PEDESTRIAN WARNING SYSTEM

Abstract

A pedestrian warning system comprises a detection system with at least one imaging sensor connected to a controller. The controller includes a processor and memory. The memory stores a pedestrian detection module and a pedestrian warning module. The pedestrian detection module contains instructions configured to instruct the processor to identify a pedestrian in an output from the at least one imaging sensor using image analysis, and instructions configured to instruct the processor to respond to a detected pedestrian by determining a reduced sensory state of the pedestrian. Based on an output from the at least one imaging sensor and a connection to at least one external data source, the system generates a model of the surrounding environment.The pedestrian warning module instructs the processor to generate a pedestrian warning that differs from the model of the surrounding environment. An external warning system implements the pedestrian warning.

Description

INTRODUCTION

[0001] The present disclosure relates to vehicles and, in particular, to vehicle systems for warning unsuspecting and sensorily impaired pedestrians of the presence of a vehicle.

[0002] Internal combustion engine vehicles (ICE) produce numerous distinct sensory outputs inherent to their operation, alerting pedestrians to their presence. These include both the engine noise resulting from the operation of the ICE and the accompanying olfactory effect of the exhaust fumes.

[0003] Due to the way electric vehicles (EVs) operate, they inherently lack at least some sensory outputs. As a result, the operation of EVs can exacerbate pedestrians' inattention or disability-related lack of perception of the EV.

[0004] Furthermore, environmental factors such as intense sunlight, strong odors, and noisy environments can overwhelm existing sensory input, regardless of whether the vehicle is an ICE, an EV, or a hybrid ICE / EV. This, in turn, exacerbates any inattention or disability-related lack of perception of the vehicle by pedestrians.

[0005] Accordingly, it is desirable to provide a system and a process to alert pedestrians to the presence of a vehicle. SUMMARY

[0006] According to an exemplary embodiment, a pedestrian warning system comprises a detection system with at least one imaging sensor connected to a controller. The controller includes a processor and memory. The memory stores a pedestrian detection module and a pedestrian warning module. The pedestrian detection module contains instructions configured to cause the processor to identify a pedestrian in an output from the at least one imaging sensor using image analysis, and instructions configured to cause the processor to respond to an identified pedestrian by determining a reduced sensory state of the pedestrian. Based on an output from the at least one imaging sensor and a connection to at least one external data source, the system generates a model of the surrounding environment.The pedestrian warning module determines at least one warning type based on the pedestrian's reduced sensory state and initial visual and / or audio data. The warning type includes an audio warning, a visual warning, an olfactory warning, and a tactile air impact warning. The pedestrian warning module causes the processor to generate a pedestrian warning of the at least one warning type, which is distinct from the model of the surrounding environment. An external warning system implements the pedestrian warning outside the vehicle. The external warning system includes at least one audio warning system, one visual warning system, one olfactory warning system, and one tactile air impact warning system.

[0007] In addition to one or more of the features described herein, the external warning system includes the audio warning system and the visual warning system, wherein the visual warning system includes a projector, the audio warning system includes a loudspeaker, and wherein causing the warning system to implement the pedestrian warning includes activating the audio warning system and / or the visual warning system.

[0008] In addition to one or more of the features described here, the audio warning system has an audio output that is adjustable at least in amplitude and frequency.

[0009] In addition to one or more of the features described here, the visual warning system has a visual output that is adjustable at least in color and brightness.

[0010] In addition to one or more of the features described herein, the external warning system includes the olfactory warning system, wherein the olfactory warning system includes a blower and the blower has an olfactory output that is adjustable at least in terms of scent.

[0011] In addition to one or more of the features described herein, the external warning system includes the tactile air warning system, the tactile air warning system including an airflow generator.

[0012] In addition to one or more of the features described herein, determining the reduced sensory state of the pedestrian includes identifying either a distracted state of the pedestrian or a detectable impairment of the pedestrian, and wherein the at least one warning type includes an audio warning in response to the pedestrian being in a distracted state.

[0013] In addition to one or more of the features described herein, determining the pedestrian's reduced sensory state includes identifying either a distracted state of the pedestrian or a detectable impairment of the pedestrian, and wherein the at least one warning type in response to the identification of a hearing impairment of the pedestrian includes a visual warning.

[0014] In addition to one or more of the features described herein, determining the reduced sensory state of the pedestrian includes identifying a distracted state of the pedestrian or a detectable impairment of the pedestrian, and wherein the at least one warning type in response to the identification of a visual impairment of the pedestrian includes an audio warning.

[0015] In addition to one or more of the features described herein, determining the reduced sensory state of the pedestrian includes identifying a distracted state of the pedestrian or a detectable impairment of the pedestrian, including identifying a detectable impairment with an impairment above a minimum threshold, and wherein the at least one warning type omits a corresponding warning of the audio warning, the visual warning, the olfactory warning, and the tactile air impact warning.

[0016] In addition to one or more of the features described herein, the detection system includes at least one distance measurement sensor, wherein the at least one distance measurement sensor includes at least one of a light and distance measurement sensor (LiDAR sensor) and one of a radio and distance measurement sensor (RADAR sensor), and the generation of the pedestrian warning is based at least partially on a distance output received by the processor from the at least one distance measurement sensor.

[0017] In addition to one or more of the features described here, generating the pedestrian warning of the specified at least one warning type includes providing the model of the surrounding environment and a set of reward / punishment weights for a weighted machine learning algorithm.

[0018] In addition to one or more of the features described here, the set of reward / punishment weights is received from a remote computing system.

[0019] In addition to one or more of the features described herein, the pedestrian warning module is further configured to cause the processor to isolate a pedestrian's response to the pedestrian warning and to provide the pedestrian's response to the remote computing system.

[0020] In addition to one or more of the features described here, generating the pedestrian warning of at least one warning type includes applying the model of the surrounding environment to a rule-based warning model.

[0021] In addition to one or more of the features described here, the olfactory blower includes a scent reservoir and a pressurized gas section, the scent reservoir containing at least one fragrant chemical.

[0022] In addition to one or more of the features described herein, the scent storage unit stores several different fragrant chemicals, and wherein generating a pedestrian warning of at least one warning type comprises selecting one of the different fragrant chemicals from the several different fragrant chemicals and a combination of different fragrant chemicals from the several different fragrant chemicals, and wherein implementing the pedestrian warning comprises distributing the selected different fragrant chemical from the several different fragrant chemicals and the combination of different fragrant chemicals from the several different fragrant chemicals using pressurized gas from the pressurized gas section.

[0023] In addition to one or more of the features described here, the model of the surrounding environment includes a time of day, a weather condition, an ambient sound profile, an ambient light profile, a vehicle speed, and a vehicle direction of travel.

[0024] According to a further exemplary embodiment, a method for warning a pedestrian of the presence of a vehicle. The method comprises receiving a set of images from a vehicle sensor system at a controller and identifying a pedestrian in the set of images using image analysis performed by the controller, determining a reduced sensory state of the pedestrian using the controller, receiving a set of environmental conditions from the vehicle sensor system at the controller and generating a model of the surrounding environment, and determining at least one warning type corresponding to the reduced sensory state, wherein the at least one warning type includes at least one audio warning, one visual warning, one olfactory warning, and one tactile air impact warning.Generating a pedestrian warning based on at least one warning type and a model of the surrounding environment using a pedestrian warning module, wherein the pedestrian warning is distinct from the surrounding environment defined by the model of the surrounding environment, and implementing the pedestrian warning using at least one external warning system of the vehicle.

[0025] In addition to one or more of the features described herein, the pedestrian warning module is a weighted machine learning algorithm, and a set of reward / punishment weights for the weighted machine learning algorithm is recovered from a remote computing system.

[0026] The above features and advantages and other features and advantages of the disclosure are easily evident from the following detailed description when considered in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Further features, advantages and details appear in the following detailed description only as examples, the detailed description referring to the drawings; they show: Fig. 1 a schematic representation of a vehicle which includes systems for warning a pedestrian of the presence of the vehicle; Fig. 2 a process for warning pedestrians of the presence of a vehicle, which includes acoustic, visual and olfactory warnings; Fig. 3. A process for generating an audio warning of the presence of a vehicle, according to an example; Fig. 4. A process for generating a visual warning about a vehicle, according to an example; Fig. 5. An isometric view of a vehicle generating a visual warning according to an example; Fig. 6. A process for generating an olfactory warning of a vehicle according to an example; Fig. 7 a schematic view of a vehicle providing an olfactory warning according to an example; and Fig. 8 a structure of a vehicle bumper containing odor blowers according to an example. DETAILED DESCRIPTION

[0028] The following description is merely exemplary and is not intended to limit the present disclosure, its application, or use. It should be recognized that throughout the drawings, corresponding reference numerals denote identical or equivalent parts and features. As the term "module" is used herein, it refers to a processing circuit arrangement that may include an application-specific integrated circuit (ASIC), an electronic circuit, a processor (shared, dedicated, or grouped), and memory (e.g., read / write memory, solid-state storage, disk drives, or any other computer-readable non-transient data storage medium) that executes one or more software or firmware programs, a combinational logic circuit, and / or other suitable components that provide the described functionality.

[0029] As the term controller is used here, it can refer to a single dedicated controller containing memory and a processor configured to implement a control scheme; a grouping of processors and memories that are interconnected and configured to work cooperatively to implement a control system; a remote processing structure, such as a cloud computing system, configured to implement a control scheme; and any similarly configured processor and memory configuration for implementing a particular control scheme.

[0030] As used here, the term generative artificial intelligence (AI) refers to artificial computer models that can generate original content, including text, images, audio, code, and the like, in response to a given prompt. Generative AI models use weighted machine learning to correlate sets of input data with a desired output. The generative AI uses a training dataset to learn the correlations using neural transformer networks, with the training dataset typically being a significantly large dataset. As used in this example, weighted machine learning employs a machine learning-based model to correlate the received environmental parameters with a distinct desired alert type.Weights in the form of rewards and penalties are applied to each received parameter and determine how much influence that parameter has on the specific standout output.

[0031] As used here, a reduced sensory state is a condition in which a person's ability to perceive a vehicle using one or more senses is diminished compared to an average person. The reduced sensory state may result from inattention due to distraction, a physical disability, a material obstacle (e.g., headphones), or a similar factor.

[0032] In a general example, a vehicle incorporates a pedestrian detection system capable of identifying pedestrians who exhibit a degree of inattention to their surroundings, whether due to a lack of perception or a disability. The vehicle further incorporates several systems and structures configured to alert identified pedestrians to the vehicle's presence through audible, visual, and olfactory warnings.

[0033] Regarding the general example, it illustrates Fig. 1. A vehicle 10 containing a controller 20 configured to identify pedestrians using a pedestrian detection module 22 and to determine when the pedestrians are inattentive. The communication signals to and from the controller 20 can be wireless or wired using any conventional communication system, the illustration of the communication signal connections being omitted for the sake of visual clarity. Fig. 1 is omitted.

[0034] The controller 20 further includes a pedestrian warning module 24, which is configured to activate at least one fan 32, one projector 34 (e.g., a light and / or image projector), and one loudspeaker 36 (collectively referred to as external warning systems 32, 34, 36). As used herein, the projector 34 and all accompanying components for operating the projector 34 are referred to as a visual warning system; the fan 32 and all accompanying components for producing an olfactory warning are referred to as an olfactory warning system; the fan 32 and all accompanying components for producing a tactile air warning are referred to as a tactile air warning system; and the loudspeaker 36 and all accompanying components for producing an audio warning are referred to as an audio warning system.

[0035] Each speaker 36 can output an adjustable amplitude and frequency. Similarly, each projector 34 can project an adjustable color and brightness, including the simultaneous projection of multiple colors, with each olfactory blower being able to distribute several different chemicals or combinations of chemicals at different pressures.

[0036] Several fans 32, projectors 34, and loudspeakers 36 are distributed throughout the vehicle 10, with each fan 32, projector 34, and loudspeaker 36 being controllably coupled to the controller 20. According to some examples, the external warning systems 32, 34, 36 can be directional, with each instance of a given external warning system 32, 34, 36 being able to direct a corresponding output type toward a specific zone adjacent to the vehicle 10.

[0037] In addition to the external warning systems 32, 34, 36, the vehicle 10 includes at least one audio sensor 42, such as a microphone, which communicates with the controller 20, and at least one light sensor 44 which communicates with the controller 20. The audio sensor 42 and the light sensor 44 are each configured to detect the environmental conditions outside the vehicle 10 and to provide the detected condition profiles to the controller 20. The combination of sensors that provide the controller 20 with information about the external environment is collectively referred to as a vehicle sensor system or detection system. According to some examples, the vehicle sensor system may also include sensors for detecting both vehicle features and internal aspects of the vehicle 10 as well as an external environment.

[0038] The vehicle 10 can include a global navigation satellite system (GNSS) 52 and a wireless link 50 that connects the controller 20, either directly or indirectly via cloud computing networks, to remote networked information repositories. In one example, the GNSS 52 uses a global positioning system (GPS) architecture. According to alternative examples, any alternative GNSS architecture can be used with the same effect. The wireless link 50 enables the controller 20 to query a wide information network, such as the internet, and / or a set of specific remote databases for current status information about the vehicle 10's location.

[0039] A set of imaging devices 60 is arranged around the vehicle 10. According to one example, the imaging devices 60 include video cameras. Each imaging device 60 defines a corresponding field of view 62 and provides the generated images to the controller 20. According to some examples, the imaging devices 60 can also be provided as an assembly with one or more distance measurement sensors, such as a light and distance measurement sensor (LiDAR sensor) and / or a radio and distance measurement sensor (RADAR sensor).

[0040] The controller 20 applies the object and pedestrian detection processes of the pedestrian detection module 22 to the outputs of the imaging devices 60 and identifies both the pedestrians outside the vehicle 10 and the pedestrian's position relative to the vehicle 10. The pedestrian detection module 22 operates using any known object and / or pedestrian detection processes. If the imaging devices 60 are provided with distance sensors as an assembly, or if the vehicle 10 contains distance sensors located elsewhere in the vehicle 10, the output of the distance sensors can be used to further improve the detection of the relative position of the detected pedestrians and the vehicle 10.According to other examples, the controller 20 can contain software that can determine the distances between the vehicle 10 and the pedestrians using only an image analysis at the output of the imaging devices 60.

[0041] Furthermore regarding Fig. 1 illustrates Fig. 2. A process 200, which is operated by the pedestrian warning module 24 based on an output from the pedestrian detection module 22 according to an example. The process 200 begins with a pedestrian detection check 202 on an output of the pedestrian detection module 22. If no pedestrian is detected near the vehicle 10, the pedestrian warning process 200 takes no action at a stop 204.

[0042] When a pedestrian is detected, the pedestrian warning process 200 proceeds to determine, using a distraction / headphone presence check 206, whether the pedestrian is distracted or otherwise inattentive. Check 206 is performed using image analysis of the pedestrian to identify whether the pedestrian is distracted and / or otherwise concealing their hearing. For example, check 206 may use image analysis to determine that the pedestrian is not facing the vehicle, is wearing headphones or other ear protection, and / or is using a smart device or other technology that could distract the pedestrian. The image analysis is performed on images generated by the imaging devices 60.According to some alternative examples, where the vehicle 10 is vehicle-to-vehicle (V2V) or vehicle-to-other (V2X) capable, the image analysis can include images of the pedestrian generated by imaging sensors outside the vehicle 10 that are connected to the vehicle 10. For example, the image sensors could include still cameras, video cameras, or any similar sensor capable of capturing an image.

[0043] If the test 206 determines that the pedestrian is not distracted and their hearing is not obstructed, the process 200 initiates an audio warning 208. The audio warning 208 uses several linear regression models, such as slope coefficients and intercepts, to model the ambient noise around the vehicle 10 and the vehicle noise from the vehicle 10 received by the audio sensor 42, as shown in an example. Additionally, reward / penalty weights for the operating modes of the linear regression are received from a remote computer system 212. The reward / penalty weights provided by the remote computer system 212 are generated based on previously confirmed successful audio warnings and are iterated and improved by the remote computer system 212 using a post-processing system 224 as more successful audio warnings are provided to the remote computer system 212.

[0044] The modeled ambient and vehicle noises, along with the reward / penalty weights, are used in an audio warning generation step 210 to create a warning sound that stands out from the ambient and vehicle noises and is therefore more noticeable to the pedestrian. The audio warning is played through one or more speakers 36 at a volume configured to exceed the volume of the ambient noise. According to one example, the standout sound is generated using a generative AI model. According to another example, the standout sound is generated using rule-based generation models or statistically based models.The generated audio warning and the one or more images of the pedestrian's reaction are provided back to the remote computer system 212, where the reaction is stored and used to improve future reward / punishment weights.

[0045] According to the examples in which directional speakers 36 or multiple speakers 36 are included in the vehicle 10, the controller 20 can cause only the one or more speakers 36 that are directed towards the pedestrian to emit the audio warning, thus providing a more targeted warning.

[0046] If check 206 determines that the pedestrian is distracted, process 200, in a disability detection check 214, determines whether a detectable disability is present. If check 206 does not detect a disability, process 200, in a multiple alert step 216, generates multiple alerts. As used here, a detectable disability includes any disability, including visual identifiers, that are visible in an image of the pedestrian. For example, visual identifiers might include a white cane, a guide dog, external hearing aids, or a similar assistive device.

[0047] The multiple alerts include audio alerts, generated as described in steps 208 and 210, and either a visual or an olfactory alert. As with the audio alert, a model of the ambient light or surrounding olfactory environment, along with reward / punishment alert data from the remote computing system 212, is used in step 218 to generate contrasting visual or olfactory alerts in the same general manner as the generation of audio alerts described in steps 208 and 210.

[0048] If the impairment check 214 identifies a detectable impairment and the detectable impairment exceeds a minimum threshold, process 200, in step 220, generates all available types of warnings, using the same general linear regression model generation processes (step 222) described for audio warnings. The minimum threshold is set, according to some examples, at a level where one or more warning types may be ineffective, and as much redundancy as possible is desired when issuing warnings.

[0049] In any case, a corresponding external warning system 32, 34, 36 generates a warning that is perceptible to pedestrians outside the vehicle 10 and stands out from the corresponding ambient levels of the warning type.

[0050] According to some examples, after detecting the presence of a detectable disability and before generating the warning, Process 200 may include another type of disability check. According to these examples, the disability type check may omit warning types that would not be of use to a pedestrian with the detected disability. For example, if the detected disability is a visual impairment or blindness, the disability type check may cause Process 200 to omit a visual warning and rely solely on audio and olfactory warnings. As visual impairment is used here, it includes disabilities and any other characteristics that affect the user's vision. For example, visual impairment may include visual impairment due to biological disabilities and visual impairment due to obstacles, such as...Included are augmented reality glasses, protective screens, and the like.

[0051] If the detected disability is a hearing impairment or deafness, the disability type assessment, similar to Process 200, may result in the omission of an audio alert and reliance solely on visual, olfactory, and / or tactile air alerts. As used here, a hearing impairment includes disabilities and any other features that affect the user's hearing. For example, a hearing impairment may include hearing loss due to biological disabilities and hearing loss due to obstructions such as headphones, earplugs, earmuffs, and the like.

[0052] Furthermore regarding the Fig. 1 and Fig. 2 illustrates Fig. 3 An exemplary audio warning generation process 300 using a generative AI model to generate a distinctive audio warning. Initially, a pedestrian is detected in a pedestrian detection step 310. The pedestrian detection includes a distance from the vehicle 10 to the pedestrian and provides the distance to a generative AI model 320. After the pedestrian is detected, the process 300 initiates a step 330 of continuous acoustic monitoring of the environment.

[0053] Step 330 of the continuous audio monitoring process monitors the ambient sound received by microphone 42, the location data received by GNSS 52 or any other available location source, and a time of day 332. Step 330 of the continuous audio monitoring process combines the monitored data into an output model 334 of the surrounding acoustic environment and provides the output model 334 to the generative AI model 320.

[0054] The generative AI model 320 communicates with the remote computer system 212 to receive reward / punishment weights, which define the reward and punishment weights for the components of the output model 334, from the remote computer system 212, and generates an audio warning output 340 that stands out from the surrounding environment.

[0055] The distinctive audio warning 340 contains a unique sound signature that differs from the ambient sounds. In some cases where the vehicle 10 contains multiple and / or directional loudspeakers 36, the audio warning signal 340 can be projected towards and oriented in the direction of the pedestrian.

[0056] According to further examples, the contrasting sound is configured to be sharply defined and emitted suddenly, startling the pedestrian and drawing their attention to the vehicle 10. During the audio warning 340, the pedestrian's reaction time and the length of the audio artifacts are determined based on the images from the imaging sensors 26 and the audio feedback from the audio sensor 42 and provided back to the remote computer system 212. The remote computer system 212 then uses the reaction time and the length of the audio artifacts to refine the reward / punishment weights of the generative AI structure.

[0057] According to alternative implementations, the generative AI model 320 can be replaced by a rule-based model using predefined reward / punishment weights applied to a sound signal from information obtained from monitoring the surrounding environment, with the rules defining specific standing-out signals and responses to potential environmental sound profiles.

[0058] Furthermore regarding the Fig. 1-3 illustrates Fig. 4 a process 400 to generate visual warnings 502 and illustrates Fig. Five exemplary visual warnings 502, generated by the projectors 34 arranged on the vehicle 10, are described in an example. Initially, in an image sensor step 402, the image sensors 26 generate images of the visual environment surrounding the vehicle 10. During image sensor step 402, the images are analyzed to identify any pedestrians using existing pedestrian and / or object detection processes. As part of the visual environment, a distance between the vehicle 10 and the pedestrian is determined using the images, the distance measurement sensors arranged in an assembly within the vehicle, or a combination thereof.

[0059] The detected visual environment is then provided to an obstruction detector 404. The obstruction detector 404 applies image analysis processes to the visual environment, and in particular to the identified pedestrian, and identifies any detectable obstructions that may impair the pedestrian's attention or ability to see the vehicle 10. The obstruction detector 404 then forwards the visual environment, which includes the distance to the pedestrian and any detected obstructions, to a light environment model 406.

[0060] The light environment model 406 also receives input from a vehicle obstacle detection system 408, a weather detection system 410, and an ambient light value 412. The ambient light value 412 can be obtained from a lumen sensor on the vehicle 10, from a database that correlates the time of day and the vehicle's location with an expected ambient light condition, or a combination thereof.

[0061] According to further alternatives, additional factors influencing the ambient light can be provided to the light environment model 406, using the combination of inputs to construct a model of the light environment surrounding the vehicle 10.

[0062] Once the light environment model 406 has been established, process 400 provides the light environment model 406 to a warning generation module in a warning generation step 414. The warning generation step 414 uses a generative AI module to generate a distinctive visual warning 502. Fig. 5, to generate, which can be projected by the projector 34 onto the environment surrounding the vehicle 10.

[0063] The contrasting visual warning is based on the light environment model 406 and a stored set of reward / punishment weights from the remote computer system 212, and provides a strong contrast to the ambient light around the vehicle 10. According to one example, laser light or a similar system is used to project a striking pattern in pronounced colors. According to another example, the striking pattern may include directional indicators, movement, or other features designed to draw the pedestrian's attention to the presence of the vehicle 10. Once the visual warning 502 has been designed, it is projected in a visual warning projection step 416.

[0064] According to an alternative example, the visual warning 502 can be generated using both rule-based processes that contrast the distance and the red-green-blue (RGB) colors of the ambient light, and rule-based processes that identify clear areas near the pedestrian where a visual warning 502 is most likely to be visible. In such a case, the reward / penalty weights are not recovered from the remote computer system 212; instead, the predefined rules are used.

[0065] Furthermore regarding the Fig. 1-5 illustrate Fig. 6 an exemplary process 600 for generating a (in Fig. 7 shown) olfactory warning 602. Fig. Figure 7 illustrates a schematic view of vehicle 10 from top to bottom, projecting two possible olfactory warnings, and Fig. Figure 8 illustrates the blowers 32 which are positioned in a section of the bumper 802 of a vehicle, such as vehicle 10.

[0066] Initially, in a context creation step 604, process 600 creates a contextual model of the air currents and flows in the surrounding environment. According to one example, the contextual model considers information about the wind speed provided by a weather service, the direction of travel of vehicle 10, the location of vehicle 10, the speed of vehicle 10, the orientation of vehicle 10 relative to the detected pedestrian, the ambient air pressure, and the distance to the pedestrian. The combined factors represent a general airflow of the surrounding environment and are provided to a generative AI model in step 606 to generate an olfactory alert. According to alternative examples, additional factors influencing the general airflows can be included in context creation step 604 if they are available to controller 20.

[0067] Step 606, generating an olfactory warning, uses a generative AI model to create an optimized regression model for estimating a desired air pressure and odor warning from the fans 32. The odor warning includes an odor selected from a database of warning odors available to be generated by one of the multiple fans 32. For example, the warning odors could include chemical odors, smoke odors, citrus odors, fruit odors, herbal odors, and the like, with the specific odor chosen to stand out from the surrounding environment. Additionally, the olfactory warning includes a desired airflow orientation and pressure to warn the pedestrian.

[0068] Once a pressure and an odor have been generated by the generative AI model in step 606 to create an olfactory alert, the alert information is provided to the blower 32. For example, the blower 32 contains a storage 804 of essential oils and / or other highly fragrant fluids that can safely produce an odor using an atomizer and a compressor or a compressed air section 806. The olfactory blower receives the olfactory alert profile and outputs the olfactory alert 602 in an alert output step 608, see [reference]. Fig. 6.

[0069] As with audio and visual alerts, olfactory alerts can be generated using a rule-based process instead of the generative AI model, according to some examples.

[0070] According to some alternative examples, the blower 32 may include a tactile airflow feature that can generate an airflow oriented towards the person to be warned, such as a gust of wind. The tactile airflow may be present instead of or in addition to the odor features and is referred to as a tactile air warning. According to one such example, the sections of the blower 32 used to generate the tactile airflow are collectively referred to as an airflow generator.

[0071] According to some alternative examples, a single model of the surrounding environment can be generated from a combination of sensor types and data sources. This single model provides the data required to generate all types of alerts.

[0072] The terms "one" and "an" do not denote a quantity limitation, but rather indicate the presence of at least one of the designated element. The term "or" means "and / or" unless the context clearly indicates otherwise. References throughout the application text to "an aspect" mean that a particular element (e.g., a feature, structure, step, or property) described in connection with that aspect is contained in at least one aspect described therein and may or may not be present in other aspects. Furthermore, it should be understood that the described elements in the various aspects can be combined in any suitable way.

[0073] When an element, such as a layer, film, area, or substrate, is described as being "on" another element, it may be located directly on top of that element, or there may be intervening elements. Conversely, when an element is described as being "directly on" another element, there are no intervening elements.

[0074] Unless otherwise specified herein, all testing standards are the most recent standard in force since the date of filing of this application or, if priority is claimed, the date of filing of the earliest priority application in which the testing standard appears.

[0075] Unless otherwise defined, the technical and scientific terms used herein have the same meanings as generally understood by a person skilled in the field to which this disclosure belongs.

[0076] While the above disclosure has described exemplary embodiments, those skilled in the art recognize that various modifications can be made and elements can be replaced by their equivalents without altering its scope of protection. Furthermore, many modifications can be made to adapt a particular situation or material to the teachings of the disclosure without altering its essential scope of protection. Therefore, it is intended that the present disclosure is not limited to the specific embodiments disclosed, but includes all embodiments that fall within its scope of protection.

Images