Method and system for determining an emergency path

Abstract

This disclosure relates to a method and system for determining an emergency path for an autonomous driving function of a vehicle. The method includes several steps. In a first step, map information relating to infrastructure features of drivable lanes in a road segment to be driven is acquired. Furthermore, attributes of the drivable lanes are determined based on the acquired map information. Further, an emergency path is calculated based on the determined attributes of the drivable lanes.

Description

Technical Field

[0001] This disclosure relates to a method and system for determining an emergency path for an autonomous driving function of a vehicle, and also provides a vehicle having a system according to this disclosure. Additionally or alternatively, a computer program is provided, comprising instructions that, when executed by a computer, cause the computer to perform at least partially the method. Background Technology

[0002] SAE Level 3 autonomous driving systems control the vehicle autonomously within their defined operating conditions. When the driving system reaches the boundaries of the operating conditions or encounters a critical situation, a Transition Demand (TD) is triggered, requiring the driver to take over control of the vehicle. If the driver does not respond, a Minimum Risk Maneuver (MRM) is initiated, the purpose of which is to bring the vehicle to a safe stop.

[0003] According to regulatory requirements, such as UNECE R157, MRM can be executed in different lanes or within a defined stopping area, such as on emergency lanes / shoulders, side lanes, or slowest traffic flow lanes. International standards, such as ISO 23739:2024, stipulate that MRM must also be activated in the event of a serious error / malfunction in the automated driving system or vehicle.

[0004] Existing driving systems used for MRM (Moving Responsibility Management) typically employ rule-based decision logic or static decision trees, which determine whether a vehicle should stop in its current lane or change lanes to reach a safer parking position. However, these decision logics are rigid and do not take into account all dynamic influencing factors, such as current traffic conditions, the likelihood of performing a safe lane change, road infrastructure (including the availability and quality of emergency lanes or parking areas), and / or the behavior of other traffic participants—who may affect the safety of the parking process.

[0005] Another issue is that existing systems only calculate the emergency path after the TD (Time Toll) has ended, wasting valuable time and potentially failing to optimally select a potential parking location. Furthermore, the severity of system errors is not adequately considered, so vehicles do not always achieve the best balance between minimizing rapid risk and a safe parking location.

[0006] Current MRM (Match Making and Ranging) systems have several limitations, related to a lack of dynamic consideration of traffic conditions. Existing systems do not evaluate in real-time whether lane changes are meaningful or necessary, but rather make decisions based on a set of static rules. Furthermore, there may be situations where the selection of safe stopping locations is limited, so the system often only stops within the current lane or performs a predefined lane change, without comprehensively evaluating the best available stopping options. Additionally, early calculation of emergency routes is not provided. Current methods often calculate the optimal stopping point only after the TD (Traffic Response Time) has ended, rather than developing a strategy for optimal stopping locations in advance. Summary of the Invention

[0007] The objective of this disclosure is to provide a method and / or system, each suited to enrich the prior art.

[0008] This task is accomplished through the features of the independent claims. The dependent claims and dependent sub-claims each contain improvements of this disclosure.

[0009] According to the first aspect, this task is addressed by a method for determining an emergency path for the vehicle's autonomous driving function. This method includes several method steps.

[0010] According to this disclosure, an emergency path can refer to an optimal parking spot and / or one or more trajectories. Furthermore, an emergency path can comprise a series of trajectories and / or these trajectories can be planned along a route. Additionally, an emergency path can be understood as a dynamically calculated trajectory that guides a vehicle to a safe parking location along an optimized route within the framework of Minimum Risk Operation (MRM) or a similar emergency strategy. This emergency path is based on continuous analysis of static map information, infrastructure conditions, and optional real-time traffic data, and allows for adaptive vehicle control considering current environmental conditions. The calculation of the emergency path involves weighting various influencing factors to ensure that the parking process achieves the best possible balance between safety, availability, and feasibility. Static attributes such as the number and type of lanes, the availability of emergency lanes / shoulders or building separation facilities can be considered, as well as dynamic parameters such as current traffic density, the speed of adjacent vehicles, available time windows for lane changes, or temporary closures. Processing this information allows for the precise determination of the optimal driving route, enabling the vehicle to stop with minimal risk. To technically implement emergency routes, various algorithms can be used, including route optimization methods (which calculate the shortest, safest, and most energy-efficient routes possible) and predictive models (which predict future traffic patterns based on historical and current sensor data). Furthermore, fleet data can be leveraged to incorporate verified stop points of other vehicles into the calculation, increasing the probability of MRM success.

[0011] In the first step, map information is acquired, which includes infrastructure features of potential drivable lanes in the road segment to be traversed. To determine an emergency path, map information containing infrastructure features of the road segment and drivable lanes is first acquired. This information may include the number of lanes, lane widths, the location of emergency lanes, lane markings, and other relevant features crucial for safe vehicle navigation. This step lays the foundation for subsequent safety assessments by utilizing pre-available static information. This accelerates the calculation of emergency paths because the system does not have to rely entirely on real-time sensor data. According to this disclosure, a drivable lane can be understood as a lane on a road segment defined by infrastructure and traffic-related conditions, suitable for regular or temporary vehicle use. Lane drivability can be influenced by both static features (e.g., road geometry and markings) and dynamic factors (e.g., traffic conditions or temporary closures).

[0012] In another step, the attributes of drivable lanes are determined based on the acquired map information. Specific attributes are assigned to the lanes based on the acquired map information. These attributes represent characteristics such as lane type, permissible speed, gradient, curvature, and potentially risk-based factors, which may be derived from past experience and / or fleet data. By assigning attributes, not only is a purely infrastructure description of the lane provided, but its suitability for potential minimum risk operations (MRM) is further categorized. This improves the decision-making process because a structural assessment of the environment is conducted before acute emergencies occur.

[0013] Lane classification is used to assess their suitability for minimum risk operation (MRM) and allows for differentiated assessments of the safety of road segments as potential parking locations. This classification is achieved by assigning attributes that can take into account both static and dynamic factors to enable accurate assessment of road infrastructure and current traffic conditions. Lanes can be classified according to numerical grades / scales or qualitative ratings, where lower values ​​or better ratings indicate safe parking areas, while higher values ​​or poorer ratings may indicate higher hazard and poorer suitability. One possible classification range could start from primary parking areas (such as emergency lanes or emergency parking zones), progress to secondary parking areas (such as right-hand lanes with moderate traffic volume), and then to areas to be avoided (such as left-hand overtaking lanes and / or lanes with no possibility of safe avoidance).

[0014] Classification can be based on a combination of predefined static attributes and dynamic real-time data. Static attributes are based on high-resolution map information that records infrastructure features such as the number of lanes, the availability and width of emergency lanes, the presence of emergency parking zones or other parking areas, and the visibility of parking areas to other road users. In addition, national traffic regulations (such as right-hand drive rules and / or special use regulations for emergency lanes) can also be incorporated into the assessment. Dynamic attributes can supplement this classification by acquiring current traffic data, including traffic density, relative speeds between adjacent lanes, available time windows for lane changes, temporary obstacles (such as construction or accidents), and weather conditions that may affect road conditions.

[0015] In addition to a combination of static and dynamic factors, fleet data can be integrated into the continuous adaptation of classification. Historical data from other vehicles can be used to assess the suitability of specific areas as parking locations with a certain probability of success. If a particular road segment has proven to be a safe parking area in the past, it can be prioritized for a higher category, while areas that repeatedly cause problems can be downgraded. Therefore, fleet data can be used to achieve a learnable classification adjustment that is further optimized with each execution of MRM.

[0016] By classifying lanes early, the system can identify and assess potential parking areas before emergencies occur, rather than waiting until an acute MRM situation arises to determine the appropriate location. This enables faster and safer decision-making because the vehicle already has a list of potential parking areas that can be compared with real-time traffic data. This allows for continuous reassessment of emergency routes and context-based adaptive adjustments.

[0017] Furthermore, it stipulates that emergency paths are determined based on the attributes of the identified accessible lanes and the acquired traffic data. Based on the adapted lane attributes and current traffic data, an emergency path is determined that describes the optimal movement the vehicle will take to minimize risk during MRM (Most Responsive Management). This emergency path considers both the best available parking location and the possibility of reaching a safe lane. This ensures that the vehicle does not simply make static decisions, but adapts to the optimal parking point in real time. This is significantly different from deterministic systems that rely on fixed decision logic and cannot flexibly respond to changing conditions.

[0018] This method offers several advantages. Firstly, it can accurately and efficiently determine emergency paths for autonomous driving functions by acquiring road segment map information and infrastructure features, and deriving the attributes of each lane. Secondly, by acquiring additional real-time traffic data, it ensures that lane attributes are not only based on static information but also adapted to specific traffic conditions. This allows for situational assessment of lane suitability for parking, enabling the targeted selection of an optimal parking location. Compared to traditional methods based on predefined decision trees or fixed rules, this method allows for flexible and dynamic determination of emergency paths, taking current conditions into account. Consequently, vehicles can pre-plan and adapt strategies for minimum-risk operations, thereby improving the safety of occupants and other road users.

[0019] Furthermore, this method allows for prior calculations through map-based evaluation and pre-classification of road segments, assessing potential safe parking corridors before actual driving decisions are made. These corridors serve as the basis for emergency path formation, eliminating the need to calculate all possible options in real-time and instead drawing upon a predefined, optimized data foundation. Combining this static prior information with real-time acquired environmental and traffic data enables rational decision-making that considers both the theoretical suitability of parking spots and current traffic conditions. This ensures that the calculated parking areas are actually usable under appropriate circumstances and can be dynamically adapted as needed.

[0020] Compared to deterministic, rule-based methods based on fixed decision trees, Unlike deterministic, rule-based methods that rely on fixed decision trees and define predetermined responses for each combination of errors, the method disclosed herein allows for adaptive and flexible responses to unforeseen events. While existing systems, for example, may exhibit critical limitations when evaluating situations based on an error matrix, such as when eliminating a particular action option based on a single criterion, the method disclosed here provides a holistic trade-off of all available action options. For example, an obstacle, such as a pothole, is not merely considered a static constraint making a certain driving path impossible, but can be dynamically evaluated based on traffic density, location, size, and possible alternative routes within emergency paths. Thus, the system maintains situational flexibility, enabling it to identify the optimal route with the lowest risk and preventing the neglect of reasonable options due to rigid decision-making logic.

[0021] Furthermore, the method disclosed herein reduces the computational workload required because by combining static prior data with dynamic real-time information, it is unnecessary to tediously calculate all potential parking areas in real time. Instead, potential corridors are classified before the actual calculation of emergency routes, so only the most relevant options are included in the real-time analysis. This leads to faster and more efficient decision-making, especially in critical situations where delayed responses can increase the risk of accidents.

[0022] The method may be a computer-implemented method, meaning that one, more, or all steps of the method may be performed at least in part by a computer or data processing device, optionally by a control device, particularly by the processing unit or control device of the system of this disclosure. It may be specified that, where meaningful, these steps may be performed in a different order than described.

[0023] The following sections will describe in detail the possible extensions to the various aspects of the above disclosure. Each extension listed below can be used, individually or in combination, to further develop and specify the methods described above.

[0024] It can be stipulated that, in another step, traffic data for the road segment to be traversed is acquired. In addition to static map information, traffic data is acquired in real time. This traffic data includes current vehicle position, traffic density, relative speed within lanes, potential lane change time windows, and detection of obstacles or other relevant road condition changes. Acquiring this data allows for supplementing static evaluations with dynamic components, enabling the system to make situation-adaptive decisions. Unlike existing systems that typically rely solely on static decision trees or predetermined actions, this step allows emergency routes to be personalized to the current traffic conditions.

[0025] Furthermore, it can be specified that the attributes of drivable lanes are adapted based on the acquired traffic data. Specific lane attributes can be updated based on current traffic data. For example, if a lane considered a potential parking area based on static information has high traffic density or temporary construction, it can be identified as unsuitable in real time. This step ensures that parking location assessment is based not only on predefined parameters but also on continuous analysis of traffic conditions. This avoids situations where MRM-guided parking appears safe but is unsuitable due to temporary changes.

[0026] It can be stipulated that an escape vector for an emergency path is determined based on the attributes of the determined drivable lanes and the acquired traffic data.

[0027] The method can be specified to include real-time, continuous adaptation of emergency routes during Transitional Demand (TD) and / or Minimum Risk Operation (MRM). This capability ensures that emergency routes are not determined only once, but are adapted to changing conditions throughout the process. This guarantees that vehicles can react to new traffic conditions, particularly when potential parking locations change or dynamic obstacles appear.

[0028] The method may also include providing control commands to control the vehicle based on the calculated emergency path. Incorporating this feature allows the calculated optimal actions to be directly passed to and implemented by the vehicle system, ensuring more precise and safer control of the vehicle within the MRM. Within the scope of this disclosure, control commands can be understood as electronic commands generated by the vehicle's control system and passed to the appropriate vehicle system to selectively influence vehicle motion or other driving dynamics. These commands can be calculated based on sensor data, map information, predefined rule algorithms, and / or real-time traffic data to achieve precise and safe vehicle control. In the context of autonomous vehicles, control commands typically include commands for longitudinal and lateral control of the vehicle. Longitudinal control involves controlling speed by adjusting powertrain power or braking force, while lateral control involves influencing the steering angle to guide the vehicle along a specific trajectory. The calculation of control commands is performed by a central processing unit that continuously evaluates environmental conditions and internal vehicle systems to make optimal operational decisions.

[0029] It can be specified that the attributes of adjacent lanes are adapted based on the severity of the detected error, which is assigned to a predefined category from multiple error categories. Adaptation of lane attributes can involve weighting existing risk assessments and influencing the selection or prioritization of specific lanes within an emergency path.

[0030] This has the following technical effects: lane selection and evaluation are not static or solely dependent on traffic conditions, but also take into account the specific technical system state of the vehicle. Therefore, the system's response to error states of varying severity can become significantly more robust, as risk mitigation measures can be taken in route planning based on the type and severity of the identified error. Incorporating error severity into attribute assignment allows vehicles to enter a defined safe state in advance and selectively exclude lanes unsuitable for specific error categories, or prioritize suitable lanes.

[0031] Within the scope of this disclosure, error severity can be understood as a technical metric that describes, quantitatively or qualitatively, the degree of functional impairment of an automated vehicle. For example, it might be caused by the loss of function of a single sensor, limited braking, or errors in the steering system or communication. Error severity can be assessed sequentially or hierarchically, for example, based on predefined system diagnostic parameters.

[0032] A predefined category can be understood as a structured entry within a classification system, which is assigned a range of error severity values ​​or error types. Each category represents a risk category associated with identifying an error situation and can be associated with a specific response in the emergency path algorithm.

[0033] Multiple error categories can represent a hierarchical or logically organized classification system that distinguishes different manifestations of technical errors. These error categories can be divided into levels such as "minor," "moderate," and "critical," or functionally defined categories such as "sensor error," "actuator limitation," or "communication interruption." Each of these categories can be associated with specific attribute assignments and behavioral patterns in emergency path calculations to achieve targeted, state-related risk minimization.

[0034] It can be specified that the remaining travel time before reaching the Emergency Response Mechanism (MRM) be considered in the real-time continuous adaptation of the emergency route. Incorporating this feature allows for proactive route planning, thereby optimizing the use of remaining time to allow the vehicle to decelerate and / or change lanes and / or stop as safely as possible. This enables the vehicle to react in advance to the necessity of the MRM and achieve more precise risk minimization control.

[0035] It can be stipulated that if traffic data identifies a lane in the drivable lanes of a road segment as having a lower risk value than the currently drivable lane, the control instruction includes a lane change. This proactively guides the driver to a safer lane, rather than statically remaining in the current lane, allowing for better adaptation to traffic conditions. The risk value can be understood as a numerical assessment of the degree of danger a lane or driving strategy poses to the safety of an automated vehicle. It is determined by combining static infrastructure data with dynamic traffic data analyzed in real time. Static factors include, for example, lane characteristics such as the presence of an emergency lane, road curvature, lane width, and visibility, while dynamic factors include current traffic density, the speed of nearby vehicles, the available time window for lane changes, and temporary obstacles such as construction sites. Risk values ​​can be represented in a tiered scoring system, such as from 0 to 10 or from 0 to 100, where low values ​​represent low risk and high values ​​represent high risk. For calculation, weighted algorithms are used, which consider various parameters differently depending on the current traffic conditions. This weighting can be dynamically adapted, for example, when changes in traffic conditions or system errors affect sensor data. Furthermore, historical fleet data and machine learning can be utilized to pre-assign high-risk values ​​to lanes known to be hazardous. The determined risk values ​​are continuously updated and serve as the basis for driving decisions, including lane changes, braking strategies, and parking location selection. By using adaptive risk values, the vehicle can flexibly respond to changing conditions and achieve a safe and optimized driving strategy by always selecting the best available route with the lowest risk potential.

[0036] The control commands can be further specified to include a braking strategy used to decelerate the vehicle within areas assessed / graded as safe within the calculated emergency path. This feature helps ensure that the vehicle does not stop abruptly or in inappropriate locations, but rather selectively chooses an optimal parking location within available infrastructure. The assessed areas within the calculated emergency path refer to specific areas along the calculated trajectory where the vehicle has been evaluated for safety and suitability for minimum risk operation (MRM) based on static and dynamic factors. This assessment / grading is achieved through a weighted analysis of infrastructure data, traffic conditions, and historical fleet data, assigning a risk value to each part of the emergency path to determine the optimal parking location. These assessed areas, as priority parking zones, enable vehicles to ensure safe, compliant, and / or efficient and / or controlled parking.

[0037] It can be stipulated that lane attributes are determined based on a risk assessment that combines infrastructure characteristics and traffic data. This allows for an objective assessment of the suitability of lanes as parking locations, ensuring that both static and dynamic factors are considered in the decision-making process.

[0038] It can be stipulated that adjusting lane attributes includes assessing safety risks based on calculated risk values ​​to determine priority parking locations within emergency routes. This function can more accurately prioritize potential parking areas by considering not only individual attributes but also the overall risk assessment of each segment.

[0039] It can be specified that the acquired traffic data includes the current location of vehicles. This allows for precise adaptation of emergency routes to location conditions.

[0040] It can also be specified that the acquired traffic data includes traffic density for each lane. This feature helps the system identify lower-risk areas for safer MRM execution. This improves operational safety.

[0041] It can also be specified that the acquired traffic data includes a time window that can be used for lane changing. This allows the system to determine whether a lane change is safe to make or whether alternative parking locations should be prioritized.

[0042] It can also be specified that the acquired traffic data includes the speed and / or direction of movement of nearby vehicles. This enables more accurate hazard analysis to optimize MRM in consideration of current traffic dynamics.

[0043] It can be stipulated that, when determining an emergency path, a weighted average of static and dynamic attributes is applied based on the type and severity of the detected vehicle error. This allows for targeted prioritization of decision parameters optimal for specific error scenarios. Within the scope of this disclosure, an error can be understood as a technical or functional defect identified within the vehicle's autonomous driving functions that affects normal control or driving behavior and necessitates adjustments to the emergency path. Such errors can be categorized into different types, relating to both vehicle hardware and software-based control systems.

[0044] System errors can be categorized into various types that affect different components and functions of a vehicle. A major category might be sensor errors, where malfunctions or performance degradation can occur in cameras, lidar, or conventional radar systems, significantly limiting environmental perception. Similarly, delays or errors in data fusion between multiple sensor channels can cause the system to receive incorrect or inaccurate environmental information. Another issue in this category is errors or insufficient accuracy in GPS signals, which can lead to incorrect location determination, particularly problematic in navigation and emergency route calculation.

[0045] Another type of error is actuator error, which directly affects vehicle control. This includes malfunctions in the steering, braking, or drive systems, potentially leading to complete or partial limitation of vehicle control. Of particular critical is erroneous or delayed braking response, which can cause the vehicle to brake unexpectedly or fail to stop in time.

[0046] Communication errors are also safety-related because they can interfere with data transmission, both inside and outside the vehicle. Errors or interruptions in the CAN bus system can prevent control units from communicating properly, directly impacting safety-critical functions such as braking or steering commands. Similarly, V2X communication errors can cause the vehicle to stop receiving information about its current environment, making it unable to process essential information such as construction site warnings or traffic data.

[0047] To ensure the best possible response to detected system errors, static and dynamic attributes are adaptively weighted when determining the emergency path. This weighting ensures that different decision priorities are set according to the type of error, so as to bring the vehicle to a stop as safely as possible.

[0048] In the event of sensor malfunctions, the system can rely more heavily on static map information and fleet data to compensate for missing or inaccurate sensor data. Because environmental perception can be affected by malfunctioning or failed sensors, the vehicle will depend more on existing high-resolution map information and historical traffic data from other vehicles to reliably identify potential parking areas. In the event of actuator malfunctions, the system can prioritize the accessibility of safe parking locations by considering alternative braking or control strategies in advance. In this case, the system can, for example, choose a gentle braking strategy to achieve the lowest possible risk of deceleration, or selectively seek lanes with lower traffic density to ensure a safe stop. If communication errors occur, the vehicle can make redundant decisions, such as by placing greater emphasis on local sensor data and ignoring external traffic data. This ensures that even without information from external sources about current traffic conditions or infrastructure, the vehicle can still independently calculate contingency strategies.

[0049] The weighting between static and dynamic information is also related to error prioritization. A minor error, such as a temporary sensor malfunction, may cause the system to continue considering dynamic traffic conditions to select the optimal driving strategy. The vehicle is still able to process real-time data to flexibly respond to current traffic conditions. Conversely, a severe error, such as a complete failure of steering or braking, may cause the vehicle to prioritize only available static parking areas to minimize the risk of loss of control. In this case, any dynamic adaptations are revoked to support an immediate, safe parking position, allowing the vehicle to transition to a safe state as quickly as possible.

[0050] It can be stipulated that, in order to continuously adapt to emergency routes, traffic data from the vehicle environment, including camera data, conventional radar data, and / or lidar data, be used. This can provide high-precision environmental analysis, thereby providing improved risk assessment for MRM.

[0051] It can be specified that the map information includes the number of lanes. This allows for basic infrastructure classification to determine which options are available for lane changes and / or parking locations.

[0052] It can also be specified that map information includes lane types. This allows for the differentiation of different lane segments, such as main lanes, acceleration lanes, and emergency lanes, to formulate appropriate MRM enforcement strategies.

[0053] It can also be specified that map information includes the availability of emergency lanes. This ensures that the system has appropriate emergency zones so that vehicles can be safely stopped in a controlled location.

[0054] It can also be specified that the map information includes the location of the emergency lane. Including this feature allows for targeted checks during emergency path calculations to determine whether the emergency lane is within reach and suitable as a parking location.

[0055] It can also be specified that map information includes lane markings. This helps identify whether lane changes are permitted or whether there are continuous lines that might restrict maneuvering options.

[0056] It can also be stipulated that map information includes traffic signs. This allows the system to obtain additional information regarding speed limits and / or special traffic rules that may affect the MRM.

[0057] It can also be specified that map information includes traffic management systems. This feature helps the system understand existing traffic management systems that may announce temporary road changes and / or construction sites and / or other current traffic changes, thereby improving the accuracy of emergency route calculations.

[0058] It can be stipulated that map information is updated regularly using fleet data. This ensures that the system always uses the latest information, rather than relying solely on static map data.

[0059] It can be stipulated that lane attributes include at least a value related to the average speed of the corresponding lane. This allows for analysis of the relative dynamics of traffic participants and determines whether lane changes to mitigate risk are worthwhile.

[0060] It can also be specified that lane attributes be updated in real time, taking into account temporary changes to infrastructure—including construction zones, accidents, and / or weather effects. Incorporating this feature allows for responses to unexpected road conditions and the development of adaptive risk-minimizing strategies.

[0061] An emergency path can be defined as follows: the vehicle must reach a safe stopping position within a defined time window. This ensures that the vehicle is not unnecessarily exposed to higher risks for extended periods and that safe maneuvers can be completed within the specified time.

[0062] Furthermore, according to another publicly available aspect, this task is addressed through a system used to determine the emergency path for the autonomous driving function of a vehicle (motor vehicle). This system is used to determine the emergency path for the vehicle's autonomous driving function. To this end, the vehicle environment is acquired and the emergency path is dynamically adapted based on map and traffic data. This enables precise adaptive control of the vehicle in situations requiring minimum risk operation (MRM).

[0063] The system includes a first acquisition unit configured to acquire map information containing infrastructure features of drivable lanes in a road segment to be traversed. The first acquisition unit aims to acquire map information containing infrastructure features related to the road segment and lanes. The map information includes parameters such as the number and type of lanes, the presence of emergency lanes, lane markings, national traffic regulations, and specific road characteristics such as curvature or gradient. The first acquisition unit can be technically configured to directly retrieve stored map information from internal vehicle memory and / or acquire updated map data via an external cloud connection.

[0064] If the map information is stored in the vehicle's memory, the system can include a local storage unit containing high-resolution map information. This storage unit can be a solid-state drive (SSD), a flash memory card, and / or a dedicated storage partition within the system. The advantage of this solution is the low latency in accessing the map information, as no external data transfer is required. The stored maps can be updated regularly via over-the-air (OTA) downloads to account for new infrastructure changes, such as construction sites or changes in traffic management.

[0065] If map information is accessible via the cloud, the first acquisition unit can be equipped with a wireless communication interface that allows connection to a cloud-based map server. This can be achieved via LTE, 5G, or dedicated V2X communication. In this case, the system acquires updated map information as needed to always utilize the latest road infrastructure data. Cloud connectivity allows for continuous map updates, enabling dynamic changes such as temporary road closures or new traffic management measures to be incorporated into calculations in near real-time.

[0066] To improve redundancy, the first acquisition unit can combine these two methods: first, accessing locally stored map information, and then verifying or supplementing the map information via cloud queries. If a cloud connection is unavailable, the system can still rely on the last stored map information and acquire updated data once the network connection is restored. Furthermore, the first acquisition unit can be combined with sensor fusion technology to compare infrastructure features identified by cameras, LiDAR, or traditional radar systems with the stored map information. This allows for the verification or updating of inaccurate or outdated map data based on the vehicle's current environmental perception.

[0067] In addition, a processing unit is provided. This processing unit can be integrated into the vehicle itself or located on an external server. The processing unit can function as a control device for the vehicle. This control device can be part of or represent a vehicle driver assistance system. For example, the processing unit or control device could be an electronic control unit (ECU). The ECU can be an intelligent, processor-controlled unit, communicating with other modules via a central gateway (CGW), and may form a vehicle network via fieldbuses such as CAN bus, LIN bus, MOST bus, FlexRay, and / or automotive Ethernet, for example, together with telematics control units and / or actuators and / or environmental sensors. Appropriate interfaces can be formed for this purpose. It can be specified that the control device controls functions related to vehicle driving behavior, such as steering, powertrain control, power transmission, and / or braking systems. Furthermore, driver assistance systems, such as parking assist, adaptive cruise control (ACC), lane keeping assist, lane change assist, traffic sign recognition, light signal recognition, start assist, night vision assist, and / or intersection assist, or entertainment systems, can be controlled by this control device.

[0068] The processing unit is configured to determine the attributes of drivable lanes based on acquired map information. The processing unit analyzes the acquired map information and determines the attributes of each lane. These attributes include, for example, the permitted speed on the lane, visibility to other road users, or proximity to emergency lanes or emergency stopping lanes. Determining these attributes allows for a qualitative assessment of each lane, enabling the system to differentiate the suitability of a lane as a potential parking area. This early classification helps in MRM scenarios where suitable parking locations are not searched for in real-time, but rather pre-assessed options are available.

[0069] Furthermore, the processing unit is configured to calculate an emergency path based on the determined drivable lane attributes. The processing unit calculates the emergency path based on the adapted lane attributes and current traffic data. This path describes the optimal movement the vehicle will take to minimize risk during MRM (Morbidity Management). Weights are applied between static and dynamic factors to make the best possible decision regarding a safe stopping position. By calculating the emergency path, the vehicle can not only determine a suitable stopping position but also plan its movement to get there. For example, the system can analyze whether a lane change is necessary to reach a safer area or whether slowing down within the current lane is a better option.

[0070] In one embodiment of the system, a second acquisition unit may be provided, configured to acquire traffic data of the road segment to be traveled. The second acquisition unit is used to acquire traffic data of the road segment in real time. The traffic data includes parameters such as traffic density, average speed of each lane, time windows available for lane changes, and the movement direction of nearby vehicles. By continuously acquiring this data, the system can perform dynamic risk assessment of the environment. This enables the system to adaptively respond to sudden changes in traffic conditions—such as congestion, accidents, or construction sites. The combination of Die traffic data and static map information ensures that the system can flexibly respond to changing conditions by developing optimal strategies for emergency route adaptation. The second acquisition unit may include cameras, lidar, and / or traditional radar systems and / or other sensors for acquiring environmental data.

[0071] To this end, the processing unit is also configured to adapt the attributes of drivable lanes based on the acquired traffic data. The processing unit adapts previously determined lane attributes to the acquired traffic data. For example, a lane classified as safe based on map information may be identified as unsuitable due to temporary factors such as high traffic density or construction sites. This dynamic adaptation of lane attributes ensures that the system always uses the latest information and coordinates its decisions according to current traffic conditions. This prevents vehicles from choosing parking areas that are initially safe but unsuitable under current circumstances. Therefore, the processing unit is configured to determine emergency routes based on the specific attributes of drivable lanes and the acquired traffic data.

[0072] This disclosure also includes improvements to the previously described system, which have the same features as those described in the improvements incorporated into the methods of this disclosure. Therefore, the corresponding method extensions will not be repeated here, and the above description of the methods also applies to the system, and vice versa.

[0073] The above content can be summarized in other ways and is described below in more specific embodiments of the possible disclosure, but the following description should not be construed as a limitation on the disclosure.

[0074] It can be stipulated that map information (map data) is always stored and, where possible, its changes and timeliness are verified by the fleet. According to this disclosure, the stored map data is categorized. This categorization involves static risk assessments based on infrastructure conditions. Examples include (but are not limited to) the number and type of lanes, lane layout, lane markings, the presence, location, and number of emergency lanes, visibility to other road users, and national traffic regulations (e.g., right-hand drive regulations).

[0075] Static classification is based on factors such as the number of adjacent lanes and the average speed in each lane. Other factors, such as the curvature or slope of the road, may also play an important role.

[0076] In addition to static risk assessment, dynamic risk assessment is also conducted. Dynamic risk assessment determines factors such as the vehicle's current position on the road, current traffic conditions, the probability of lane changes (constrained by time windows and lane markings), and the likelihood of reaching a lane with a higher static risk assessment than the currently occupied lane under given dynamic conditions. Therefore, emergency path formation can be performed during the transition demand / takeover request (TD) phase. Furthermore, temporal and spatial changes in vehicle position can influence the direction of the emergency path. Additionally, error severity can be incorporated into the calculations, which can be achieved by pre-defining (through data) the expected range of the system under specific error or error severity conditions.

[0077] Every interaction between the driver and the vehicle, such as potential oversteer or takeover, can affect the direction of the path. In certain situations, emergency paths may align with the vehicle's direction of travel; in these cases, static risk assessments can be referenced to evaluate the best possible stopping position within the lane. Weighting of static and dynamic information can be based on the error and its severity. Furthermore, it is possible to specify the direct integration of static risk assessments into the driving function map. This allows for the direct assignment of risk scores (risk values) to stopping points for certain lane segments. It is also possible to specify the integration of fleet data into both static and dynamic risk assessments. This fleet data can also include success probabilities if other vehicles (in the presence of specific errors) have already calculated emergency paths and performed MRM. Such data can be incorporated as scores / weights into the calculation of static risk assessments.

[0078] This significantly improves the calculation of safe stopping points, thereby significantly reducing the residual risk to drivers, especially in situations where getting out of the vehicle may be necessary. It enables (rapid) integration of static data (with low computational cost) and dynamic data (potentially error-dependent), but does not require static data to calculate stopping points when MRM is triggered. Physical realities can be integrated into the calculation of stopping points.

[0079] The technical implementation disclosed herein can be used as part of an autonomous driving function or as a standalone function that interacts with the autonomous driving function.

[0080] Furthermore, a vehicle equipped with the aforementioned system is provided. This vehicle (motor vehicle) can be a passenger car, particularly an automobile, or a commercial vehicle, such as a truck. The motor vehicle can be a Level 3 automated driving – i.e., conditional automated driving – motor vehicle, where the driver does not need to continuously monitor the vehicle system. The motor vehicle automatically performs functions such as activating turn signals, changing lanes, and / or keeping lanes. The driver can engage in other activities, but will be required by the system to take over control during a warning period when necessary. The motor vehicle can be a Level 4 automated driving – i.e., highly automated driving – motor vehicle, where control of the motor vehicle is permanently taken over by the vehicle system. If the system can no longer complete the driving task, it can require the driver to take over control. The motor vehicle can be a Level 5 automated driving – i.e., fully automated driving – motor vehicle, where the driver is not necessary to complete the driving task. No human intervention is required except for setting a goal and activating the system. The vehicle may be without a steering wheel and pedals. The above description of the methods and systems also applies to this motor vehicle, and vice versa.

[0081] Furthermore, a computer program is provided that contains instructions, which, when executed by a computer, cause the computer to at least partially perform or implement the methods described above. The program code of the computer program can take any form, particularly code suitable for motor vehicle control devices. The above description of the system, vehicle, and method also applies to this computer program, and vice versa.

[0082] Furthermore, a computer-readable medium, particularly a computer-readable storage medium, is provided. This computer-readable medium contains instructions that, when executed by a computer, cause the computer to at least partially perform or implement the methods described above. That is, a computer-readable medium containing a computer program as defined above can be provided. This computer-readable medium can be any digital data storage device, such as a USB flash drive, hard disk, CD-ROM, SD card, or SSD card (i.e., an SSD drive / SSD hard disk). The computer program need not necessarily be stored on such a computer-readable storage medium for provision to a vehicle; it can also be obtained via the Internet or other external means. The above descriptions of the methods, systems, computer programs, and vehicles also apply to this computer-readable medium, and vice versa.

[0083] This disclosure also includes combinations of features of the described embodiments. Therefore, this disclosure also includes implementations that each have a combination of features of a plurality of described embodiments, provided that these features are not described as mutually exclusive. Attached Figure Description

[0084] The following describes embodiments of this disclosure. Therefore: Figure 1This illustration schematically depicts a motor vehicle in a road segment where it is to be driven, in one embodiment of the method disclosed herein. Figure 2 A motor vehicle equipped with the system disclosed herein is illustrated schematically; Figure 3 A flowchart illustrating one embodiment of the method of this disclosure is shown schematically. Detailed Implementation

[0085] The embodiments described below are preferred embodiments of this disclosure. In these embodiments, each described embodiment represents an independent and separately conceivable feature of this disclosure, which can also independently further develop this disclosure. Therefore, this disclosure should also include combinations of features other than those of the illustrated embodiments. Furthermore, the embodiments can be supplemented by other features already described in this disclosure.

[0086] The same reference numerals in the figure indicate elements with the same function.

[0087] Figure 1 The illustration schematically depicts a vehicle 1 in a road segment to be driven when performing one embodiment of the method 100 of this disclosure. Figure 1 This diagram illustrates vehicle 1 on a multi-lane road. The road comprises multiple lanes, which may have different risk values. Lane attributes are determined based on infrastructure information considering both static and dynamic factors.

[0088] Vehicle 1 is equipped with system 10, which includes a first acquisition unit 11 and a processing unit 13. The first acquisition unit 11 acquires map information relating to the infrastructure features of drivable lanes in a road segment. The information includes, in particular, the number of lanes, the type of lanes, the presence and location of emergency lanes (hard shoulders), as well as lane markings and traffic signs.

[0089] In another embodiment, system 10 may further include a second acquisition unit 12. Alternatively, only one acquisition unit may be provided, which is capable of performing the steps (especially method steps) of the first acquisition unit 11 and / or the second acquisition unit 12.

[0090] In another embodiment, the second acquisition unit 12 additionally acquires traffic data for the current road segment. The traffic data includes the current position of vehicle 1, traffic density of each lane, time windows available for lane changing, and speeds and directions of movement of nearby vehicles.

[0091] Processing unit 13 determines lane attributes based on the acquired map information. Based on these attributes, an emergency path 20 is calculated, which specifies the optimal movement of vehicle 1 for performing the minimum risk maneuver (MRM). Furthermore, processing unit 13 also determines lane attributes based on the acquired traffic data. These attributes are adapted to the current traffic data for a context-appropriate risk assessment. Emergency path 20 can then be calculated.

[0092] exist Figure 1 In the example, vehicle 1 is located in a lane with a medium risk level. Emergency path 20 shows the movement of vehicle 1 to a safer lane with a lower risk level. Emergency path 20 is continuously adapted in real time based on the remaining travel time to reach a safe stopping point and current traffic conditions.

[0093] If traffic data identifies a lane with a low risk value, a lane change can be made. Alternatively, a braking strategy can be applied to slow vehicle 1 down within the area of ​​the calculated emergency path 20 that is deemed safe.

[0094] System 10 is designed to utilize traffic data from the vehicle 1's environment, including camera data, conventional radar data, and lidar data. Map information can be updated periodically using fleet data to achieve the most accurate and real-time calculation of emergency routes 20 possible.

[0095] also, Figure 1 The image also shows a vehicle traveling ahead. This vehicle may be in the same lane or an adjacent lane. The second acquisition unit 12 identifies the vehicle traveling ahead and determines its motion. Based on this information, the system 10 can perform a hazard analysis to decide whether to change lanes or apply a braking strategy.

[0096] Processing unit 13 determines lane attributes and adapts them in real time. If the vehicle ahead is traveling slowly or brakes unexpectedly, system 10 can decide whether vehicle 1 should accelerate, decelerate, or change lanes. The calculated emergency path 20 is considered in the attribute adaptation. The attributes of adjacent lanes can be adapted based on the severity of detected errors. Predefined error severity categories can be assigned from multiple error categories.

[0097] The calculation of emergency path 20 ensures that vehicle 1 moves safely. If the current lane is blocked or unsafe, it will consider changing to an adjacent lane with a lower risk value. Alternatively, vehicle 1 can be decelerated to reach a safe stopping position.

[0098] Therefore, the vehicles traveling ahead affect the continuous adaptation of emergency path 20. If vehicle 1 is in a minimum risk operation (MRM), it checks whether stopping in the current lane or changing lanes is safer. This system can utilize camera data, conventional radar data, and / or lidar data for more precise environmental analysis.

[0099] By evaluating traffic data in real time, System 10 can react more quickly in critical situations and minimize the risk of rear-end collisions.

[0100] Figure 2 A schematic diagram of a system 10 according to one embodiment is shown, which is integrated into a vehicle 1 for determining an emergency path 20 for an autonomous driving function. System 10 includes a first acquisition unit 11 and a processing unit 13. In another embodiment, a second acquisition unit 12 may also be included.

[0101] The first acquisition unit 11 is configured to acquire map information, which includes the infrastructure characteristics of the drivable lanes of the road segment. The map information includes the number and type of lanes, the presence and location of emergency lanes, lane markings, and traffic signs. Using this static data, the system can perform a preliminary assessment of the road environment.

[0102] The second acquisition unit 12 is configured to acquire traffic data for the road segment to be traveled. The traffic data includes the position of vehicle 1, traffic density of each lane, time windows available for lane changing, and the speed and direction of movement of nearby vehicles. The acquired traffic data allows for dynamic adaptation to previously determined lane attributes.

[0103] The processing unit 13 is configured to analyze the map information acquired by the first acquisition unit 11 and determine lane attributes accordingly. These attributes may include, for example, the average speed of the lane or the accessibility of a safe stopping position.

[0104] In another embodiment, the processing unit 13 adjusts the attributes based on the traffic data acquired by the second acquisition unit 12 to perform a risk assessment that conforms to the context.

[0105] Based on the aforementioned attributes, processing unit 13 calculates emergency path 20. Emergency path 20 defines the optimal movement and / or optimal stopping point for vehicle 1 to minimize risk during minimum risk operation (MRM). During this process, the system continuously analyzes the environment, allowing vehicle 1 to choose between lane changes or reaching a safe stopping point within a defined area.

[0106] System 10 is designed to fully execute the reference. Figure 3 Method 100 for determining emergency path 20. Vehicle 1 is equipped with system 10 that can automatically execute method 100 to perform optimal MRM.

[0107] Figure 3 A schematic diagram of a method 100 for determining an emergency path 20 for an autonomous driving function of vehicle 1 is shown. Method 100 includes several method steps.

[0108] Method 100 begins with step 110, in which map information of the drivable lanes in the road segment to be traversed is acquired, the map information including infrastructure features of the drivable lanes. This information includes the number and type of lanes, the presence and location of emergency lanes and / or supplementary lanes, lane markings, and traffic signs. By acquiring this static data, a preliminary environmental analysis can be performed.

[0109] In step 120, lane attributes are determined based on the acquired map information. These attributes may include, for example, the lane's average speed or its suitability as a potential parking location. During this process, the lane's risk potential is assessed to provide a basis for subsequent decisions.

[0110] In step 130, an emergency path 20 is calculated based on the determined lane attributes. Emergency path 20 specifies the optimal movement of vehicle 1 to perform the least-risk maneuver (MRM), specifically to reach the optimal stopping point. During this process, vehicle 1 may choose to change lanes or apply braking strategies to achieve a safe stopping position.

[0111] Method 100 is continuously executed so that emergency path 20 can be adapted in real time. This enables adaptive control of vehicle 1, taking into account both static and dynamic factors.

[0112] In addition, the method may include a step of acquiring traffic data for the road segment to be traveled. This data includes the current position of vehicle 1, traffic density of each lane, time windows available for lane changing, and the speed and direction of movement of nearby vehicles. By acquiring this dynamic information, method 100 can take into account the current traffic conditions.

[0113] In another supplementary step, the previously determined lane attributes can be adapted to current traffic data. If traffic conditions change—for example, due to high traffic density or temporary obstacles such as construction sites—the lane attributes will be modified accordingly. This ensures that the lane risk assessment is always up-to-date.

[0114] The calculation of emergency path 20 can be supplemented by the attributes of the determined drivable lanes and the acquired traffic data.

[0115] Overall, these examples demonstrate how to provide a method for generating emergency paths for autonomous driving functions.

Claims

1. A method (100) for determining an emergency path (20) for an autonomous driving function of a vehicle (1), the method comprising the steps of: - Obtain (110) map information, which involves the infrastructure features of potential drivable lanes in the road segment to be driven; - Determine the attributes of drivable lanes based on the acquired map information; - Calculate the emergency path based on the properties of the determined drivable lanes (20).

2. The method (100) according to claim 1, characterized in that, The method also includes: - Obtain traffic data for the road segment to be traveled; - Adapt the attributes of drivable lanes based on the acquired traffic data, and then calculate emergency routes based on the acquired traffic data; - During transitional demand (TD) and / or minimum risk operation (MRM), the emergency path (20) is continuously adapted in real time; and / or - Provide control instructions to control the vehicle (1) according to the calculated emergency path (20). Among these measures, the attributes of adjacent lanes are adapted based on the severity of the detected errors. The severity of the error is assigned a predefined category from multiple error categories.

3. The method (100) according to claim 2, characterized in that, The real-time continuous adaptation of the emergency path (20) takes into account the remaining travel time before reaching the minimum risk operation.

4. The method according to claim 2 or 3, characterized in that, If traffic data identifies a lane with a lower risk value than the current lane in the drivable lanes of the road segment to be driven, the control command includes lane change, wherein the control command includes a braking strategy to decelerate the vehicle in a safe area within the calculated emergency path (20).

5. The method (100) according to claim 4, characterized in that, Lane attributes are determined based on a risk assessment derived from a combination of infrastructure characteristics and traffic data.

6. The method (100) according to claim 5, characterized in that, Adapting lane attributes involves assessing safety risks based on risk values ​​calculated through risk assessment to determine preferred parking locations within the emergency path (20).

7. The method according to any one of claims 2 to 6, characterized in that, The traffic data obtained includes at least: - The current position of vehicle (1); -Traffic density on each lane; - The time window available for lane changes; and / or - The speed and / or direction of movement of nearby vehicles.

8. The method according to any one of claims 2 to 7, characterized in that, In the calculation of the emergency path (20), the static and dynamic attributes are weighted according to the type and severity of the detected vehicle (1) error.

9. The method according to any one of claims 2 to 8, characterized in that, In order to continuously adapt the emergency path (20), traffic data from the vehicle (1) environment is used, including camera data, conventional radar data and / or lidar data.

10. The method according to any one of the preceding claims, wherein, Map information should include at least: - The number of lanes; - Lane type; -Availability of emergency lanes; - Location of the emergency lane; - Lane markings; - Traffic signs; and / or Traffic management system.

11. The method according to any one of the preceding claims, characterized in that, Lane attributes include at least the average speed value in the corresponding lane, and lane attributes are updated in real time to account for temporary changes in infrastructure.

12. The method according to any one of the preceding claims, characterized in that, The emergency path (20) is determined to enable the vehicle (1) to reach a safe stopping position within a defined time window.

13. A system (10) for determining an emergency path (20) for an automated driving function of a vehicle (1), the system comprising: - First acquisition unit (11), the first acquisition unit is configured to acquire map information, the map information having infrastructure features of drivable lanes in the road segment to be driven; as well as - Processing unit (13), which is configured as follows: The attributes of drivable lanes are determined based on the acquired map information. Calculate the emergency path based on the properties of the determined drivable lanes (20).

14. A vehicle (1) equipped with a system (10) according to claim 13, the system being used to determine an emergency path (20).

15. A computer program, wherein, The computer program contains instructions that, when executed by a computer, cause the computer to perform the method according to any one of claims 1-12.

Images