Methods for reducing the risk of wildlife accidents, computer program product, and motor vehicle

Abstract

Method for reducing the risk of wildlife collisions, wherein a motor vehicle (2) is equipped with a wildlife protection device (10) comprising at least one detection arrangement (12, 26) for detecting wild animals (8) in an area (14) located in the direction of travel (x) in front of the motor vehicle (2) and at least one warning device (16), with the detection arrangement (12, 26) detecting at least one wild animal (8) and emitting at least one warning signal by means of the warning device (16) in order to drive the at least one wild animal (8) away from a possible collision area (24) with the motor vehicle (2) or before it enters the possible collision area (24), wherein the wildlife protection device (10) detects the presence of at least one wild animal (8) in the possible collision area (24) in a first step by means of a control device (22) on the basis of signals from the detection arrangement (12, 26),In a second step, upon detection of at least one wild animal (8), the species or genus or species or genera of wild animals (8) in the potential collision area (24) are determined, and the warning device (16) is triggered to emit at least one warning signal tailored to the species or genus of the at least one wild animal (8), characterized in that the species or genus or species or genera of wild animals (8) are determined using a machine vision method, wherein the species or genus or species or genera of wild animals (8) are determined using an algorithm constructed according to artificial intelligence methods using convolutional neural networks (CNNs) and transformer architectures, wherein the localization is designed to be continuous in order to predetermine a corresponding movement trajectory for each identified wild animal (8).

Description

[0001] A method for reducing the risk of wildlife accidents, a computer program product, and corresponding motor vehicles are described.

[0002] Methods for reducing the risk of wildlife accidents, computer program products, and motor vehicles of the type mentioned above are known in the prior art.

[0003] The avoidance or mitigation of collisions in road traffic using passive and active systems is generally known. In particular, motor vehicles with corresponding special safety devices are known, whereby such safety devices can, for example, use generated information to issue warnings to the driver and / or intervene in the longitudinal and / or lateral dynamics of the motor vehicle. This information can be obtained, for example, by means of sensor devices and / or data receivable by the motor vehicle, such as traffic jam warnings.

[0004] A collision between a motor vehicle and any stationary or moving object is imminent. The information can relate to the motor vehicle in relation to the object, allowing for an assessment of the collision risk and, based on this assessment, the initiation of a warning message and / or intervention. The object can also be a living being, particularly a person and / or a wild animal. Wild animals often cannot accurately assess the current danger, which frequently leads to wildlife collisions with serious consequences for the animals, as well as for the occupants of motor vehicles.

[0005] German patent DE 10 2011 102 760 A1 discloses a vehicle, in particular a motor vehicle, with a safety device for preventing the consequences of a collision with a wild animal. To enable improved avoidance or mitigation of the consequences of a collision, the safety device comprises: - a detection device by means of which the presence of a collision risk with the wild animal can be determined, and - a warning device by means of which the wild animal can be alerted to the collision risk, depending on the level of the collision risk.

[0006] DE 10 2018 222 003 A1 discloses a method for emitting an acoustic signal by means of an acoustic signal generator device assigned to a motor vehicle in order to improve road safety for at least one road user in the vicinity of the motor vehicle. The method comprises detecting at least one road user in the vicinity of the motor vehicle. Subsequently, a category is determined to which at least one road user is assigned. Finally, depending on the category, the acoustic signal is generated for the at least one road user.

[0007] The disadvantage of this is that sound is emitted without taking into account the species of wild animal, which may mean that the effectiveness of the acoustic warning is not optimal.

[0008] Therefore, the task is to further develop procedures for reducing the risk of wildlife accidents, computer program products, and motor vehicles of the aforementioned type in such a way that a more effective warning of one or more wild animals is possible.

[0009] The problem is solved by a method for reducing the risk of wildlife collisions according to claim 1, a computer program product according to dependent claim 6, and motor vehicles according to dependent claims 7 and 8. Further embodiments and developments are the subject of the dependent claims.

[0010] A method for reducing the risk of wildlife collisions is described, wherein a motor vehicle is equipped with a wildlife protection device, comprising at least one detection arrangement for recognizing wild animals in an area in front of the motor vehicle in the direction of travel and at least one warning device, with which the detection arrangement recognizes at least one wild animal and emits at least one warning signal by means of the warning device in order to drive the at least one wild animal away from a possible collision area with the motor vehicle or before it enters the possible collision area, wherein the wildlife protection device, by means of a control device, detects the presence of at least one wild animal in the possible collision area in a first step based on signals from the detection arrangement.In a second step, upon detection of at least one wild animal, the species or genus or species or genera of wild animals in the potential collision area are determined, and the warning device is triggered to emit at least one warning signal tailored to the species or genus of the at least one wild animal.

[0011] Vehicles equipped with such a wildlife warning device are able to detect wild animals and issue a targeted acoustic warning without polluting the area in front of the vehicle with noise. By using a loudspeaker system, the sound is directed towards the wild animal, so that essentially only the animal can hear the warning tone and react accordingly. Such a reaction typically consists of the animal being startled and running away from the noise, generally moving out of the danger zone before a collision can occur.

[0012] In this sense, wild animals are any free-roaming animals and include field, forest and pasture animals such as deer, wild boars, bears, wolves, martens, squirrels, hedgehogs, foxes, raccoons, badgers, cows, goats, sheep, horses, donkeys, but also free-roaming domesticated animals such as dogs or cats.

[0013] The system uses acoustic signals to detect and deter wild animals and other wildlife from entering roads, thereby improving road safety. Particularly when driving on rural, forested, or alpine roads, the detection system registers the presence of animals such as cows or sheep in or near a potential collision zone. Upon detection, the vehicle's wildlife deterrent device emits targeted warning sounds designed to safely drive the animals away from the collision zone or prevent them from entering it. This system effectively prevents road blockages and reduces the risk of accidents without compromising animal welfare.

[0014] It was surprisingly discovered that acoustic deterrent methods in animal husbandry can keep certain areas clear, for example, using virtual fences that emit acoustic signals to the animals. The studies conducted showed that specific frequencies and sound patterns are effective in keeping animals away from certain areas without harming them. Based on these findings, a wildlife protection device can be developed that not only increases road safety but also considers animal welfare.

[0015] The process is designed in two stages. First, a smaller, faster model is used to detect the presence of animals in the collision area or danger zone and to perform an initial filtering. This model works specifically with a binary output that indicates whether or not at least one relevant animal is present in the corresponding area.

[0016] Following this initial selection, a more detailed analysis is performed using a more comprehensive model that enables not only the detection but also the localization of the animals. This second stage of analysis is crucial, as it forms the basis for triggering the acoustic signals. The wildlife protection device is designed to recognize and classify a wide variety of animal species, allowing for a differentiated response. Upon detecting at least one animal in the collision zone, for example, on the roadway, specific acoustic signals are triggered, tailored to the respective animal species and designed to safely drive the animal off the road.

[0017] Such signals are developed in collaboration with behavioral scientists to ensure effective deterrence without stressing the animals.

[0018] By integrating these technologies, the system offers a highly effective and efficient solution for preventing collisions with animals in road traffic, thus significantly contributing to increased road safety. At the same time, the intelligent and species-appropriate deterrent effect contributes to animal welfare. The system's flexibility and scalability allow for easy adaptation to various vehicle types and models, underscoring the broad applicability and potential benefits of this technology.

[0019] The process can be easily integrated into modern motor vehicles, providing a technologically advanced solution to the challenges associated with animals on roads.

[0020] Furthermore, it is planned that the species or genus or species or genera of wild animals will be determined using a machine vision method.

[0021] In this way it is possible not only to identify one or more wild animals, i.e., species or genus or species or genera of wild animals in the danger area, but also to locate them accordingly, whereby the localization can be designed continuously in order to predict a corresponding movement trajectory for each identified wild animal and thus be able to assess a collision risk.

[0022] Furthermore, it is planned that the species or genus or species or genera of wild animals will be determined using an algorithm built according to artificial intelligence methods.

[0023] Both established image processing techniques and newer machine learning methods are used to analyze the sensor data. The use of convolutional neural networks (CNNs) and transformer architectures enables the recognition of complex patterns and features in the image data that are relevant for animal identification. These models are trained in an adaptive process based on a large database of animal images to ensure high recognition accuracy.

[0024] In a further, more advanced version, it is provided that if at least one wild animal is present, a warning will be issued in a passenger cabin of the motor vehicle.

[0025] The system's user interaction and controls are designed for intuitive operation and situation-appropriate responsiveness. Upon detection of animals, the driver is immediately alerted via a visual signal in their field of vision and / or an audible warning signal.

[0026] Depending on the driving speed and the situation, the driver can decide, given sufficient time, whether or not to manually activate the acoustic warning system.

[0027] For situations requiring an immediate response, such as at higher speeds, the procedure can be configured, according to further training, to automatically take the appropriate measures.

[0028] This may include, among other things, the automatic playing of the acoustic warning signal and, if necessary, intervention in the vehicle's ESP, for example to automatically reduce speed and / or perform an evasive maneuver in order to maximize the safety of the passengers and at least one wild animal and to give the driver more time to react appropriately.

[0029] The driver interface can be designed in such a way that it allows for quick and easy operation when needed, without distracting the driver from the road.

[0030] In a possible further development, the wildlife protection device can also operate fully automatically, always taking into account the current traffic situation and the safety of all road users.

[0031] In a further, more advanced version, it is planned that the warning signal will be varied in real time.

[0032] The system's acoustic signals are variable in frequency and volume to induce an effective response from wildlife without causing stress or harm. The sounds are designed to safely guide animals off the road. Their development is based on in-depth studies of animal behavior and acoustics to ensure undisturbed and safe interaction. The signals are designed to have an immediate yet calming effect on the animals, prompting them to leave the road. At the same time, care is taken to ensure that the volume remains comfortable for both the driver and passengers and does not impair driving comfort.

[0033] The volume is regulated in real time to ensure that the acoustic signals are effective but not disruptive. This adaptability ensures compliance with legal requirements and promotes acceptance in various applications.

[0034] The wildlife protection device is designed, according to further training, to comply with all relevant noise protection regulations and traffic safety regulations. It can, according to further training, recognize its surroundings and automatically adjust the volume of the acoustic signals to local noise protection standards.

[0035] This adjustment is achieved through a combination of geographical location determination and a database of noise protection regulations, which enables the system to identify proximity to towns and the applicable noise protection regulations.

[0036] A first independent subject matter relates to a computer program product comprising a computer-readable storage medium on which instructions are embedded which, when executed by at least one computing unit, cause that at least one computing unit to be equipped to execute the procedure of the aforementioned type.

[0037] The process can be executed on one or more computing units, so that certain process steps are executed on one computing unit and other process steps on at least one other computing unit, whereby calculated data can be transmitted between the computing units if necessary.

[0038] Another independent item concerns a motor vehicle with a computer program product of the type described above.

[0039] Another independent item relates to a motor vehicle with a wildlife protection device, comprising at least a detection arrangement for detecting wild animals in an area in front of the motor vehicle in the direction of travel and at least a warning device, as well as a control device, wherein the wildlife protection device is configured to carry out the procedure of the type described above.

[0040] The data from these sensors is analyzed in real time to detect animals near the road.

[0041] The wildlife deterrent device can include a network of loudspeakers positioned to emit acoustic signals into the surrounding area. These signals are specifically designed to keep animals away from the roadway without harming or causing long-term impairment. The loudspeakers are engineered to transmit the signals clearly and distinctly over long distances for optimal effectiveness. The integration of these components allows the wildlife deterrent device to act proactively and minimize potential hazards posed by animals on the road.

[0042] In a first further embodiment, it is provided that the recognition arrangement includes at least a front camera, infrared camera, lidar and / or radar.

[0043] The vehicle can be equipped with a comprehensive sensor network that includes a variety of sensor types such as cameras, lidar, infrared cameras and radar.

[0044] These sensors are positioned in such a way as to ensure seamless monitoring of the vehicle's surroundings, with a particular focus on capturing the road in the direction of travel, as this is where the greatest risk of collision is located.

[0045] The data obtained from these sensors can then be processed using various machine vision methods to reliably identify the presence of animals.

[0046] Key components can include high-resolution cameras and a selection of sensors distributed around the vehicle to provide a 360-degree view. These cameras and sensors are designed to operate accurately under a wide range of environmental conditions, such as varying lighting and all weather conditions.

[0047] In a further, more advanced embodiment, it is provided that the warning device has at least one loudspeaker.

[0048] Further advantages, features, and details will become apparent from the following description, in which – possibly with reference to the drawing – at least one embodiment is described in detail. Identical, similar, and / or functionally equivalent parts are identified by the same reference numerals.

[0049] They show schematically: Fig. 1. A top view of a road with a motor vehicle equipped with a wildlife protection device and a wild animal; Fig. 2 an enlarged representation of the motor vehicle's wildlife protection device Fig. 1, as well as Fig. 3. A flowchart.

[0050] In the embodiments described below, identical or equivalent components or elements are provided with the same reference numerals for better readability.

[0051] Fig. Figure 1 shows a motor vehicle 2 traveling on a road 4 in a direction x.

[0052] To the right of road 4 is a forest 6. In the forest 6 is a wild animal, in this case a cow 8, which is moving with a movement trajectory y in the direction of road 4.

[0053] The motor vehicle 2 is equipped with a wildlife protection device 10 (outlined by a dashed line). The wildlife protection device 10 has a camera 12 that monitors an area 14. Depending on its design, the camera 12 can be a standard camera sensitive in the visible spectrum and / or an infrared camera. The camera 12 can be integrated into a camera oriented in the direction of travel x, which can also be used for other purposes, such as detecting the road, traffic signs, traffic lights, hazards, and the like. The camera 12 has a detection range 14, which is defined, among other things, by the sensor size, optical elements such as lenses, apertures, and obstacles.

[0054] The wildlife protection device 10 further comprises a loudspeaker arrangement 16, which is arranged in the front area of ​​the motor vehicle 2. The loudspeaker arrangement 16 is directed towards the wild animal 8 and has a radiation angle 18. Depending on the specific design, the loudspeaker arrangement 16 can have one or more loudspeakers, which can be fixed or movable, for example, pivotable.

[0055] Furthermore, at least one passenger cabin loudspeaker 20 is provided, through which warning tones can be emitted into a passenger compartment of the motor vehicle 2.

[0056] Camera 12, loudspeaker arrangement 16 and passenger cabin loudspeaker 20 are each connected to a control unit 22. The control unit 22 serves to evaluate the signals recorded by camera 12 and to detect whether a wild animal 8 is visible in area 14, to record the relative direction between motor vehicle 2 and wild animal 8, and to calculate the probability that the wild animal 8 will enter a collision area 24, and, if necessary, to control the loudspeaker arrangement 16 accordingly to emit a suitable warning tone in the direction of the wild animal 8.

[0057] In a two-stage process, which is explained in more detail in connection with Fig. As explained in section 3, the camera 12 is used to first detect the presence of the wild animal 8 within its corresponding detection range 14. Then, in a second step, a more computationally intensive artificial intelligence-based image analysis is performed, which identifies the wild animal as a cow. Afterwards, a warning tone effective for cows can be emitted via the loudspeaker system 16 to prevent the wild animal 8 from continuing towards the collision area 24.

[0058] If several animals are found in the detection area 14 of camera 12, the second stage of the procedure can be carried out for each individual wild animal.

[0059] Simultaneously, an acoustic warning can be issued to the driver of the vehicle 2 via the passenger cabin loudspeaker 20, thus alerting the driver of the vehicle 2 to the presence of the wild animal 8. Further warnings can be given, for example, via visual displays such as a driver display or a head-up display.

[0060] Fig. Figure 2 shows the wildlife protection device 10 in a schematic representation with more details.

[0061] The controller 22 has a computing unit 22.1 and a memory 22.2, wherein a computer program product is stored in the memory 22.2 which, when executed, is in Fig. 3 procedures described in more detail.

[0062] The control unit 22 is connected on the one hand to the camera 12 and a sensor network 26, and on the other hand to the loudspeaker arrangement 16 and the passenger cabin loudspeaker 20.

[0063] Camera 12 is a high-resolution camera oriented in the direction of travel X of the vehicle 2 and covers the area in front of the vehicle 2, including the collision area 24. Camera 22 is designed to deliver high-quality images under a wide range of lighting conditions, i.e., both in bright sunlight and in low light conditions.

[0064] The sensor network 26 can have different sensors, which may differ depending on the specific embodiment, but can basically include sensors such as additional cameras, lidar, infrared cameras and / or radar.

[0065] The additional sensor network 26 can be used to increase the coverage area of ​​the wildlife protection device X, e.g. to enable 360° all-round monitoring of the motor vehicle 2.

[0066] The data received by means of the camera 12 and the sensor network 26 are first manipulated by the computing unit 22.1 in such a way that computer vision methods can be used for data analysis, e.g. by normalizing and possibly converting the data.

[0067] The artificial intelligence implemented in the computer program can include a Convolutional Neural Network (CNN) that can process the potentially multimodal data from the sensor network 26. In this way, the presence of at least one wild animal 8, as well as the species or genus of the corresponding wild animal 8, can be detected with high reliability.

[0068] By consecutively processing data over time, the movement trajectory y of the wild animal 8 can also be determined using the computing unit 22.1, and thus the risk of the wild animal 8 entering the collision area 24 can be estimated. The risk assessment may also be favorable, so that under certain circumstances no warning tone needs to be emitted to the wild animal 8.

[0069] Furthermore, in some embodiments, the control unit 22 can take the driving state of the vehicle 2 into account. Depending on the remaining reaction time, the control unit 22 can decide whether this reaction time is sufficient to allow the driver to decide whether a warning tone should be emitted and / or whether an evasive maneuver is necessary. If there is insufficient time, for example, because the wild animal 8 appears at a very short distance in front of the vehicle 2 or because the vehicle 2 is traveling at a higher speed, all measures, including, if necessary, evasive measures by intervening in the ESP of the vehicle 2, can be carried out automatically.

[0070] Fig. Figure 3 shows a flowchart of the procedure.

[0071] The procedure is essentially two-stage. After the procedure is started, for example, together with the starting of the vehicle 2, the wildlife protection device 10 is activated. Then, while driving, the camera 12 and the sensor network 26 continuously monitor whether a wild animal 8 is present in the corresponding detection area. This is repeated until a wild animal 8 is detected.

[0072] Following this very fast and computationally efficient processing step, at least one wild animal 8 is identified using artificial intelligence methods.

[0073] A suitable warning signal is then selected from a database.

[0074] If the risk assessment is high enough, a corresponding warning signal is emitted immediately via the loudspeaker arrangement 16 and an interior warning is issued simultaneously.

[0075] Using the corresponding camera 12 and the sensor network 26, it is then monitored whether the wild animal 8 has been driven out of the collision area or the vicinity of the collision area 26, with the warning signal continuing to be emitted if this is not yet the case.

[0076] The warning signal can also be varied accordingly in this case, in order to find a suitable, alternative warning signal that will prompt the wild animal 8 to turn back.

[0077] If the wild animal 8 has been successfully driven away, the procedure will be terminated or resumed from the beginning.

[0078] Although the invention has been further illustrated and explained in detail by means of preferred embodiments, the invention is not limited by the disclosed examples, and other variations can be derived from them by a person skilled in the art without departing from the scope of protection of the invention. It is therefore clear that a multitude of possible variations exist. It is also clear that the embodiments mentioned as examples are truly only examples and are not to be understood in any way as limiting, for example, the scope of protection, the possible applications, or the configuration of the invention.Rather, the preceding description and the description of the figures enable the person skilled in the art to implement the exemplary embodiments in concrete terms, whereby the person skilled in the art, with knowledge of the disclosed inventive concept, can make various changes, for example with regard to the function or the arrangement of individual elements mentioned in an exemplary embodiment, without leaving the scope of protection defined by the claims and their legal equivalents, such as a further explanation in the description. Reference symbol list 2 motor vehicles 4th Street 6 Forest 8 Wild animal 10 Wildlife protection device 12 Camera 14 Detection area 16 speaker arrangement 18 beam angles 20 passenger cabin loudspeakers 22 Control 22.1 Computing unit 22.2 Storage 24 Collision area 26 Sensor x direction of travel y movement trajectory

Claims

[1] Method for reducing the risk of wildlife collisions, wherein a motor vehicle (2) is equipped with a wildlife protection device (10) comprising at least one detection arrangement (12, 26) for detecting wild animals (8) in an area (14) located in the direction of travel (x) in front of the motor vehicle (2) and at least one warning device (16), with the detection arrangement (12, 26) detecting at least one wild animal (8) and by means of the warning device (16) emitting at least one warning signal in order to drive the at least one wild animal (8) away from a possible collision area (24) with the motor vehicle (2) or before it enters the possible collision area (24), wherein the wildlife protection device (10) detects the presence of at least one wild animal (8) in the possible collision area (24) in a first step by means of a control device (22) on the basis of signals from the detection arrangement (12, 26),In a second step, upon detection of at least one wild animal (8), the species or genus or species or genera of wild animals (8) in the possible collision area (24) are determined and the warning device (16) is caused to emit at least one warning signal tailored to the species or genus of the at least one wild animal (8), . characterized by , that the species or genus or species or genera of wild animals (8) is determined using a machine vision method, wherein the species or genus or species or genera of wild animals (8) is determined using an algorithm constructed according to artificial intelligence methods using Convolutional Neural Networks (CNNs) and Transformer Architectures, wherein the localization is designed to be continuous in order to predetermine an appropriate movement trajectory for each identified wild animal (8). [2] Method according to claim 1, characterized by, that if at least one wild animal (8) is present a warning is issued in a passenger cabin of the motor vehicle (2). [3] Method according to any of the preceding claims, characterized by , that the warning signal is varied in real time. [4] Computer program product, comprising a computer-readable storage medium (22.2) on which instructions are embedded which, when executed by at least one computing unit (22.1), cause the at least computing unit (22.1) to be configured to execute the method according to any of the preceding claims. [5] Motor vehicle with a computer program product according to claim 4. [6] Motor vehicle with a wildlife protection device (10), comprising at least a detection arrangement (12, 26) for detecting wild animals (8) in an area (14) located in the direction of travel (x) in front of the motor vehicle (2) and at least a warning device (16), as well as a control device (22), characterized by , that the wildlife protection facility (10) is equipped to carry out the method according to one of claims 1 to 3. [7] Motor vehicle (2) according to claim 6, characterized by , that the detection arrangement (12, 26) includes at least a front camera (12), infrared camera, lidar and / or radar. [8] Motor vehicle (2) according to claim 6 or 7, characterized by that the warning device has at least one loudspeaker (16).

Images