Method for adapting the driving behavior of an autonomous vehicle, autonomous vehicle, special-purpose vehicle and system

Abstract

Method for adapting the driving behavior of an autonomous vehicle (4) comprising the steps: • Determining (100) a mission-specific characteristic of a special-purpose vehicle by the autonomous vehicle (4), wherein the determination (100) is carried out at least via a camera sensor system and the mission-specific characteristic includes at least one characteristic of a light signal of a signal light of the special-purpose vehicle, wherein reflections are used to sensorially detect an approaching special-purpose vehicle, wherein the reflections with respect to the characteristic color and / or frequency of the light signal are detected over several individual images; • Analyzing (200) an environmental situation concerning the autonomous vehicle (4) and the special operations vehicle, including a process based on the identified mission-specific characteristic by the autonomous vehicle (4) and creating data relating to the analyzed environmental situation; and • Comparing (300) data concerning the analyzed environmental situation with a predefined reference, wherein the predefined reference is assigned a criterion regarding performing an adjustment of the driving behavior of the autonomous vehicle (4).

Description

State of the art

[0001] The present invention relates to a method for adapting the driving behavior of an autonomous vehicle, an autonomous vehicle, a special-purpose vehicle and a system.

[0002] The development of automated vehicles is currently a strong trend. In the field of highly automated driving, the main challenges lie in ensuring the necessary functionality and reliability of the systems in all situations, i.e., at all times and in all places. One step towards full automation in all situations, without the need for driver handover in specific situations, is to implement full automation only within a well-defined, known infrastructure. According to the current state of technology, there are numerous situations that automated vehicles cannot reliably detect in order to react appropriately—as a human driver already can.

[0003] Patent application DE 10 2017 126 790 A1 discloses a method for a vehicle to avoid an emergency vehicle. The method comprises retrieving sensor data from a multitude of sensors on the vehicle; detecting that the emergency vehicle is approaching the vehicle on a roadway based on the retrieved sensor data; determining an avoidance strategy with which the vehicle can avoid the emergency vehicle based on additional sensor data; and automatically controlling vehicle components to cause the vehicle to implement the avoidance strategy.

[0004] Document US 9 278 689 B1 reveals autonomous vehicle detection of and response to emergency vehicles. Disclosure of the invention

[0005] According to a first aspect, the present invention relates to a method for adapting the driving behavior of an autonomous vehicle. For the purposes of this invention, "vehicles" include, for example, automobiles, in particular passenger cars and / or trucks and / or motorcycles, or aircraft and / or ships. In the context of an autonomous vehicle, "autonomous" preferably refers to an operating mode at an autonomy level of three or higher according to SAE J 3016. The method according to the invention comprises the step of detecting, in particular by means of sensors, an operational feature, for example, an acoustic signature of a siren and / or a flashing blue light and / or yellow light, of a special-purpose vehicle. In particular, an "operational feature" of a special-purpose vehicle can be understood to be a feature that is used only within the context of the operation.For example, determining an operational characteristic may be preceded by determining and recognizing a specific characteristic of the special-purpose vehicle, whereby a specific characteristic may include the size and / or shape of the vehicle. For instance, in response to the determined and recognized specific characteristic, the operational status of the special-purpose vehicle may be determined based on this characteristic. This determination is carried out by the autonomous vehicle, for example, using its acoustic sensors. The term "special-purpose vehicle" includes, in particular, fire engines, police vehicles, ambulances, emergency medical vehicles, snow removal vehicles (especially snowplows), street sweepers, heavy transport vehicles, technical relief vehicles, armored vehicles, and / or police vehicles equipped with water cannons.In particular, the identification of the special-purpose vehicle's specific characteristics is followed by the autonomous vehicle's recognition of the special-purpose vehicle as such. This can be achieved, in particular, via an evaluation unit of a corresponding sensor, specifically a CPU and / or a microcontroller and / or an electronic control unit. Optionally, the method can also include the autonomous vehicle wirelessly receiving, in particular via radio and / or GPS, mission-specific information transmitted ahead of the special-purpose vehicle. Mission-specific information includes, in particular, the mission status (for example, an acute emergency or a training exercise) of a special-purpose vehicle. The information transmitted by the special-purpose vehicle can be sent, in particular, via any suitable wirelessly communicating transmitting units.In particular, wireless communication can take place via "special-purpose vehicle-to-X communication," where "X" can comprise either the autonomous vehicle or an intermediate storage unit, especially a mobile edge cloud. When communicating via an intermediate storage unit, the mission-specific information transmitted by the special-purpose vehicle can be retrieved by an autonomous vehicle. In a further step of the method according to the invention, an environmental situation concerning the autonomous vehicle and the special-purpose vehicle is analyzed based on the determined mission-specific feature. Optionally, the analysis of an environmental situation concerning the autonomous vehicle and the special-purpose vehicle can also be performed based on the transmitted mission-specific information. For example,The analysis includes the distance between the autonomous vehicle and the emergency response vehicle, the trajectory of the emergency response vehicle, and / or the operational status of the emergency response vehicle. Specifically, the analysis encompasses the aggregation of all available sensor-acquired information regarding both the emergency response vehicle and the autonomous vehicle. In particular, the collected information, determined and / or received with respect to the mission-specific characteristic, can be assigned to the current traffic situation of the autonomous vehicle. Within the autonomous vehicle, data concerning the analyzed environment is then generated for further internal data processing. This analysis is also based on traffic data from the autonomous vehicle.Such traffic data can include the position, speed, direction of travel, planned trajectory, and / or route of the autonomous vehicle. Furthermore, traffic data can include road type (e.g., highways and / or rural roads), contextual details (e.g., tunnels, roundabouts, lanes, and / or brightness, e.g., day or night), traffic regulations (e.g., speed limits, one-way signs, and / or traffic light phases). Thus, such traffic data can comprise either ego data (i.e., data from the autonomous vehicle) and / or environmental data. Therefore, the traffic data can be acquired either by the autonomous vehicle's sensors and / or provided via a wireless network by the special response vehicle.In other words, the autonomous vehicle can analyze information gathered by its sensors and, optionally, information transmitted by the emergency response vehicle to determine whether a particular environmental situation exists, and which environmental parameters, such as distance, are relevant to the autonomous vehicle in relation to the emergency response vehicle. After analyzing the environmental situation and generating data related to it—specifically the positions of the emergency response vehicle and the autonomous vehicle—this data can be compared with a predefined reference, stored, for example, on the autonomous vehicle's memory or in the cloud. This comparison determines whether driving behavior needs to be adjusted.In particular, a criterion regarding the implementation of an adjustment to driving behavior is assigned to the predefined reference. Specifically, the criterion may state that the emergency vehicle must be in operation. For example, the criterion may additionally or alternatively state that the proximity of the emergency vehicle and / or its use of the same lane and / or a foreseeable encounter between the emergency vehicle and the autonomous vehicle, especially at a traffic light, necessitates an adjustment of the autonomous vehicle's driving behavior.For example, if a comparison of the data concerning the analyzed environmental situation with the predefined reference reveals that an emergency vehicle is in the immediate vicinity, particularly at a distance of 30 meters from the autonomous vehicle, the predefined reference can include the criterion that an adjustment of the autonomous vehicle's driving behavior is necessary at a distance of ≤ 50 meters, which could, for example, result in a reduction in speed. In particular, an autonomous vehicle can additionally or alternatively form an emergency lane using the method according to the invention if, for example, a blue flashing light signal has been detected by the autonomous vehicle and, moreover, the autonomous vehicle would obstruct the emergency vehicle due to its position in the respective lane.The inventive method thus has the advantage that interaction between autonomous vehicles and emergency vehicles is possible without hindering the latter. Furthermore, an emergency vehicle can be explicitly integrated into the traffic context of the autonomous vehicle, thereby avoiding unnecessary adjustments to its driving behavior by the autonomous vehicle. Additionally, driving comfort can be improved, as no driver handover with regard to an emergency vehicle is necessary within an autonomous operating mode.

[0006] The dependent claims contain advantageous further developments and embodiments of the method according to the invention.

[0007] According to an advantageous embodiment of the method according to the invention, this includes the step of generating a specification for driving behavior adapted to the analyzed environmental situation by the autonomous vehicle when the criterion of the predefined reference is met. In particular, this specification can include slowing down and / or forming an emergency lane and / or driving further to the right within the given driving corridor and / or stopping at a traffic light despite the green phase and / or taking a wide detour around hazards. For example, adapting the driving behavior can include adapting a driving trajectory, wherein an adaptation of a driving trajectory includes lateral guidance for following this trajectory and / or providing information and / or warnings to the vehicle occupant, for example, by controlling a human-machine interface.

[0008] According to the invention, the detection is carried out using camera sensors. The camera sensors can, in particular, detect and recognize the external appearance of the special-purpose vehicle, for example, its shape and / or size, specific features such as optical signal lights and / or vehicle markings and / or roof structures and / or the presence of aluminum roller shutters. If such a specific feature is present, the sensors can, in particular, detect an operational feature. Additionally or alternatively, the detection and, in particular, the recognition of a specific feature of a special-purpose vehicle can be carried out using radar sensors. This can be done, in particular, based on a radar signature, such as that expected for special-purpose vehicles.Special-purpose vehicles exhibit distinctive external features that can be used for detection, depending on the resolution of the radar sensor and the performance of the subsequent evaluation unit. Roof-mounted equipment is particularly suitable for this purpose. Additionally or alternatively, the special-purpose vehicle can be detected using LiDAR sensors. In this case, a specific LiDAR signature, similar to that of radar sensors, is particularly relevant. Furthermore, a mission-specific characteristic of the special-purpose vehicle can be determined and, in particular, detected using acoustic sensors. Suitable acoustic sensors include a microphone and / or an ultrasonic sensor, which can be installed on autonomous vehicles. In this case, the signal characteristics of the special-purpose vehicle, especially if it is equipped with a siren, can be used to determine a mission-specific characteristic.Thus, according to the invention, special emergency vehicles in particular can be recognized both on the basis of their external appearance and on the basis of emergency signaling, for example a siren, depending on the sensor technology.

[0009] According to an advantageous embodiment of the method according to the invention, indirect communication can take place between the special-purpose vehicle and the autonomous vehicle, such that the mission-specific information transmitted by the special-purpose vehicle is first stored on an intermediate storage unit and made available for reception, and in particular for retrieval, by the autonomous vehicle. In particular, communication between the special-purpose vehicle and X is provided here. Here, X can represent, in particular, the autonomous vehicle (in the case of direct communication) or the intermediate storage unit. In particular, the mission-specific information can include the type of special-purpose vehicle and / or its operational status. Furthermore, the information can include the position and / or speed of the special-purpose vehicle and / or a planned route or trajectory of the special-purpose vehicle.For example, vehicle-to-vehicle communication can take place, enabling direct communication between the emergency response vehicle and the autonomous vehicle. Vehicle-to-intermediate storage unit communication additionally or alternatively enables, in particular, a vehicle-to-X infrastructure (for example, in a transition with a mobile network or mobile edge cloud communication). Furthermore, the transmission of information can include multicasting, i.e., transmission to selected recipients. In particular, selective autonomous vehicles located near the emergency response vehicle can receive the transmitted information. Additionally, broadcasting can take place, i.e., transmission of the information transmitted by the emergency response vehicle to all reachable recipients. Additionally or alternatively, geo-broadcasting can take place, i.e., transmission to all vehicles within a relevant region.This has the particular advantage that the operational area of ​​the special-purpose vehicle-to-X communication, for example via radio channels, is independent of the optical detection range of the autonomous vehicle's sensors. In particular, this method therefore also works in situations with permanent and / or temporary obstruction of vision and with a very long range.

[0010] According to an advantageous embodiment of the method according to the invention, the detected specific feature, which prompts the determination of an application-specific feature, can comprise a size and / or a shape and / or lettering of the special-purpose vehicle and / or a radar signature. The application-specific feature can comprise an activated signal light and / or an acoustic signature of the special-purpose vehicle. According to the invention, the application-specific feature comprises a characteristic of a light signal from the signal light. A characteristic of a light signal can, in particular, relate to a reflection from the active signaling system of the special-purpose vehicle. In particular, these special signals of the active signaling system can be reflected from objects, especially trees and / or vehicles and / or building walls or road surfaces and / or mirrors.According to the invention, these characteristic reflections are used to detect an approaching special-purpose vehicle using sensors and thus to determine an operational feature of the special-purpose vehicle. According to the invention, the reflections are detected across multiple individual frames with respect to the characteristic color and / or frequency of the light signal. This frequency of the light signal can be determined, for example, with the aid of a Fourier transform in the color space of the relevant image area using a camera sensor, even with just a few individual frames. Furthermore, the distance or the change in distance can be estimated by measuring the change in the intensity of the reflection.Using a localization unit, which is based on GPS and / or mobile communications, for example, the autonomous vehicle can determine its current position, and with the help of a database, it can obtain data concerning the region-specific characteristics of the light signals of special emergency vehicles.

[0011] According to an advantageous embodiment of the method according to the invention, this further comprises a machine learning step of the autonomous vehicle depending on the data generated by the autonomous vehicle concerning the analyzed environmental situation. Furthermore, the machine learning can, in particular, include neural networks, such as decision trees and / or Bayesian learning and / or genetic / evolutionary algorithms and / or reinforcement learning and / or deep learning.

[0012] The following aspects of the invention exhibit the advantageous embodiments and further developments with the features mentioned above, as well as the general advantages of the method according to the invention and the associated technical effects. To avoid repetition, a further listing is therefore omitted. Naturally, the aspects of the invention described below exhibit the technical features, advantages, and technical effects of the first aspect of the invention accordingly.

[0013] According to a second aspect, the present invention relates to a special-purpose vehicle comprising a transmitter unit configured to transmit mission-specific information to an autonomous vehicle in order to carry out the method according to the first aspect of the invention. In particular, all transmitter units known to those skilled in the art (e.g., optical and / or acoustic and / or antennas for wireless communication, especially for mobile communications and / or WLAN) are suitable as transmitter units.

[0014] According to a third aspect of the invention, the present invention relates to an autonomous vehicle comprising a receiving unit which is configured to receive mission-specific information from the special-purpose vehicle in order to carry out the method according to the first aspect of the invention.

[0015] According to a fourth aspect of the invention, the present invention relates to a system comprising an autonomous vehicle according to the third aspect of the invention and a special-purpose vehicle according to the second aspect of the invention. In particular, country-specific adaptations can be made to the method according to the invention. This can be implemented dynamically, in particular, by providing additional data when the autonomous vehicle enters a specific geographical region (for example, via geofencing) and subsequently reparameterizing the method according to the invention in accordance with an infrastructure. Brief description of the drawings

[0016] Exemplary embodiments of the invention are described in detail below with reference to the accompanying drawing. The drawing shows: Fig. 1 an illustration of a variant of the method according to the invention; Fig. 2 a flowchart according to a variant of the method according to the invention; Fig. 3 a variant of a special-purpose vehicle according to the invention; Fig. 4 a variant of a system according to the invention; and Fig. 5 a variant of an autonomous vehicle according to the invention. Embodiments of the invention

[0017] Fig. Figure 1 illustrates a situation according to a variant of the method according to the invention. In particular, a police vehicle 1 in operation emits an acoustic signal 5 in the form of a sound wave, which emanates from a siren 2. Therefore, in the present case, the acoustic sound waves 5 are the operational feature of the police vehicle 1. The sound waves can be detected by the autonomous vehicle 4 via a first and a second microphone 3a, 3b. The exact position of the police vehicle 1 can be determined by triangulation of the signal received by the first and second microphones 3a, 3b, taking the Doppler effect into account. Thus, the position of the police vehicle 1 can be determined recognizably for the autonomous vehicle 4. The criterion for adapting the driving behavior, which is associated with the predefined reference, can be fulfilled in this way.In the present case, a predefined reference can be assigned a requirement to relinquish the right-of-way rule. Thus, in particular, the autonomous vehicle 4 can relinquish its right-of-way rule vis-à-vis the police vehicle 1, which is in operation, by initiating an adjustment of its driving behavior in the form of braking, according to the invention.

[0018] Fig. Figure 2 shows a flowchart according to a variant of the method according to the invention. In a first step 100, mission-specific information in the form of a trajectory of a special-purpose vehicle is received by the autonomous vehicle 4. In particular, this mission-specific information is transmitted by the special-purpose vehicle via vehicle-to-vehicle communication. Based on this trajectory information and its own speed and / or traffic light phase, the autonomous vehicle 4 can analyze the surrounding situation in a second step 200 and generate data concerning the surrounding situation based on this analysis. For example, a predefined reference can be assigned a criterion that the trajectories of the special-purpose vehicle and the autonomous vehicle 4 must not intersect. In the present case, the traffic light phase for the autonomous vehicle is green.However, due to the analysis of the environmental situation, the criterion according to the predefined reference is violated by the comparison in the third step 300. With regard to the predefined reference, a driving behavior adapted to the analyzed environmental situation is now defined for the autonomous vehicle 4 in a fourth step 400. In this case, the adapted driving behavior includes stopping at the red light so that the emergency vehicle, which also has a red light phase, can pass through, thus preventing the autonomous vehicle 4 from obstructing the emergency vehicle's trajectory. In a fifth step 500, the driving behavior of the autonomous vehicle 4 is adjusted so that it brakes to allow the emergency vehicle to pass.

[0019] Fig. Figure 3 shows a variant of a special-purpose vehicle according to the invention in the form of a fire engine 20. The fire engine 20 includes, for example, an operation-specific feature in the form of a blue light 15, which can be detected by the autonomous vehicle, as already fully described. Furthermore, the fire engine 20 includes a transmitter 14, which is configured to transmit operation-specific information to an autonomous vehicle 4.

[0020] Fig. Figure 4 shows a system according to the invention. In particular, the system according to the invention comprises a fire engine 20, which has an activated blue light 15. This fire engine 20 specifically comprises a transmitter 14, which is configured to send mission-specific information, here its own trajectory, as shown here by the arrow in front of the fire engine 20, to a receiver 16 of the autonomous vehicle 4. In response to the transmission of the information, the autonomous vehicle 4 can, by carrying out the method according to the invention, veer to the right, as indicated by the corresponding arrow, in order to allow the fire engine 20 to pass.

[0021] Fig.Figure 5 shows an embodiment of an autonomous vehicle 4 according to the invention. This autonomous vehicle 4 according to the invention comprises a receiving unit 16, which is configured to receive mission-specific information transmitted by a special-purpose vehicle in order to carry out the method according to the invention.

Claims

[1] Method for adapting the driving behavior of an autonomous vehicle (4) comprising the steps: • Determining (100) a mission-specific characteristic of a special-purpose vehicle by the autonomous vehicle (4), wherein the determination (100) is carried out at least via a camera sensor system and the mission-specific characteristic includes at least one characteristic of a light signal of a signal light of the special-purpose vehicle, wherein reflections are used to sensorially detect an approaching special-purpose vehicle, wherein the reflections with respect to the characteristic color and / or frequency of the light signal are detected over several individual images; • Analyzing (200) an environmental situation concerning the autonomous vehicle (4) and the special operations vehicle, including a process based on the identified mission-specific characteristic by the autonomous vehicle (4) and creating data relating to the analyzed environmental situation; and • Comparing (300) data concerning the analyzed environmental situation with a predefined reference, wherein the predefined reference is assigned a criterion regarding performing an adjustment of the driving behavior of the autonomous vehicle (4). [2] Method according to claim 1, further comprising the step of generating (400) a specification of driving behavior adapted to the data relating to the analyzed environmental situation by the autonomous vehicle (4) assigned to the predefined reference, if the criterion of the predefined reference is met. [3] Method according to claim 2, further comprising adapting (500) the driving behavior of the autonomous vehicle (4) to the analyzed environmental situation according to the specification. [4] Method according to one of the preceding claims, wherein the detection (100) is carried out via RADAR sensors and / or LIDAR sensors and / or acoustic sensors. [5] Method according to one of the preceding claims, wherein mission-specific information transmitted by the special-purpose vehicle is received directly by the autonomous vehicle (4) and / or is received by the autonomous vehicle (4) via an intermediate storage unit. [6] Method according to one of the preceding claims, wherein the mission-specific feature is determined by responding to the detection of a specific feature of a size and / or a shape and / or a signal light and / or a lettering and / or a RADAR signature of the special-purpose vehicle. [7] Method according to one of the preceding claims, further comprising a step of machine learning by the autonomous vehicle (4) depending on the data relating to the environmental situation of the special emergency vehicle and the autonomous vehicle (4). [8] Method according to one of the preceding claims, wherein the analysis (200) and creation of data relating to the analyzed environmental situation is further based on traffic data of the autonomous vehicle (4). [9] Special-purpose vehicle comprising a transmitter unit (14) which is configured to transmit mission-specific information to an autonomous vehicle (4) in order to carry out the method according to any one of claims 1 to 8. [10] Autonomous vehicle (4) comprising a receiving unit (16) which is equipped to receive mission-specific information transmitted by a special operations vehicle in order to carry out the method according to any one of claims 1 to 8. [11] System for adapting the driving behavior of an autonomous vehicle (4) comprising an autonomous vehicle (4) according to claim 10 and a special-purpose vehicle according to claim 9.

Images