Autonomous driving safety system and autonomous driving control method that share a driving design area based on risk level

Abstract

To provide an autonomous driving safety system for sharing a risk-based operation design domain capable of securing autonomous driving safety by complementing operation design domains each unique to each autonomous driving vehicle.SOLUTION: The present disclosure includes an autonomous driving safety system for sharing a risk-based operation design domain including an autonomous driving system 110 configured to control a vehicle driving unit 140 according to information detected by a sensor unit 120 to perform autonomous driving, in which the autonomous driving system 110 includes an operation design domain update unit 111 configured to evaluate a risk of at least one of a static operation design domain and a dynamic operation design domain recognized by the sensor unit 120 while driving to update the operation design domain.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an autonomous driving safety system and an autonomous driving control method for sharing a risk-based operation design area, which can design a risk-based operation design area and share the designed risk-based operation design area via an edge infrastructure. [Background technology]

[0002] According to the Ministry of Land, Infrastructure and Transport of Korea's "Level 4 Autonomous Vehicle Manufacturing Safety Guidelines," the Operation Design Domain (ODD) is defined as "the operating area (roads, weather, traffic, etc.) in which the autonomous driving system's functions can be performed normally and safely," and it is recommended that the Operation Design Domain related to the safe operation of the autonomous driving system be clearly indicated.

[0003] The safety guidelines clearly indicate the operational design area and provide system status information when the operational design area is exceeded, including road type, geographical area, weather environment, speed range, and other constraints.

[0004] Additionally, the US NHTSA's "Automated Driving Systems: A Vision for Safety 2.0" recommends that each ADS (Automated Driving System) define and document an ODD for testing or deployment for use on public roads.

[0005] Additionally, the Operation Design Domain (ODD) must describe the specific conditions under which a given ADS or function is intended to operate, and the ADS must be able to operate safely within the designed ODD.

[0006] However, in the Operation Design Domain, ODD (Operation Design Domain) design is recommended to ensure actual road-based autonomous driving operation and functional safety, but complementary settings such as hierarchical classification and detailed attribute value definitions for the Operation Design Domain are required.

[0007] Autonomous driving on actual roads can be divided into static ODD and dynamic ODD. For example, road width, lanes, signs, and structures can be considered static ODD characteristics, while objects (people, vehicles, etc.) and weather conditions can be considered dynamic ODD characteristics.

[0008] The dynamic and static characteristics of such operational design domains require updating and supplementing the initially designed ODD design in accordance with the flow of time (changes in ODD characteristics between day and night, deformation of structures over time, etc.) and changes in traffic (traffic volume, unexpected objects, etc.).

[0009] However, in the past, there has been a lack of updates to these Operation Design Domains or follow-up after updates, which has led to a decrease in safety.

[0010] In addition, since the design conditions and response strategies of each vehicle are different in the Operation Design Domain, there is a troublesome problem that the characteristics of each different Operation Design Domain must be differentiated. [Prior art documents] [Patent documents]

[0011] [Patent Document 1] Korean Patent Publication No. 10-2020-0101517 (2020.08.28) Summary of the Invention [Problem to be solved by the invention]

[0012] The present invention has been made to solve the above-mentioned conventional problems, and an object of the present invention is to provide an autonomous driving safety system and an autonomous driving control method that share an Operation Design Domain based on risk, utilizing a method for calculating the risk of an Operation Design Domain and vehicle-to-vehicle communication technology, thereby complementing the Operation Design Domains that differ for each autonomous vehicle and ensuring the safety of autonomous driving. [Means for solving the problem]

[0013] In order to achieve the above object, the present invention can include the following embodiments.

[0014] An embodiment of the present invention includes an autonomous driving system that controls a vehicle drive unit based on information detected by a sensor unit to drive autonomously, and the autonomous driving system includes an operation design domain update unit that evaluates the risk level of at least one of a static operation design domain (Static Operation Design Domain) and a dynamic operation design domain (Dynamic Operation Design Domain) recognized by the sensor unit during driving and updates the operation design domain (ODD), thereby providing an autonomous driving safety system that shares operation design domains based on the risk level.

[0015] In the above embodiment, the autonomous driving system transmits the updated operation design domain to at least one of a roadside device, another autonomous driving vehicle, and a control server, and shares the updated operation design domain.

[0016] In the above embodiment, the autonomous driving system can include a hierarchical attribute information DB that divides the operation design area into multiple hierarchies and stores hierarchical attributes and weights set for each attribute, a risk calculation module that calculates hierarchical risk by applying the attribute weights stored in the hierarchical attribute information DB according to the recognized operation design area and sums the hierarchical risk to calculate a final risk, and an update module that updates the risk calculated by the risk calculation module to the operation design area.

[0017] In the above embodiment, the Operation Design Domain (ODD) can be classified into a road geometry hierarchy, which is classified into the operation design domain of road structure and road type, road shape and form, and road surface shape, type, and condition; a social infrastructure hierarchy, which is classified into the operation design domain of traffic signals and structures; a temporary restricted area hierarchy, which is classified into the operation design domain of traffic accidents, construction sections, and emergencies due to emergency stopping vehicles, and protected areas; an object hierarchy, which is classified into the operation design domain of roads and surrounding static objects, dynamic objects, and emergent objects; an environmental conditions hierarchy, which is classified into the operation design domain of seasons, weather (climate), and light sources; and a connectivity (communication) hierarchy, which is classified into the operation design domain of communications and control.

[0018] In the above embodiment, the temporary restricted area hierarchy is characterized in that attributes are set that include at least one of the distance between the autonomous vehicle and the start and end points of the protected area, the protected area type, the distance between the autonomous vehicle and the sudden section, the left and right lane positions of the sudden section from the autonomous vehicle, the relative position of the sudden section from the autonomous vehicle, and the sudden section type.

[0019] In addition, the operation design domain of road structure and road type is characterized in that the accident occurrence rate and fatality are set as attributes, and the accident occurrence rate and fatality are applied with different weights depending on the degree of death and injury.

[0020] In addition, the operational design area for the road shape is characterized in that at least one of the following is set as an attribute: lane width, the number of lanes on the entire road, the lane number on which the autonomous vehicle is traveling, the road curvature radius, and the road design speed.

[0021] In addition, the operational design area of ​​the road surface can be set to at least one of the following attributes: road surface material type, road surface friction coefficient (μ), damage type, road surface damage severity (high, medium, low), and distance between the autonomous vehicle and the road surface damage section.

[0022] In the embodiment, the risk level of the operation design area for the road structure and road type is calculated by the following formula:

number

[0023] The weather operation design domain is characterized in that the percentage of clouds, wind speed, precipitation amount, snowfall amount, and visibility distance are set as attributes, the percentage of clouds is set as the percentage of clouds in sunny and cloudy conditions, wind, rain, and snow are classified into multiple categories based on the amount, and fog is classified into multiple categories based on the visibility distance.

[0024] In addition, the object hierarchy is classified into vehicle, pedestrian, and animal operation design areas, and the vehicle operation design area has at least one of the following set as attributes: left / right lane position of the target vehicle from the autonomous vehicle, peripheral position of the target vehicle from the autonomous vehicle, absolute speed (X, Y axes), relative speed from the autonomous vehicle (X, Y axes), distance (X, Y axes), travel angle (direction), detection coordinates, and vehicle type; and the pedestrian and animal operation design area has at least one of the following set as attributes: relative speed from the autonomous vehicle, absolute speed, distance, travel angle, detection coordinates, and type.

[0025] In the above embodiment, the risk level of the operation design area of ​​the road shape is

number

[0026] In another embodiment, the present invention includes a control server that collects operation design areas in real time from multiple autonomous vehicles and roadside devices, and the control server transmits the operation design areas including the risk levels to the multiple autonomous vehicles and roadside devices to update the operation design areas.

[0027] In the above embodiment, the control server can receive operation design areas based on risk levels from autonomous vehicles, compare and evaluate the collected operation design areas, and send updated operation design areas to autonomous vehicles whose risk levels are missing or have errors.

[0028] Here, the control server provides an autonomous driving control method for sharing an operation design area based on risk level, which includes an information collection unit that collects information on the operation design area and road and surrounding conditions from multiple roadside devices and autonomous vehicles, an operation design area design unit that models an operation design area that includes at least one of the static operation design area and dynamic operation design area collected from the autonomous vehicles, and a server communication unit that transmits the operation design area modeled by the operation design area design unit to multiple autonomous vehicles and roadside devices. [Effects of the Invention]

[0029] The present invention allows an autonomous vehicle to assess the level of risk based on road conditions recognized in real time and update the operation design area, and the updated operation design area can be shared with other autonomous vehicles and roadside devices, making it easy to update the operation design area regardless of the vehicle type, system specifications, or year of manufacture.

[0030] Furthermore, the present invention allows an operation design area including risk levels to be designed using information collected in real time by a control server connected to the infrastructure, and this information can be shared with autonomous vehicles and roadside devices via the infrastructure, making it easy to update the operation design area. [Brief explanation of the drawings]

[0031] [Figure 1] 1 is a block diagram showing an overview of an autonomous driving safety system and an autonomous driving control method for sharing a risk-based operation design domain according to the present invention. [Figure 2] 1 is a block diagram of an autonomous vehicle according to the present invention; [Figure 3] FIG. 10 is a block diagram showing the operation design domain update unit. [Figure 4] FIG. 2 is a block diagram showing a control server. [Figure 5] FIG. 2 is a block diagram showing an operation design area design unit. [Figure 6] 1 is a flowchart showing an autonomous driving method for sharing a risk-based operation design area according to the present invention. [Figure 7] 10 is a flowchart showing step S100. [Figure 8] 10 is a flowchart showing step S200. [Figure 9] FIG. 1 is a diagram illustrating a response scenario according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0032] Although the present invention may be modified in various ways and may have various embodiments, a specific embodiment will be illustrated in the drawings and described in detail. This is not intended to limit the present invention to the specific embodiment, but it should be understood that the present invention covers any one of modifications, equivalents, or alternatives within the spirit and technical scope of the present invention for connecting and / or fixing structures extending in different directions.

[0033] The terms used in this specification are merely used to describe particular embodiments and are not intended to limit the present invention. The singular expressions include the plural expressions unless otherwise clearly indicated in the context.

[0034] It should be understood that in this specification, the use of terms such as "comprise" or "have" is intended to specify the presence of a feature, number, step, operation, component, part, or combination thereof stated in the specification, but does not preclude the presence or possible addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0035] The term "Operation Design Domain (ODD)" in this specification refers to the operating domain (e.g., roads, weather, traffic, etc.) in which the functions of an autonomous driving system can be performed normally and safely, as presented in the "Level 4 Autonomous Driving Vehicle Manufacturing Safety Guidelines" of the Ministry of Land, Infrastructure, Transport and Tourism of Korea.

[0036] Hereinafter, preferred embodiments of an autonomous driving safety system and an autonomous driving control method for sharing a risk-based operation design domain according to the present invention will be described with reference to the accompanying drawings.

[0037] FIG. 1 is a block diagram showing an overview of an autonomous driving safety system and an autonomous driving control method that share a risk-based operation design domain according to the present invention.

[0038] Referring to FIG. 1, the present invention may include a control server 200, a plurality of autonomous vehicles 100, and a roadside device 300.

[0039] A plurality of autonomous vehicles 100 autonomously travel on roads while detecting the conditions around the roads, recognize elements of an Operation Design Domain while traveling, evaluate the degree of risk, update the Operation Design Domain to include the degree of risk, and share it via infrastructure (e.g., edge infrastructure). A detailed configuration will be described later.

[0040] The roadside unit (RSU) is, for example, configured of a plurality of devices capable of communication so as to configure a V2X infrastructure and capable of generating, transmitting, and receiving traffic information, as is well known.

[0041] Here, the roadside device 300 can share information with autonomous vehicles via infrastructure, and relay the shared information to communicable autonomous vehicles, other roadside devices 300, and the control server 200 at the control center.

[0042] Control server 200 collects operation design domain information from autonomous vehicles and roadside devices in real time, and compares and evaluates the collected operation design domains. Control server 200 then transmits the updated operation design domain to any autonomous vehicles in which an error or omission in the risk level for the operation design domain has been found, and updates the operation design domain of that autonomous vehicle.

[0043] Here, the control server 200 can model the operation design domain based on the risk level, and use the modeling results as a standard to compare and evaluate the collected operation design domains of autonomous vehicles.

[0044] For example, autonomous vehicles have operation design domains (static operation design domain, dynamic operation design domain) set according to different formats and standards depending on the manufacturer and type of vehicle, and some or all of these must be updated from the originally designed operation design domain in response to changes in time (changes in the characteristics of the operation design domain between day and night, deformation of structures over time, etc.) and traffic (traffic volume, unexpected objects, etc.).

[0045] However, due to various factors such as the personal circumstances of the vehicle owner, differences in autonomous driving systems, differences in vehicle models, and the circumstances of the manufacturer, updates to the operation design domain may not be carried out.

[0046] Therefore, the control server 200 transmits an operation design domain designed based on the latest information to the autonomous vehicle 100 in which an operation design domain that does not match the current situation has been set. In this case, the updated operation design domain can be received via the autonomous vehicle 100 and / or infrastructure, or can be designed by the control server itself.

[0047] Preferably, the control server 200 can compare and evaluate the operation design domain it has designed with the operation design domain of the autonomous vehicle 100 it has received, calculate the latest information, and transmit this information to autonomous vehicles and roadside devices 300 that have not been updated to the operation design domain.

[0048] FIG. 2 is a block diagram of an autonomous vehicle according to the present invention.

[0049] Referring to FIG. 2, the autonomous vehicle 100 includes an autonomous driving system 110, a sensor unit 120, a communication unit 130, and a vehicle driving unit 140.

[0050] The sensor unit 120 is composed of multiple sensors that can detect objects and surrounding conditions in front, behind, and / or on the left and right sides of the autonomous vehicle 100, such as a LiDAR sensor, a radar sensor, a temperature sensor, a humidity sensor, and a rain sensor (Rain).

[0051] The communication unit 130 communicates with the control server 200, the roadside device 300, and other autonomous vehicles 100.

[0052] The vehicle driving unit 140 is configured with devices that perform the vehicle's specific functions such as steering, driving, deceleration, braking, etc. Here, the vehicle driving unit 140 operates under the control of the autonomous driving system 110.

[0053] The autonomous driving system 110 includes a driving control unit 112 that controls the vehicle driving unit 140 to drive autonomously based on scenarios set according to various situations and information detected by the sensor unit 120, and an operation design domain update unit 111 that evaluates and updates the risk level of the recognized operation design domain.

[0054] The driving control unit 112 corresponds to a known autonomous driving system 110, sets a route to an input destination, and controls the vehicle driving unit 140 to drive autonomously. Here, the driving control unit 112 controls the vehicle driving unit 140 according to a scenario set according to various surrounding environments detected during driving, such as weather (snow, rain, fog), road surface conditions (wet, dry, icy, damaged), unexpected objects (vehicles, pedestrians, animals, fallen rocks on the road), and road structures (tunnel, two-lane road, three-lane road, elevated road, back road, paved, unpaved), as well as static operation design areas or dynamic operation design areas (D1, D2, D3, see FIG. 9) detected in real time.

[0055] The operation design domain update unit 111 updates the operation design domain by modeling the static operation design domain or the dynamic operation design domain detected by the sensor unit 120 based on the degree of risk. In addition, the operation design domain update unit 111 updates the operation design domain shared by other autonomous vehicles 100 and / or the control server 200 via the communication unit 130 to the latest operation design domain in accordance with the flow of time and changes in traffic.

[0056] For a detailed description of the operation design domain update unit 111, please refer to FIG.

[0057] FIG. 3 is a block diagram showing the operation design domain update unit.

[0058] Referring to FIG. 3, the operation design area update unit 111 may include a level-specific attribute information DB 111a in which operation design area attribute information is stored, a risk calculation module 111b, and an update module 111c.

[0059] The hierarchy-based attribute information DB 111a hierarchizes the operation design domain and stores attributes and weight information set for each hierarchy.

[0060] In the present invention, the Operation Design Domain is divided into six layers (6-Layer), 6Layer 1 to Layer 6, and each layer is divided into a single layer. The risk level is defined as the sum of the risk levels of each single layer.

[0061] The classification of these hierarchical operation design areas was done by dividing each layer into road geometry, social infrastructure facilities, temporary restricted areas, objects, environmental conditions, and connectivity (communication), as shown in Table 1 below, and classifying the operation design areas by layer.

[0062] [Table 1]

[0063] In addition, the present invention classifies the operation design area for each level into large, medium, and small levels by applying criteria such as type, form, state, and type, and sets attributes for determining risk assessment for each small level. Tables 2 to 9 below are classification tables in which the attributes for each level in Table 1 are set. Specifically, Tables 2 to 4 list detailed classifications and attributes for the road geometry level, Table 5 for the social infrastructure level, Table 6 for the temporary restricted area level, Table 7 for the object level, Table 8 for the environment level, and Table 9 for the connectivity (communication) level.

[0064] [Table 2]

[0065] In Table 2, the attributes of the operational design area for road structure are set to accident rate and fatality, and the risk level is calculated according to the accident rate and fatality level for each road section. The risk level based on the risk factors of road structure is shown in Table 3 below. Table 3 classifies dangerous sections according to the fatality rate for each road section, the Korean Ministry of Land, Infrastructure and Transport standard, and EPDO (Equivalent Property Damage Only).

[0066] [Table 3]

[0067] In addition, the operational design area for road structure types (single roads, intersections, back roads, etc.) can be calculated by combining fatality rates, EPDO, and the standards of the Ministry of Land, Infrastructure, Transport and Tourism of Korea for single roads such as inside tunnels, on bridges, on elevated roads, underground roads, and back roads, and intersections such as three-way intersections, four-way intersections, multi-way intersections, and roundabouts.

[0068] In addition, the operational design domain for road shape (configuration) was set with attributes such as lane width, number of lanes on the entire road, road curvature radius, road design speed, road type (flat or mountainous), longitudinal gradient, and road design speed, depending on the curvature and gradient of the lanes and each road.

[0069] With these attributes, the risk level can take into account the radius of curvature, the longitudinal gradient, and the driving speed. For example, the risk level can be calculated according to the radius of curvature, the longitudinal gradient, and the current driving speed of the autonomous vehicle 100 on the road on which the autonomous vehicle 100 is currently traveling. An example is shown in Table 4 below.

[0070] [Table 4]

[0071] Table 4 above shows the risk level for different driving speeds on a highway with a 5% longitudinal gradient, with the risk level being set higher when the driving speed and longitudinal gradient AMF (Accident Modification Factor) values ​​are low.

[0072] The operational design domain for road surfaces and conditions is classified into pavement, road surface condition, and lane. Among these, the road surface friction coefficient according to the type of pavement (e.g., asphalt, concrete, block, unpaved) is set as an attribute.

[0073] Among these, road surface conditions are further classified into dry, wet, snowy, icy, and damaged. Of these, dry, wet, snowy, and icy have road surface friction coefficients set as attributes. The friction coefficients of dry, wet, snowy, and icy can be applied differently depending on the type of pavement.

[0074] For example, the road surface friction coefficient can be assigned attributes of dry (0.8), wet (0.6-0.7), snow (0.3-0.6), and icy (0.05-0.3) for asphalt pavement, dry (0.8), wet (0.4-0.6), snow (0.3-0.6), and icy (0.05-0.3) for concrete pavement, dry (0.7), wet (0.3-0.4), snow (0.3-0.6), and icy (0.05-0.2) for block pavement, and dry (0.5), wet (0.3-0.4), snow (0.3-0.6), and icy (0.05-0.2) for unpaved roads.

[0075] The above-described friction coefficient is used to calculate the stopping distance for each traveling speed of the autonomous vehicle 100 while traveling on a road with a corresponding road surface condition. The calculated stopping distance is then used as an evaluation criterion for risk calculation.

[0076] The attributes of a lane include the lane type (for example, solid line, dotted line, double solid line, solid and dotted line, zigzag lane, bus lane), the left / right position of the lane from autonomous vehicle 100, and the lane status (normal or abnormal). An abnormal lane status here refers to a situation where the lane cannot be clearly visually identified, such as when the lane has disappeared or is unclear.

[0077] Table 5 below further breaks down the social infrastructure facility hierarchy and sets attributes for each classification.

[0078] [Table 5]

[0079] The traffic signals in the above content can be categorized into caution signs, directional signs, and road signs. Caution signs include warning signs for protected areas (e.g., crosswalks, child protection, wildlife protection), gradients (e.g., uphill, downhill), accident risk areas, constantly congested areas, merging roads, branching roads, intersections, start of median strip, end of median strip, road width limit, loss of lane, two-way traffic, speed bumps, priority roads, roads with falling rocks, construction areas, railroad crossings, tunnels, bridges, crosswinds, etc.

[0080] The type, details and coordinates of the signs are set as attributes for the above-mentioned caution signs, regulatory signs and instruction signs.

[0081] The road markings have the following attributes set: road marking type, detailed road marking content, distance, and sign coordinates.

[0082] In addition, as a subcategory of structures, as shown in the table above, they were further subdivided into road and surrounding road structures such as traffic signal signs, guardrails, median strips, streetlights, speed bumps, utility poles, cones, curbs, manholes, etc., and the attributes of each subcategory were set as road facility type, distance and size, and structure coordinates.

[0083] The temporary restricted area hierarchy is categorized into protected areas and emergency-related operation design areas, and attributes are set for each, as shown in Table 6. Of these, the protected areas (e.g., elderly protection area, child protection area, school zone protection area) have attributes set for the start and end points of each protected area, the distance between the autonomous vehicle 100 and the protected area, and the protected area type (e.g., speed reduction, caution).

[0084] [Table 6]

[0085] An unexpected occurrence refers to a traffic accident, a construction section, or an emergency vehicle stop, and the attributes are set to the distance between the autonomous vehicle 100 and the unexpected section, the position of the unexpected section, the relative position, and the unexpected section type.

[0086] The object hierarchy includes objects that may appear while driving on a road, as shown in Table 7 below, and is classified into vehicles, pedestrians, animals, etc.

[0087] [Table 7]

[0088] Of these, vehicles include, for example, commercial vehicles, passenger cars, agricultural machinery, motorcycles, motor vehicles, bicycles, special vehicles, freight vehicles, police cars, and special vehicles such as fire engines, and attributes are set such as relative lane position, lane position, absolute speed, relative speed, travel angle, and detection coordinates.

[0089] Pedestrians can also be classified as adults, elderly people, children, and others such as police officers or construction workers.

[0090] Attributes of pedestrians and animals include relative speed from autonomous vehicle 100, absolute speed, distance, travel angle, detected coordinates, and type.

[0091] The environmental hierarchy is a hierarchy regarding the environment when autonomous vehicle 100 is traveling on a road, and can be classified into seasons, weather, and light sources as shown in Table 8 below.

[0092] [Table 8]

[0093] The weather in the above example has attributes set to cloud percentage, wind speed, precipitation, snowfall, and visibility. Here, cloud percentage is a value used to determine whether the weather is sunny or cloudy. For example, a sunny state is one in which clouds account for 0-50% of the sky, and a cloudy state is one in which clouds account for 90-100% of the sky.

[0094] Additionally, among the weather attributes, wind is set to wind speed (e.g., strong: 25m / s or more, medium: 20-25m / s, low: 15m / s), rain is set to precipitation amount (e.g., strong: 30mm or more, medium: 15-30mm, low: 1-15mm), snow is set to snowfall amount (e.g., strong: 30mm or more, medium: 15-30mm, low: 1-15mm), and fog is set to visibility distance (e.g., strong: 40m or less, medium: 40-200m, low: 200-1000m).

[0095] The connectivity (communication) layer is a layer for communication with the control server 200 and / or surrounding infrastructure for operating the autonomous driving system 110, and is classified and attributed as shown in Table 9 below.

[0096] [Table 9]

[0097] The risk calculation module 111b models an operation design domain based on the risk depending on whether the static operation design domain or the dynamic operation design domain is recognized during driving, updates the operation design domain to suit the site, and reflects it in the existing operation design domain. As an example, this risk calculation module will be described using an embodiment in which a risk is calculated in a road geometry hierarchy.

[0098] The risk calculation module 111b can apply weights when calculating the risk according to the above-mentioned hierarchical attributes. Here, the risk can be calculated using a formula such as Equation 1 below.

[0099]

number

[0100] In the present invention, the risk level (Total Layer Risk Value) of the situation in the operation design domain recognized while the autonomous vehicle 100 is traveling corresponds to the sum of the risk levels (Single Layer Risk Value) for each layer.

[0101] In addition, the Single Layer Risk Value is an attribute value (ODD) classified and calculated according to the layer variables. element ) was applied with a weight.

[0102] The update module 111c updates the risk assessment result by the risk calculation module 111b to the current operation design domain. The driving control unit 112 drives the autonomous vehicle according to the best scenario based on the risk included in the updated operation design domain.

[0103] Hereinafter, an embodiment of the risk calculation and / or evaluation process as described above will be described. Note that the final risk is determined by the sum of the risk levels for each layer, and the embodiment will be described using only some layers selected from the total 6 layers as an example.

[0104] [First embodiment] The risk level according to the road structure (type) of the road geometry layer can be calculated by applying different weights to traffic accidents, for example, fatal accidents, serious injury accidents, and minor injury accidents. In this case, the risk level can be calculated using the following Equation 2.

[0105]

number

[0106] In the above formula 2, a weight of 1 was set for fatal accidents, 0.7 for serious injury accidents, and 0.3 for minor injury accidents.

[0107] In addition, the risk level according to the severity of accidents by road type is calculated by applying the EPDO (Equivalent Property Damage Only) methodology according to the average number of accidents, which is calculated using the following Equation 3.

[0108]

number

[0109] Here, a weight of 12 was applied to fatal accidents and a weight of 3 to injury accidents.

[0110] The weights in Equation 2 and Equation 3 are those applied when Korea Road Traffic Authority selects accident hotspots. This is just an example and is not intended to be limiting. In other words, the weights can be changed depending on the intentions of the designer or operator, traffic regulations, and the characteristics of the road or area.

[0111] The risk level reflecting the results of Equation 2 and Equation 3 can be calculated as shown in Table 10 below.

[0112] Table 10 shows examples of calculated fatality rates for single roads and intersections among road types, as well as risk levels according to traffic risk and EPDO (by number of accidents).

[0113] [Table 10]

[0114] The risk calculation module 111b evaluates the risk by taking into account both the traffic risk and the EPDO result value, and reflects any difference with the existing operation design domain. Such an update of the operation design domain can change the scenario set according to the recognized risk.

[0115] For example, if a scenario in the existing Operation Design Domain is set to slow down the driving speed to 70% or less at 100 km / h when the inter-vehicle distance is 50 m, but if the risk is assessed to be high in the updated Operation Design Domain, the scenario can be changed to slow down to 30% or less or stop the vehicle.

[0116] That is, the present invention can improve the safety of autonomous driving by evaluating the risk level depending on the situation of the operation design domain at the road site and updating the operation design domain.

[0117] Second Embodiment The factors (attributes) that can affect the safety of vehicle driving in the shape elements of the road are the curve radius and the longitudinal gradient. Therefore, the risk calculation module calculates the risk due to the curve radius using the curve radius accident correction coefficient, which is calculated using the accident correction coefficient formula in Table 11.

[0118] [Table 11]

[0119] In addition, the risk calculation module can calculate the risk using the accident modification coefficient for longitudinal gradient. The accident modification coefficient for longitudinal gradient is calculated using the formula in Table 12 below.

[0120] [Table 12]

[0121] Examples of values ​​calculated based on curve radius and longitudinal gradient are shown in Table 13.

[0122] [Table 13]

[0123] In addition, on expressways where many vehicles travel at speeds of 100 km / h or more, in order to calculate the risk level based on curve radius, accident modification factors are calculated for speeds of 110 km / h and 100 km / h, and the curve radius AMF score is as disclosed in Table 14 below.

[0124] [Table 14]

[0125] The risk calculation module then calculates the road shape risk using the curve radius and longitudinal gradient AMF score as described above. The road shape risk is calculated using the following equation 4.

[0126]

number

[0127] Therefore, the road shape risk can be finally calculated as shown in Table 15 below. The table below evaluates the risk for each driving speed of 110km / h and 100km / h on a two-lane expressway, and the higher the score, the higher the risk.

[0128] [Table 15]

[0129] Third Embodiment The road surface is a risk that can arise depending on the pavement material and road surface condition, such as an increase in stopping distance due to changes in the frictional force of the road surface.

[0130] Therefore, the risk calculation module 111b calculates the risk based on the stopping distance by applying the friction coefficient for each road surface condition (dry, wet, icy). First, the stopping distance according to the road surface condition for each driving speed is as shown in Table 16 below.

[0131] [Table 16]

[0132] The risk calculation module can set a higher risk as the stopping distance increases depending on the traveling speed and road surface conditions.

[0133] Furthermore, the risk calculation module sums up the results of the first to third embodiments as described above, and finally calculates the risk for the currently detected situation in the Operation Design Domain. However, for convenience, the above explanation has been given using the first to third embodiments, and the final value for the actual risk is determined as the sum of the risk of all related floors according to the recognized ODD.

[0134] Furthermore, the update module 111c updates the operation design domain according to the risk level. Therefore, the driving control unit 112 can control autonomous driving with a scenario suited to the site according to the operation design domain in which the risk level is reflected by the update module 111c.

[0135] The control server 200 is described in more detail with reference to FIGS.

[0136] FIG. 4 is a block diagram showing the control server, and FIG. 5 is a block diagram showing the operation design area design unit.

[0137] Referring to FIGS. 4 and 5, the control server 200 may include an information collection unit 210, an operation design domain design unit 220, and a server communication unit 230.

[0138] The information collection unit 210 collects information from the roadside device 300 and the autonomous vehicle 100. The collected information may be information about the state of the operation design area recognized while the autonomous vehicle 100 is traveling on a road.

[0139] In addition, the information collection unit 210 can collect various information including the operation design area, road surroundings, and traffic conditions shared by the autonomous vehicle and roadside devices.

[0140] The operation design domain design unit 220 models an operation design domain including the situations of ODDs (e.g., static ODDs, dynamic ODDs, or combined static and dynamic ODDs) collected from an autonomous vehicle.

[0141] For this purpose, the traffic design domain design unit 220 may include a traffic design domain design module 221 , an evaluation module 222 , and an information providing module 223 .

[0142] The operation design domain design module 221 models an operation design domain based on risk. The process and method may be the same as the modeling process of the operation design domain of the autonomous vehicle 100 described above (e.g., the sum of risk levels for each 6 layers).

[0143] The evaluation module 222 compares and evaluates the result values ​​modeled by the operation design domain design module 221 with the operation design domain of the autonomous vehicle and roadside device 300. Here, the evaluation module 222 compares the collected operation design domain with the operation design domain modeled by the operation design domain design module 221, and extracts operation design domains that lack the best update information, such as risk assessment.

[0144] Alternatively, the evaluation module 222 compares the operation design domain design module 221 with the operation design domain of the autonomous vehicle 100, and if the operation design domain of the autonomous vehicle is newer information, updates the operation design domain that it has designed to the operation design domain of the autonomous vehicle.

[0145] The information providing module 223 provides information to the autonomous vehicle 100 and the roadside device 300. In this case, the provided information may include information on real-time incidents and accidents, traffic volume, weather, etc., as well as the latest update information designed by the operation design domain design module.

[0146] In particular, the information providing module 223 can transmit the latest update information of the Operation Design Domain to an autonomous vehicle 100 that is missing the latest update information from a previously collected Operation Design Domain.

[0147] The present invention includes an autonomous driving method capable of updating an operation design domain achieved by the above-described configuration. The autonomous driving method capable of updating an operation design domain will be described below.

[0148] FIG. 6 is a flowchart showing an autonomous driving method capable of updating an operation design domain according to the present invention, and FIG. 7 is a flowchart showing step S100.

[0149] Referring to Figures 6 and 7, the present invention includes step S100 in which the autonomous vehicle 100 shares ODD modeling and information, and step S200 in which the control server 200 transmits ODD modeling and updated ODD modeling information.

[0150] Among these, step S100 is a step in which the autonomous vehicle 100 evaluates the level of risk based on the Operation Design Domain recognized at the site, updates the Operation Design Domain accordingly, and performs autonomous driving according to a scenario based on the level of risk of the on-site situation.

[0151] More specifically, step S100 includes step S110, in which the autonomous vehicle drives autonomously; step S120, in which the surrounding conditions and objects are detected; step S130, in which the level of danger is calculated according to a hierarchy set according to the conditions recognized while driving; step S140, in which the operation design area is updated based on the level of danger; and step S150, in which the update information is shared.

[0152] Step S110 is a step in which autonomous driving of autonomous vehicle 100 begins. The autonomous driving system sets a route according to the input destination, and controls vehicle driving unit 140 to begin driving to the destination.

[0153] Step S120 is a step for detecting an event defined in the Operation Design Domain while autonomous vehicle 100 is traveling. For example, as shown in Fig. 9, autonomous vehicle 100 detects that there is a construction section (D1) in which a worker is present in the lane ahead of the vehicle, that there is an emergency stop vehicle (D3) ahead of the construction section, and that there is a speed bump (D2) ahead of the emergency stop vehicle.

[0154] Here, the construction area (D1) corresponds to a dynamic ODD classified as a temporary restricted area, the emergency stop vehicle (D3) corresponds to a dynamic ODD / event as a sudden object, and the speed bump (D2) corresponds to a static ODD as a social infrastructure facility.

[0155] Step S130 is a step for calculating the risk level for the static ODD and dynamic ODD detected while traveling by the autonomous vehicle 100. For example, a construction section (D1) is classified as a temporary restricted section, an emergency stop vehicle (D3) is classified as a sudden object, and a speed bump (D2) is classified as a social infrastructure facility.

[0156] Therefore, the autonomous vehicle 100 searches for attributes for each layer classified into 6 layers, such as road surface condition, road structure, type, weather, wind, etc., and applies a set weight to each attribute.The autonomous vehicle 100 then calculates a final result value by summing up the risk levels for each layer.

[0157] Step S140 is a step in which the autonomous vehicle 100 updates the operation design domain based on the risk level. The autonomous vehicle 100 applies the calculated risk level to update the existing established operation design domain.

[0158] For example, the updated Operation Design Domain contained a higher level of risk due to the presence of a sudden object (D3) with an emergency stop and a speed bump (D2) in succession than the construction zone ahead (D1).

[0159] Therefore, in an existing scenario, if the autonomous driving system 110 is configured to decelerate to 80% when it reaches a set distance a from a construction section in the lane ahead and then move and drive in the adjacent lane, it controls driving in a scenario in which it decelerates from an even further distance depending on the danger level, and passes through the section at a speed of 10 km / h or less.

[0160] Step S150 is a step in which the autonomous vehicle 100 shares the updated Operation Design Domain. The autonomous vehicle 100 transmits the updated Operation Design Domain to other nearby autonomous vehicles 100, roadside devices 300, and the control server 200. The transmitted information is evaluated by the control server 200 and can be retransmitted to and shared with other autonomous vehicles 100. Alternatively, the autonomous vehicle 100 directly shares information with other autonomous vehicles 100 through one-to-one communication.

[0161] Step S200 will be described with reference to the flowchart of FIG.

[0162] FIG. 8 is a flowchart showing step S200.

[0163] Referring to Figure 8, step S200 includes step S210 in which the control server 200 collects information from the autonomous vehicle 100 and the roadside device 300, step S220 in which the operation design area is modeled, step S230 in which the operation design area of ​​the autonomous vehicle 100 is evaluated, and step S240 in which the operation design area is updated.

[0164] Step S210 is a step in which the control server 200 collects information. The control server 200 collects information on traffic accidents, traffic volume, weather and incidents around the road, etc., including the Operation Design Domain, via infrastructure such as the autonomous vehicle 100 and the roadside device 300.

[0165] Step S220 is a step in which the control server 200 models an operation design domain. Here, the control server 200 designs and / or shares the operation design domain in various ways.

[0166] For example, before step S210, the control server 200 can model the operation design domain based on the information of a specific road section and store it by itself.

[0167] Alternatively, the control server 200 can collect information on the Operation Design Domain collected in step S210, check whether there is an event on the road and surrounding conditions, and model the Operation Design Domain based on information on the section where the event occurred.

[0168] Alternatively, in step S210, the control server 200 collects road section and surrounding information (information classified and attributed at a 6-layer hierarchy) collected through infrastructure such as multiple autonomous vehicles 100 and roadside devices 300, and models an operation design domain, and can model the operation design domain each time information different from existing information is collected.

[0169] Step S230 is a step in which the control server 200 evaluates whether the collected operation design domain of the autonomous vehicle 100 has been updated to the latest information. Since different operation design domains may be installed for the autonomous vehicle 100 depending on the manufacturer and product specifications, updates to the static operation design domain or the dynamic operation design domain may be delayed due to events, weather, road type, or structure.

[0170] Therefore, the control server 200 can collect operation design domains from the plurality of autonomous vehicles 100 in this manner and then evaluate whether each operation design domain has been updated to the latest information. In this case, the comparison object can be either the operation design domain modeled by the control server 200 itself based on information collected in real time, or the operation design domain based on the latest road information collected from the autonomous vehicles 100, or the collected operation design domains can be compared and evaluated to derive an operation design domain updated to the latest information.

[0171] Step S240 is a step in which the control server 200 compares and evaluates the collected operation design domains and transmits the updated operation design domains to the autonomous vehicles 100 that have not been updated to the latest information.

[0172] In other words, the control server 200 can transmit the operation design domain that it has modeled itself or the latest operation design domain received via infrastructure to other autonomous vehicles 100, and provide the latest operation design domain that can respond to static operation design domains and dynamic operation design domains such as sudden situations on the road or sudden objects due to surrounding conditions.

[0173] Therefore, the present invention can solve the conventional problem of the autonomous vehicle 100 being difficult to update due to various reasons such as the system or the manufacturer, thereby creating a safer autonomous driving environment.

[0174] Although one embodiment of the present invention has been described above, a person having ordinary knowledge in the art can modify and change the present invention in various ways by adding, changing, deleting, or adding components within the scope of the concept of the present invention as set forth in the claims, and this can also be said to be within the scope of the present invention. [Explanation of symbols]

[0175] 100 Autonomous Vehicles 110 Autonomous Driving System 111 Operation Design Area Update Department 111a Hierarchical attribute information DB 111b Risk calculation module 111c update module 112 Travel control unit 120 Sensor unit 130 Communications Department 140 Vehicle drive unit 200 Control Server 210 Information Gathering Department 220 Operation Design Area Design Department 230 Server Communication Department 300 Roadside equipment

Claims

1. An autonomous driving system that controls a vehicle drive unit based on information detected by a sensor unit to drive autonomously, The autonomous driving system and an operation design domain update unit that calculates a risk level of at least one of elements of a static operation design domain and a dynamic operation design domain recognized by the sensor unit during traveling, and updates the operation design domain by modeling based on the calculated risk level; The operation design domain update unit A layer-by-layer attribute information DB that divides an operation design domain into multiple layers and stores layer-by-layer attributes and weights set for each attribute; a risk calculation module that calculates a risk level for each level by applying a weight of an attribute stored in a level attribute information DB according to a recognized operation design domain, and calculates a final risk level by summing the risk levels for each level; an update module that updates the risk calculated by the risk calculation module to an operation design domain; The operation design domain is: a road geometry hierarchy categorized into Operation Design Domains of road structure and road type, road shape and configuration, and road surface shape, type, and condition; A social infrastructure facility layer classified into an Operation Design Domain of traffic signals and structures; A temporary restricted area hierarchy is classified into an Operation Design Domain for traffic accidents, construction zones, and emergencies due to emergency vehicle stops, as well as a protected area; an object hierarchy categorized into an Operation Design Domain of road and surrounding dynamic objects; An environmental conditions hierarchy categorized into Operation Design Domains of season, weather (climate), and light source; and the connectivity (communication) layer, which is classified into the Operation Design Domain of communications and control. The risk of the Operation Design Domain of the road shape is [Equation 1] It is calculated by the formula: The AMF curve radius is calculated using the following formula: [Equation 2] is calculated by The AMF longitudinal gradient is calculated using the following formula: [Equation 3] An autonomous driving safety system that shares an operation design area based on risk, characterized in that the area is calculated by the following:

2. The autonomous driving system 2. The autonomous driving safety system for sharing an operation design domain based on risk levels according to claim 1, wherein the updated operation design domain is shared with at least one of a roadside device, another autonomous driving vehicle, and a control server.

3. The temporary restricted area hierarchy is as follows: The autonomous driving safety system for sharing an operational design area based on risk levels described in claim 1, characterized in that attributes are set including at least one of the following: the distance between the autonomous vehicle and the start and end points of the protected area; the protected area type; the distance between the autonomous vehicle and the sudden departure section; the left and right lane positions of the sudden departure section from the autonomous vehicle; the relative position of the sudden departure section from the autonomous vehicle; and the sudden departure section type.

4. The Operation Design Domain of road structure and road type is: Accident occurrence rate and fatality are set as attributes, The autonomous driving safety system according to claim 1, wherein the accident occurrence rate and fatality rate are weighted differently depending on the degree of death and injury.

5. The Operation Design Domain of the road shape is:

2. The autonomous driving safety system for sharing a risk-based operation design domain as described in claim 1, wherein at least one of the following is set as an attribute: lane width, the number of lanes on the entire road, the lane number on which the autonomous vehicle is traveling, the road curvature radius, and the road design speed.

6. The Operation Design Domain of the road surface is: The autonomous driving safety system for sharing a risk-based operation design domain described in claim 1, characterized in that at least one of the following is set as an attribute: road surface material type, road surface friction coefficient (μ), damage type, road surface damage severity (high, medium, low), and distance between the autonomous driving vehicle and the road surface damaged section.

7. The risk level of the operation design domain (ODA domain) of the road structure and road type is calculated by the following formula: The autonomous driving safety system for sharing a risk-based operation design domain as described in claim 1, characterized in that in the formula, a weight of 1 is set for a fatal accident, a weight of 0.7 for a serious injury accident, and a weight of 0.3 for a minor injury accident.

8. The Weather Operation Design Domain is Attributes include cloud percentage, wind speed, precipitation, snowfall, and visibility distance. The cloud percentage is set to the percentage of clouds in sunny and cloudy conditions. Wind, rain, and snow are categorized into multiple categories based on volume. The autonomous driving safety system for sharing a driving design area based on risk levels according to claim 1, wherein fog is classified into a plurality of categories based on visibility distance.

9. The object hierarchy is divided into Operation Design Domains for vehicles, pedestrians, and animals. The vehicle operation design area is At least one of the left / right lane position of the target vehicle from the autonomous vehicle, the peripheral position of the target vehicle from the autonomous vehicle, the absolute speed (X, Y axes), the relative speed (X, Y axes) from the autonomous vehicle, the distance (X, Y axes), the traveling angle (direction), the detected coordinates, and the vehicle type is set as an attribute; The Pedestrian and Animal Operation Design Domain is The autonomous driving safety system that shares a risk-based operation design domain as described in claim 1, characterized in that at least one of the relative speed from the autonomous driving vehicle, absolute speed, distance, traveling angle, detection coordinates, and type is set as an attribute.

10. A control server that collects an operation design domain in real time from a plurality of autonomous vehicles and roadside devices, The control server is The operation design domain collected from the autonomous vehicle is modeled based on a calculated risk level of at least one of elements of a static operation design domain and a dynamic operation design domain recognized by a sensor unit during driving, and the calculated risk level is calculated. The operation design domain including the risk level is transmitted to a plurality of autonomous vehicles and roadside devices to update the operation design domain; The operation design domain is: a road geometry hierarchy categorized into Operation Design Domains of road structure and road type, road shape and configuration, and road surface shape, type, and condition; A social infrastructure facility layer classified into an Operation Design Domain of traffic signals and structures; A temporary restricted area hierarchy is classified into an Operation Design Domain for traffic accidents, construction zones, and emergencies due to emergency vehicle stops, as well as a protected area; an object hierarchy categorized into an Operation Design Domain of road and surrounding dynamic objects; An environmental conditions hierarchy categorized into Operation Design Domains of season, weather (climate), and light source; and the connectivity (communication) layer, which is classified into the Operation Design Domain of communications and control. The risk level is calculated as follows: The risk level for each level is calculated by applying the weight of the attributes set according to the recognized operation design domain, and the risk levels for each level are summed to calculate the final risk level. The risk of the Operation Design Domain of the road shape is [Equation 4] It is calculated by the formula: The AMF curve radius is calculated using the following formula: [Equation 5] is calculated by The AMF longitudinal gradient is calculated using the following formula: [Equation 6] An autonomous driving safety system that shares an operation design area based on risk, characterized in that the area is calculated by the following:

11. The control server is receiving an operation design area based on the risk level from the autonomous vehicle; The autonomous driving safety system for sharing operational design areas based on risk levels described in claim 10, characterized in that the collected operational design areas are compared and evaluated, and an updated operational design area is sent to an autonomous driving vehicle whose risk level is missing or has an error.

12. The control server is an information collection unit that collects information on an operation design domain, roads, and surrounding conditions from a plurality of roadside devices and autonomous vehicles; A static operation design domain and a dynamic operation design domain are collected from the autonomous vehicle. The static operation design domain and the dynamic operation design domain are collected from the autonomous vehicle. The static operation design domain and the dynamic operation design domain are collected from the autonomous vehicle. The operation design domain modeling unit models the static operation design domain and the dynamic operation design domain. and a server communication unit that transmits the operation design domain modeled by the operation design domain design unit to a plurality of autonomous vehicles and roadside devices.

13. The Operation Design Area Design Department an operation design domain design module that models an operation design domain based on risk; an evaluation module that compares and evaluates an operation design domain modeled by the operation design domain design module with an operation design domain collected from an autonomous vehicle and roadside devices, and extracts an operation design domain that does not include a risk; and an information providing module that transmits the updated operation design domain to the autonomous vehicle and the roadside device via a server communication unit.

13. An autonomous driving safety system that shares an operation design domain based on the risk level according to claim 12.

14. a) recognizing elements of an operation design domain in real time while the autonomous vehicle is traveling, calculating the risk of the elements, and updating the operation design domain by modeling based on the calculated risk; b) The autonomous vehicle transmits and shares the updated operation design domain with at least one of the control server, other autonomous vehicles, and roadside devices; In step a), The operation design domain is: a road geometry hierarchy categorized into Operation Design Domains of road structure and road type, road shape and configuration, and road surface shape, type, and condition; A social infrastructure facility layer classified into an Operation Design Domain of traffic signals and structures; A temporary restricted area hierarchy is classified into an Operation Design Domain for traffic accidents, construction zones, and emergencies due to emergency vehicle stops, as well as a protected area; an object hierarchy categorized into an Operation Design Domain of road and surrounding dynamic objects; An environmental conditions hierarchy categorized into Operation Design Domains of season, weather (climate), and light source; and the connectivity (communication) layer, which is classified into the Operation Design Domain of communications and control. The risk level is calculated as follows: The risk level for each level is calculated by applying the weight of the attributes set according to the recognized operation design domain, and the risk levels for each level are summed to calculate the final risk level. The risk of the Operation Design Domain of the road shape is [Equation 7] It is calculated by the formula: The AMF curve radius is calculated using the following formula: [Equation 8] is calculated by The AMF longitudinal gradient is calculated using the following formula: [Equation 9] An autonomous driving control method for sharing an operation design area based on risk, characterized in that the risk is calculated by:

15. a) the step a-1) a step in which an autonomous vehicle autonomously drives; a-2) detecting an event defined in an operation design domain while the autonomous vehicle is autonomously traveling; a-3) calculating the risk of an event that includes at least one element of the static operation design domain and the dynamic operation design domain detected by the autonomous vehicle while traveling; a-4) calculating the risk of elements of the operation design area recognized by the autonomous vehicle, modeling the area based on the calculated risk, and updating the operation design area; and autonomously driving according to a scenario set in the updated operation design area.

16. The autonomous driving control method for sharing an operation design domain based on a risk level according to claim 14, further comprising the step of: the control server receiving the updated operation design domain and transmitting the updated operation design domain to other autonomous vehicles and roadside devices.

17. 15. The autonomous driving control method for sharing operation design domains based on risk levels according to claim 14, further comprising the steps of: a control server collecting operation design domains in real time from a plurality of autonomous vehicles and roadside devices, evaluating and updating the risk levels of the collected operation design domains, comparing and evaluating the updated operation design domains with the collected operation design domains, and transmitting the updated operation design domains to the autonomous vehicles and roadside devices that transmitted operation design domains with missing or erroneous risk levels.

Images