadside.

[0081] In vehicle-mounted deployment, the vehicle-mounted system primarily controls the vehicle itself and receives emergency vehicle requests and other vehicle statuses via vehicle-to-vehicle communication. In roadside or edge deployment, the roadside system can issue collaborative control commands to authorized vehicles and broadcast yield prompts to unauthorized vehicles. In cloud deployment, the cloud is used for cross-regional path and task management, while real-time vehicle control is performed at the edge or on the vehicle. Traffic signal status can be used as a time window constraint input, but this system does not require early red light termination, extended green light, or other phase adjustments to achieve the corridor control.

[0082] The system can convert raw perception data into the status of traffic participants in a unified coordinate system and use map matching to determine lane relationships. When external information and the vehicle's perception information are inconsistent in terms of position, speed, or lane occupancy, a more conservative value or an expanded occupancy envelope can be used instead of directly overriding the vehicle's perception results.

[0083] The following will further elaborate with typical examples: Example 1: Urban multi-lane mixed traffic scenario: This embodiment is an exemplary traffic simulation used to illustrate the execution process of the present invention. The target area is a one-way three-lane urban road, with an emergency vehicle approaching a congestion queue along the middle lane. A total of ten motor vehicles are identified within the area, of which four are directly controllable autonomous vehicles, three are vehicles that can receive V2X prompts but are not directly controlled by the system, and three are manually driven vehicles that cannot communicate. There is also one pedestrian and one non-motorized vehicle on the side of the road.

[0084] The system generates a spatiotemporal occupancy envelope based on the message timestamps and perception status of each vehicle. For one vehicle that can communicate, the message is 0.8 seconds outdated, so the system increases its longitudinal envelope accordingly. For the two non-communicating vehicles that are tracked only by roadside vision and their own sensors, the system sets a larger lateral uncertainty for them.

[0085] The system divides the emergency vehicle's future 180-meter path into six spatiotemporal channel units, generating two candidate solutions: a central corridor and a right-side corridor. The central corridor requires eight vehicles to take action; the right-side corridor, after avoiding two non-communicating vehicles, only requires three directly controllable vehicles and one communicable prompting vehicle to release space. The target release vehicle set is determined by selecting the right-side corridor and four participating vehicles.

[0086] Three directly controllable vehicles respectively execute deceleration to widen the gap, change lanes to the right, and drift left within the lane. The system can communicate with the vehicles to indicate they have received yielding instructions to pull over and decelerate, but retains the original spatiotemporal occupancy envelope until the vehicles confirm their execution. Because the target envelopes of the lane-changing vehicle and the lateral drifting vehicle overlap, a spatiotemporal conflict relationship edge is established between them in the spatiotemporal conflict graph. Deceleration to create a gap is executed first, followed by the lane change and lateral drift.

[0087] Before an emergency vehicle arrives at each spatiotemporal channel unit, the system continuously verifies the effective width and time coverage of the corresponding unit. If a vehicle fails to fully move to the edge as suggested by communication, the system maintains the corridor by reducing the emergency vehicle's target speed and increasing the offset of adjacent directly controllable vehicles, without sending cooperative control commands to non-communication vehicles. After the emergency vehicle passes, the system restores vehicle speed and lanes in stages according to the restored spatiotemporal conflict relationship graph.

[0088] Example 2: Signalized intersection and vulnerable road users scenario: The target area is a crossroads where an emergency vehicle is expected to reach the stop line in twenty seconds. There is a queue of vehicles at the intersection entrance, and pedestrians and cyclists are on the crosswalk. The system reads the existing traffic signal phases and uses them as constraints for the estimated arrival time and passage window, setting up multiple spatiotemporal channel units along the entrance lanes, the intersection interior, and the exit lanes. This embodiment does not actively adjust the traffic signal phases, nor does it pre-designate fixed priority lanes.

[0089] For spatiotemporal passage units near pedestrian crossings, the system generates an expanded envelope based on pedestrian walking speed, orientation, reachable area, and possible stopping area, and restricts this envelope to the sidewalk, crossing area, and their dynamic reachable range. For non-communicating vehicles in a queue, the system shortens the effective prediction time domain based on low-speed following status and road boundaries to avoid expanding all possible long-term actions to the entire intersection at once. If the high-confidence envelope of pedestrians or non-motorized vehicles still overlaps with the recent arrival window of emergency vehicles, the corresponding passage unit is determined to be infeasible at the current stage.

[0090] The system generates multiple candidate corridors from the central space zone of the approach lane, the connecting space zone between adjacent lanes, and the segmented bypass space zone, and establishes occupancy relationships only for vehicles whose envelopes overlap with the neighborhood of the top three high-priority candidate corridors. The solver reuses feasible solutions from the previous cycle and first generates a heuristic initial solution that satisfies safety constraints; if further optimization is not completed within the preset computation time budget, the currently known feasible solution with the lowest cost and passing the safety constraint verification is directly used. If a complete intersection corridor is still unavailable, the system first forms a first-stage corridor from the current position of the emergency vehicle to the safe waiting unit before the stop line. When the current traffic signal allows passage, traffic management authorization is obtained, or the vehicle at the front of the queue can safely complete its action before the stop line, a directly controllable vehicle is arranged to release forward; if the above conditions are not met, deceleration, stopping, or lateral space release actions are adopted, and directional yielding prompts are sent to communicable prompting vehicles and roadside information equipment.

[0091] When a roadside sensor detects that a pedestrian's dwell time exceeds the predicted value, the corresponding envelope is updated within its physically reachable area and a shorter prediction time domain. The system keeps the emergency vehicle in the first-stage safe waiting unit and recalculates the second-stage corridor from the intersection to the exit lane. If a complete scheduling sequence does not yet exist in the spatiotemporal conflict relationship graph, the largest conflict-free subset of actions that can release the continuous space on the exit side is executed first. Subsequent actions on the entrance side are activated only after the pedestrian envelope leaves the conflict zone. The system outputs a minimum-risk stopping instruction only when neither the first-stage corridor nor the safe waiting unit can satisfy braking and boundary constraints.

[0092] After an emergency vehicle passes, the system continues to follow the original traffic signal control logic and regenerates the recovery action according to the vehicle's current position, without requiring all vehicles to return to their original lanes before coordination.

[0093] Example 3: Communication scenario on a highway: The target area is a one-way three-lane highway with an emergency lane on the right. The fire truck approaches the slow-moving traffic flow along the middle lane. Some vehicles in the area have vehicle-to-vehicle communication capabilities, while others can only be tracked by roadside sensors. A large truck is also obstructing the view of vehicles behind it.

[0094] The system expands the occupancy envelope of the truck and the unknown area behind it based on the perception uncertainty caused by the truck's size and obstructions. Since the central corridor overlaps with the envelope, the system does not select the option requiring the truck to make an emergency lane change, but instead generates a segmented spatiotemporal corridor formed by connecting the right emergency lane and a partial middle lane.

[0095] The target release vehicle set includes two directly controllable vehicles and two communicable prompting vehicles. The directly controllable vehicle in front performs a forward release, while the directly controllable vehicle behind decelerates to create an entry gap; the communicable prompting vehicles receive yielding prompts to maintain lane position and reduce speed. The system sets dynamic safety gaps based on the relative speed in high-speed scenarios and adds a larger longitudinal occupancy envelope to lane-changing actions.

[0096] When a communicable vehicle fails to provide confirmation of execution, the system does not consider it to have released space. Instead, it preserves the original envelope and delays lane changes by adjacent directly controllable vehicles. The subsequent action window is only activated after the system detects that the vehicle has actually decelerated and the target clearance has met the requirements.

[0097] After the fire truck passes, the system re-plans the recovery sequence according to the relative speed of high-speed vehicles and the safety clearance, to prevent participating vehicles from returning to the main lane at the same time.

[0098] Example 4: Communication degradation, recovery, and cross-regional connectivity scenarios: The objective involves crossing two roadside cooperative zones. When an emergency vehicle enters the first zone, the system operates in full-zone cooperative mode; however, as it approaches the zone boundary, some roadside links experience increased latency and continuous packet loss.

[0099] When communication metrics fall below the degradation threshold, the system expands the spatiotemporal occupancy envelope of the affected vehicles, shortens the planning range to the locally verifiable area, and changes two lane-changing maneuvers originally planned to be executed simultaneously to be executed sequentially. At this time, the second area has not yet obtained sufficient status coverage, so the system controls emergency vehicles to reduce their target speed to prevent them from entering channel units that have not formed continuous exits.

[0100] When vehicle-to-vehicle direct communication is still available, the system switches to a local communication cooperative mode, issuing cooperative control commands only to directly controllable vehicles within a one-hop range. When external links are completely unavailable, the vehicle-side perception local mode only controls the vehicle itself and authorized vehicles, treating all other traffic participants as uncertain dynamic constraint objects with spatiotemporal occupancy envelopes.

[0101] After communication is restored, the system performs a consistency check between the roadside status and the vehicle-side perception status. After five consecutive update cycles that meet the recovery threshold, the system gradually reduces the uncertainty expansion and, without changing the vehicle safety actions being performed, transfers the subsequent lane unit planning to the second area.

[0102] By using the hysteresis interval between the degradation threshold and the recovery threshold, the system avoids frequent mode switching when communication indicators fluctuate near the critical value, while maintaining the continuous connection of the spatiotemporal passage corridor.

[0103] The solution of the present invention can be replaced by the following: (1) The spatiotemporal occupancy envelope can be a rectangle, ellipse, polygon, occupancy grid, probability occupancy region, reachability set or other forms that can characterize the space that traffic participants may occupy; the uncertainty expansion can be obtained from the upper bound of the deterministic error, the probability confidence interval, the risk field or the learning prediction model, and the effective range of the envelope can be limited by the allowed activity area corresponding to the type of traffic participant, the reachability set of traffic participant dynamics and the adaptive prediction time domain.

[0104] (2) The spatiotemporal channel unit can be divided according to fixed length, lane segment, road topology node, conflict zone, or adaptive spatial range; the target release vehicle set can be solved using integer programming, graph search, set coverage, branch and bound, greedy heuristic, local search, or equivalent solution methods. During real-time operation, the number of candidate vehicles and candidate corridors can be limited, the feasible solution of the previous cycle can be reused, and the feasible solution with the minimum known cost and passing the safety constraint verification can be output when the computation time budget expires.

[0105] (3) In addition to the intersection of target envelopes, spatiotemporal conflict relationships can also be established based on responsibility-sensitive safety constraints, traffic rules, shared lane gaps, signal phases, execution order dependencies, and control permissions; execution windows can be represented by discrete time slots, continuous time intervals, or event triggering order.

[0106] (4) The information source, deployment entity, and communication method can vary individually or in combination. When some communication is unavailable, the system expands the spatiotemporal occupancy envelope and narrows the coordination range based on the remaining communication and local perception; when all external communication is unavailable, the system does not assume the ability to control other vehicles, but degenerates into local control of its own vehicle or authorized vehicles. The above changes do not alter the core technical concept of this invention, which constructs a corridor based on spatiotemporal occupancy envelope, spatiotemporal channel unit, target release vehicle set, and spatiotemporal conflict scheduling.

[0107] (5) When the complete corridor, the complete set of released vehicles, or the complete action scheduling sequence are temporarily unavailable, the target corridor can be divided into multiple stages, with each stage ending at a safe waiting unit that meets the conditions for emergency vehicles to stop or maintain low speed. The system first executes the subset of conflict-free actions that contribute the most to the release, and then continues to solve the next stage after the vehicle state is updated. This phased mechanism still uses road boundaries, dynamic reachability, safety of vulnerable traffic participants, and emergency vehicle braking distance as hard constraints.

Claims

1. A method for coordinated control of spatiotemporal passage corridors for emergency vehicles under mixed traffic conditions, characterized in that, include: Obtain status information of emergency vehicles and traffic participants within the target road area, and classify motor vehicles into directly controllable vehicles, communicable vehicles, and non-communicable vehicles based on their control authority and communication capabilities. The spatiotemporal occupancy envelope of each traffic participant in the prediction time domain is generated by combining state uncertainty. The spatiotemporal occupancy envelope is expanded by uncertainty from the predicted occupancy region and then restricted to the intersection of the dynamic reachability set and the allowed activity region of the corresponding traffic participant type. The generation of the spatiotemporal occupancy envelope specifically includes: determining the predicted occupancy region based on the predicted position, heading, and outline dimensions of the traffic participants at the prediction time; determining the uncertainty expansion amount based on data staleness, communication delay, positioning error, velocity error, acceleration error, traffic participant type, and prediction duration, and determining the dynamic reachability set based on the movement capabilities of the traffic participants; geometrically expanding the predicted occupancy region according to the uncertainty expansion amount, and restricting the expansion result to the intersection of the dynamic reachability set and the allowed activity region corresponding to the traffic participant type, thus obtaining the spatiotemporal occupancy envelope. When the uncertainty of the state exceeds a preset threshold, the effective prediction time domain is shortened and the subsequent spatiotemporal occupancy envelope is generated in a segmented rolling manner. Without pre-setting fixed priority lanes, the road space area is divided along the priority passage path according to the preset spatial length or road topology nodes, and the corresponding time window is set according to the estimated time for emergency vehicles to arrive at each road space area, forming a spatiotemporal channel unit. Based on the overlap relationship between the spatiotemporal occupancy envelope and the spatiotemporal channel unit, candidate spatiotemporal passage corridors are generated from different passable spatial zones; when there are no candidate spatiotemporal passage corridors that meet the conditions in the complete prediction time domain, segmented spatiotemporal passage corridors covering the current continuous confirmable area and ending at the safe waiting unit are generated. Based on the effective width, spatiotemporal continuity, and executable actions of the candidate spatiotemporal passageways, a target set of vehicles for releasing the passageway space is determined. Based on the spatiotemporal conflict relationship between the target spatiotemporal occupancy envelopes corresponding to the candidate actions of vehicles in the target release vehicle set, the execution order or execution window of the candidate actions is determined to form a collaborative control scheme. The cooperative control scheme is subjected to safety constraint verification, cooperative control commands are output to vehicles with control authority, yield prompts are output to vehicles with communication capabilities but no control authority, and the spatiotemporal occupancy envelope and spatiotemporal passage corridor are updated according to the updated traffic participant status. The safety constraint verification specifically includes: performing collision safety verification based on whether the spatiotemporal occupancy envelopes of any two traffic participants intersect or whether the minimum envelope distance is less than a preset safety threshold.

2. The method according to claim 1, characterized in that, The permitted activity areas for motor vehicles include motor vehicle lanes and authorized shoulder areas; the permitted activity areas for non-motor vehicles include non-motor vehicle lanes and mixed traffic areas; and the permitted activity areas for pedestrians include sidewalks and crosswalks.

3. The method according to claim 1, characterized in that, The generation of the spatiotemporal channel unit and the candidate spatiotemporal passage corridor further includes: the effective width of any spatiotemporal channel unit in the candidate spatiotemporal passage corridor is not less than the sum of the outer width of the emergency vehicle and the safety buffer distance on both sides; Adjacent spatiotemporal channel units have overlapping areas in space or meet the dynamic reachability conditions of emergency vehicles, and continuously cover the expected arrival range of emergency vehicles in time; The safety waiting unit is a spatiotemporal channel unit that accommodates emergency vehicles to stop or maintain low speed, meets safe braking distance and road boundary constraints, and maintains a preset safe distance from downstream unconfirmed conflict areas.

4. The method according to claim 1, characterized in that, The determination of the target set of released vehicles specifically includes: Candidate vehicles whose spatiotemporal occupancy envelopes overlap with candidate spatiotemporal passage corridors or their preset neighborhoods are selected, and the occupancy relationship between candidate vehicles and spatiotemporal passage units is established. The channel release contribution is calculated based on the target spatiotemporal occupancy envelope after the candidate vehicle completes its action. The channel release contribution is determined by at least one of the following: the number of newly added spatiotemporal channel units that meet the effective width condition, the increase in the effective width of the corridor, the increase in the continuous passable length, or the increase in the available time margin of the corridor. Under the conditions of satisfying the effective width, spatiotemporal continuity and safety constraints of the spatiotemporal channel unit, the first optimization objective is to minimize the number of actively participating vehicles, and the second optimization objective is to minimize traffic disturbance and action execution cost, and the target release vehicle set is determined. Within a preset computation time budget, the feasible solutions from the previous update cycle are reused in an incremental solution manner to generate an initial feasible set, and the initial feasible set is updated through heuristic search or local improvement. When the computation time budget expires, output the feasible set that has the minimum known cost and passes the security constraint check.

5. The method according to claim 1, characterized in that: The candidate actions include at least one of the following: maintaining driving, decelerating, stopping, lateral deviating within the lane, changing lanes to the left or right, pulling over to give way, and forward release; Based on whether the target spatiotemporal occupancy envelopes of candidate actions intersect, whether the candidate action trajectories cross, whether they compete for the same release space, and whether their execution time windows overlap, spatiotemporal conflict relationships are established between corresponding candidate actions. For candidate actions that have spatiotemporal conflicts, different start times, completion deadlines, or mutually exclusive execution windows are assigned, and execution priorities are determined according to the expected arrival order of emergency vehicles and the release order of spatiotemporal channel units. If there is no executable sequence covering all target actions within the preset calculation time budget, select a subset of conflict-free actions that are executable within the current time window and whose channel release contribution meets the preset conditions. First, release the preceding spatiotemporal channel units that emergency vehicles can continuously reach, and then redetermine the execution order or execution window for the remaining candidate actions.

6. The method according to claim 1, characterized in that: Send at least one of the following cooperative control commands to the directly controllable vehicle: target speed, target acceleration, target lane, lateral offset, target trajectory, execution window, and failure time; Send a yielding prompt message to the communicable prompting vehicle, consisting of at least one of the following: suggested yielding direction, suggested speed, suggested execution time, and risk warning; Instead of sending cooperative control commands to the non-communicating vehicles, the spatiotemporal occupancy envelope is generated based on their perception state, motion boundary, and prediction error. This envelope is used as an uncertain dynamic constraint object, and the candidate spatiotemporal passage corridors and other vehicles' candidate actions are adjusted accordingly.

7. The method according to claim 1, characterized in that, The safety constraint verification specifically includes: verifying whether the vehicle's target spatiotemporal occupancy envelope is within the drivable area and has not entered the restricted area; verifying whether the vehicle's longitudinal acceleration, lateral acceleration, yaw rate and target trajectory curvature are within the allowable range of vehicle dynamics; Safety verification of vulnerable road users is conducted based on the probability of trajectory conflict, expected collision time, or degree of risk exposure. Verify the effective width and spatiotemporal continuity of candidate spatiotemporal passageways when emergency vehicles are expected to arrive at each spatiotemporal passage unit.

8. The method according to claim 1, characterized in that, The method further includes: Based on regional status coverage, communication latency, packet loss rate, data obsolescence time, and status consistency, the system switches between full regional collaborative mode, local communication collaborative mode, and vehicle-side perception local mode. Local incremental updates are performed using a first update frequency, and global re-solution is performed using a second update frequency lower than the first update frequency, in order to update the spatiotemporal occupancy envelope, related occupancy relationships, and spatiotemporal conflict relationships of traffic participants whose states have changed. When the actual state of traffic participants exceeds the corresponding spatiotemporal occupancy envelope, the effective width of the spatiotemporal channel unit is lower than the preset threshold, the communication capability or control authorization status of vehicles in the target release vehicle set changes, a new spatiotemporal conflict relationship occurs, or the expected arrival time deviation of emergency vehicles exceeds the preset threshold, the spatiotemporal passage corridor is updated. The mode switching adopts a hysteresis control mechanism, and the degradation threshold and recovery threshold are set with different values ​​to avoid frequent switching of control mode caused by communication link fluctuations.

9. The method according to claim 1, characterized in that, The method also includes traffic restoration steps: After the emergency vehicle has passed, the order or window for restoring vehicle speed and lane status is determined based on the current position of the participating vehicles, the available space in the target lane, and the spatiotemporal conflict between the restoration actions. Output collaborative control commands to vehicles with control authority, and output recovery prompt messages to vehicles with communication capabilities but no control authority; When a recovery action causes an overlap of the spatiotemporal occupancy envelope or a new traffic bottleneck, the recovery time of the corresponding vehicle is delayed or the current safe state is maintained.

10. A spatiotemporal collaborative control system for emergency vehicle passage corridors under mixed traffic conditions, characterized in that, The system for implementing the method according to any one of claims 1 to 9 comprises: The status and occupancy envelope processing module is used to obtain status information of emergency vehicles and traffic participants, and classify motor vehicles according to their control permissions and communication capabilities. The spatiotemporal occupancy envelope of each traffic participant in the prediction time domain is generated by combining state uncertainty. The spatiotemporal occupancy envelope is constrained to the intersection of the dynamic reachability set and the allowed activity area of ​​the corresponding traffic participant type after uncertainty expansion of the predicted occupancy region. When the uncertainty of the state exceeds a preset threshold, the effective prediction time domain is shortened and the subsequent spatiotemporal occupancy envelope is generated in a segmented rolling manner. The corridor and release vehicle set determination module is used to divide the road space area along the priority passage path of emergency vehicles without pre-setting fixed priority lanes and set corresponding expected arrival time windows to form a spatiotemporal channel unit. Candidate spatiotemporal passage corridors are generated from different passable space zones based on the spatiotemporal occupancy envelope, and a target set of release vehicles for releasing corridor space is determined. When there are no candidate corridors that meet the conditions within the complete prediction time domain, a segmented spatiotemporal passage corridor is generated, covering the current continuous verifiable area and ending at the safe waiting unit. The spatiotemporal conflict scheduling and safety verification module is used to determine the execution order or execution window based on the spatiotemporal conflict relationship between the target spatiotemporal occupancy envelopes of the candidate vehicle actions in the target release vehicle set, form a collaborative control scheme and perform safety constraint verification. The control output and corridor maintenance module is used to output collaborative control commands to vehicles with control authority, output yield prompts to vehicles with communication capabilities but no control authority, and continuously update the spatiotemporal occupancy envelope and spatiotemporal passage corridor based on the updated traffic participant status.

Images