Method for establishing road guidance instructions

Abstract

The invention relates to a method for establishing voice guidance instructions for a person, these voice guidance instructions being formulated in the most natural manner possible, the method comprising the steps of: - determining a path to be followed by the person by means of a computer, - acquiring at least one image of the environment of the person by means of an image acquisition device, - processing said image to detect at least one object therein and to characterize this object, and - establishing voice guidance instructions, thereby providing a piece of information to the user to perform a maneuver to follow the path. According to the invention, a step for determining the complexity of the maneuver is provided, and in the establishing step, the voice guidance instructions are formulated using the indications deduced from the characterization of the object only if the complexity of the maneuver is greater than a first threshold.

Description

Technical Field

[0001] This invention relates generally to guidance systems, and more specifically (but not only) to guidance systems integrated into motor vehicles.

[0002] More specifically, the present invention relates to a method for generating voice-guided instructions for an individual, the method comprising the following steps:

[0003] -Use computers to determine the route to take.

[0004] - Obtain at least one image of the individual's environment.

[0005] - Process the image to detect at least one object present in the environment and characterize the object, and

[0006] - Establish and send voice guidance instructions, thereby informing the individual how to maneuver to follow the stated route.

[0007] The present invention also relates to a device for generating voice guidance instructions.

[0008] This invention is preferably applicable to situations where the individual is a motor vehicle driver. Background Technology

[0009] Satellite navigation and positioning systems are a particularly effective and popular way for drivers of vehicles to determine the route they want to take.

[0010] The advantages over paper maps are significant, especially since an individual's location is updated in real time. Most navigation systems display directions on the map corresponding to a schematic representation of the vehicle's surroundings.

[0011] Typically, such navigation systems are designed to send route guidance instructions to the driver in two different ways: visually via a display screen and verbally via the vehicle's speakers.

[0012] Because maps distort the distances and angles between roads, the visual representation of the maneuver to be performed on the display screen is not always clear. Drivers sometimes have difficulty correctly understanding the maneuver to be performed.

[0013] Navigation system instructions (due to their visual representation and associated voice commands) are not always consistent, leading to driver misunderstandings about the maneuvers to be performed and their perception of road infrastructure, its configuration, and the driving environment.

[0014] It is particularly important to note that currently, voice guidance commands are systematically generated based on the distance between the vehicle and the next maneuver to be performed, much like a route. Figure 1 Sample.

[0015] Therefore, these instructions often become unsuitable for the environment, repetitive and predictable, or even inconsistent with the driver's expectations.

[0016] Another drawback of known navigation systems is that the instructions are generated without any connection to each other. These instructions are independent of each other. The content of each instruction does not take into account the information conveyed in the preceding instructions or the information to be conveyed in the next instruction depending on the maneuver the driver must perform.

[0017] In practice, due to these shortcomings, most users have stopped using the voice function of navigation systems. Summary of the Invention

[0018] To overcome the aforementioned deficiencies of the prior art, the present invention provides a method for generating voice guidance instructions as defined in the introduction, wherein a step of determining the complexity of the maneuver is specified, and wherein, in the establishment step, the voice guidance instructions are formulated using a prompt inferred from the representation of the object only when the complexity of the maneuver is higher than a first threshold.

[0019] In other words, this invention proposes to generate voice guidance instructions, the level of detail of which depends on the complexity of the relevant circumstances.

[0020] For example, if the driver is facing a simple situation such as a crossroads and there are no objects obstructing the driver's view of the crossroads, then voice guidance instructions are generated without adding any specific details.

[0021] In contrast, in more complex situations, such as when an object (like a truck) partially obscures an intersection, or when the intersection itself is more complex, voice guidance instructions are generated with additional cues to allow the driver to determine their location. These can be visual cues (e.g., trucks, signs, road infrastructure) or other types of cues, such as contextual cues (to tell the driver, for example, when they must turn).

[0022] Therefore, this invention proposes generating voice guidance instructions in a manner as similar as possible to that of a human co-pilot. In this way, the comprehensibility of these instructions is increased, and their frequency is appropriate, which reduces the cognitive burden on the driver and minimizes driver errors and hesitation.

[0023] The following are other advantageous and non-limiting features of the method according to the invention, which may be implemented individually or in any technically possible combination:

[0024] - The cue is a visual cue that is currently visible to the individual or will soon be visible to the individual, or a contextual cue that provides the individual with information about when he must perform the maneuver;

[0025] - The voice guidance instruction is constructed such that at least one action and the prompt are inserted therein;

[0026] The calculation of complexity depends on at least one of the following criteria: the vehicle's location relative to the area in which the maneuver must be performed, the type of road infrastructure present in the area in which the maneuver must be performed, the individual's perceived stress, the density of road traffic, the number of objects detected in at least a portion of the acquired images, the individual's knowledge of the road infrastructure, and the current stage of the vehicle.

[0027] - If the road infrastructure is a roundabout, the calculation of complexity depends on at least one of the following criteria: the angle between the axis from which an individual arrives at the roundabout and the axis from which the individual leaves the roundabout, the angle between at least two roads leading to the roundabout, the number of roads leading to the roundabout, the difference between the number of entrances to the roundabout and the number of exits to the roundabout, the number of lanes present in the roundabout, and the outer diameter of the roundabout.

[0028] - The cue is a visual cue visible to the individual. If the visual cue disappears from the environment visible to the individual between the moment the voice guidance instruction is sent in audible form and the moment the maneuver must be performed, then steps are specified for generating a new voice guidance instruction without using the visual cue.

[0029] - The prompt is a visual prompt visible to the individual. If, between the moment the voice guidance instruction is sent in audible form and the moment the maneuver must be performed, another visual prompt that is indistinguishable from the visual prompt used in the voice guidance instruction appears in the environment visible to the individual, then the steps for generating a new voice guidance instruction that is different from the voice guidance instruction sent in audible form are specified.

[0030] - If the complexity of the maneuver is higher than a second threshold that is strictly higher than the first threshold, then two prompts are used to generate the voice guidance command;

[0031] - The regulations specify the costs associated with the additional time or distance that would result from an error while following the route, and the level of complexity is determined based on these costs;

[0032] - It is stipulated that at least two distinct and consecutive phases shall be distinguished before the execution of the maneuver, namely, a maneuver prediction phase in which voice guidance instructions may be sent to instruct the individual on how to position himself to approach the maneuver, and a description phase in which voice guidance instructions are sent to instruct the individual on how to execute the maneuver, wherein the voice guidance instructions sent in these phases use different prompts;

[0033] - During the prediction phase, if the road to be taken or to be taken contains multiple lanes, a voice guidance command should be given.

[0034] - The prompts used in the voice guidance instructions during the pre-judgment phase only involve signs when the individual is outside the building area;

[0035] - Further specifies at least one other stage following the description stage and in which voice guidance instructions can be generated, said other stage being an accompanying stage performed concurrently with the maneuver, or a dispelling doubt stage following the maneuver and allowing the individual to be instructed on whether the maneuver was performed correctly, or a continuation stage allowing the individual to be informed of the active characteristics of the generation method;

[0036] - If the voice guidance instruction sent during the description phase does not involve any visual indicators and the vehicle decelerates, then it is stipulated that a voice guidance instruction using prompts be sent during the accompanying phase;

[0037] - In this dispelling doubts phase, it is stipulated that the voice guidance instructions will be sent using visual indicators that are different from one or more visual cues used in the one or more voice guidance instructions sent during the previous phase; and

[0038] - In this continuation phase, different voice guidance instructions are generated depending on whether the radius of curvature of the road being traveled is below or above a threshold.

[0039] According to one aspect of the invention, the step of determining a route by computer includes: retrieving static visual indicators located on the route and attributes associated with the static visual indicators from a map database, and determining attributes common to a plurality of static visual indicators. When the object is characterized as including the common attribute in the image processing step, a prompt including the common attribute is used in the establishment step to formulate the voice guidance instruction.

[0040] According to one aspect of the invention, the static visual indicator is a road sign, the common attributes of which include the name written on the road sign or the background color of the road sign, and the object is a road sign.

[0041] According to one aspect of the invention, the step of processing the image includes detecting a plurality of objects, and the generation method further includes the step of classifying the detected objects, wherein objects characterized as having the common attribute are given priority, and the objects selected for the generation step are the objects given priority.

[0042] The present invention also provides an apparatus for generating route guidance instructions suitable for an environment, the apparatus comprising:

[0043] - A device for acquiring an image of the environment.

[0044] -Map database,

[0045] -Navigation and geolocation systems

[0046] - A processing module adapted to receive and combine elements from the acquisition device, the map database, and the navigation and geolocation system, and programmed to implement the method described above.

[0047] - A voice transmitter, which is used to send instructions to the individual.

[0048] Of course, the various features, variations and embodiments of the present invention can be associated with each other in various combinations, as long as they are not incompatible or mutually exclusive. Attached Figure Description

[0049] The description given with reference to the accompanying drawings, by way of non-limiting examples, will make it easy to understand what the invention includes and how it can be practiced.

[0050] In the attached diagram:

[0051] Figure 1 This is a schematic diagram of the front of a motor vehicle equipped with a device according to the present invention for generating route guidance instructions. Detailed Implementation

[0052] Preferably, the present invention can be used in any motor vehicle (car, truck, motorcycle, etc.).

[0053] exist Figure 1 The diagram schematically shows the passenger compartment of a motor vehicle 1, which is equipped with a device for generating route guidance instructions that allows the implementation of the present invention.

[0054] The device includes means for acquiring images of the vehicle environment, a guidance unit 5, and an interface 11 for sending guidance commands to the driver of the motor vehicle 1.

[0055] For example, the image acquisition device here includes a camera 10 that faces forward of the vehicle and can acquire an image of the environment as the driver’s gaze turns toward the road.

[0056] Interface 11 includes at least one audio signal transmission system 12. Additionally, the interface includes a touchscreen 14. As a variant, it may not have a screen.

[0057] The audio signal transmission system 12 is designed to send voice instructions to an individual. It is placed within the passenger compartment at a distance from the individual so that the individual can clearly hear the transmitted voice instructions. For example, it may be a loudspeaker installed in a vehicle for, for example, listening to a radio, or actually a loudspeaker for a portable system (e.g., a telephone or external navigation device).

[0058] The guidance unit 5 includes a map database 20, a navigation and geolocation system 22, a computer memory (hereinafter referred to as "memory 24") and a computer (hereinafter referred to as "processing module 25").

[0059] Map database 20 includes two-dimensional or even three-dimensional map data, especially the location (longitude and latitude) of roads in a given space (e.g., Europe). Map database 20 also includes additional information such as the shape of buildings, the location of service stations and typical environmental locations (e.g., landmarks or rivers), and the type of road surface (paved, dirt, asphalt, etc.). The map data is designed to be readable by processing module 25.

[0060] Advantageously, the map database 20 includes static visual indicators, such as road signs and attributes associated with each road sign, such as the background color and orientation of the sign or the name written on it.

[0061] The navigation and geolocation system 22 is capable of determining the current spatial location of the vehicle and, based on data stored in the map database 20, determining the route between the vehicle 1's current location (or starting location) and the desired destination. The navigation and geolocation system 22 may be, for example, a Global Positioning System (GPS). The vehicle's current location and route are intended to be readable by the processing module 25.

[0062] The route determined by the navigation and geolocation system 22 is stored in the memory 24. The memory 24 also stores parameters, such as thresholds used in embodiments of the invention. All elements stored in the memory 24 can be read by the processing module 25. These elements will be described in further detail in this specification.

[0063] The processing module 25 is programmed to receive and combine various elements from the camera 10, map database 20, navigation and geolocation system 22, and memory 24. The processing module 25 is also programmed to send data to the speaker 12 and screen 14 to convey guidance instructions to the individual. In practice, this processing module 25 is formed using a microcontroller.

[0064] The device described above for generating boot instructions allows the following methods to be implemented.

[0065] Step S1: Determine the route

[0066] The process begins when the vehicle 1 is started. Then, in the first step, the processing module 25 determines the route that the vehicle 1 must take.

[0067] For this purpose, the current (starting) location is determined by navigation and geolocation system 22. As a variation, an individual can specify a starting location different from their current location by manually entering the target starting location. For example, an individual can enter the target starting location via screen 14.

[0068] Then, the individual specifies their desired destination. For example, the desired destination can be entered via screen 14.

[0069] Processing module 25 combines the starting location, desired destination, and map data corresponding to the relevant geographic area. Processing module 25 sends this information to navigation and geolocation system 22, which then generates a route.

[0070] The determined route is stored in memory 24.

[0071] Advantageously, processing module 25 retrieves information about road signs located on the route, as well as attributes associated with the signs, from map database 20. As will be better explained below, processing module 25 is able to determine whether multiple signs have common attributes in order to identify the common attributes and signs that include the common attributes.

[0072] A route consists of multiple segments. For example, a segment is defined as a part of a highway that extends between two consecutive intersections.

[0073] Each road segment can then be characterized by multiple parameters, such as the number of lanes it includes, one or more directions of travel it allows, the type of intersection, and the number of entrance and exit lanes at each intersection.

[0074] This route passes through numerous waypoints, such as:

[0075] - Road infrastructure (overpasses, roundabouts, intersections, etc.),

[0076] - Municipalities or localities

[0077] -Important locations (pharmacies, bakeries, shopping malls, fountains, bus shelters),

[0078] - Prominent locations (museums, historical sites, cathedrals, train stations, airports, etc.).

[0079] At this stage, before describing the rest of the method, other concepts that are useful for understanding this description may be described.

[0080] The first concept is the concept of "action". This is a single movement performed by the vehicle (e.g., "entering the right lane", "following the red car").

[0081] The concept of a "target" refers to the area to which a motor vehicle must move. It can be represented by elements of road infrastructure (traffic lights, stop signs, directional signs, highways, etc.).

[0082] The concept of a "maneuver" refers to a series of one or more actions that allow for the achievement of an "objective." A maneuver is performed to change the position of a motor vehicle relative to road infrastructure and to achieve that objective. For a given maneuver, multiple pairs consisting of an action and an objective can be cascaded. For example, a maneuver can be expressed as: "Enter the left lane and then turn left onto highway X."

[0083] The concept of a “cue” encompasses any linguistic element that may be used in voice guidance instructions, and when a cue is used, it is defined based on images captured by a camera to facilitate the vehicle driver’s understanding of the situation.

[0084] The concept of a "visual cue" (or "visual indicator") refers to static elements (roads, signs, road infrastructure, and more generally any element contained in map database 20) or dynamic elements (vehicles, pedestrians, or more generally any element that processing module 25 can detect in the images acquired by camera 10) that can be used to describe the action to be performed or when the action should be performed. These elements allow guidance instructions to be enriched by adding details to the guidance instructions (e.g., "turn right after the traffic light") or by allowing guidance instructions to be given by referring to the element (e.g., "continue along the blue building").

[0085] The concept of a "contextual cue" (or "temporal cue") refers to when a maneuver must be performed. This information indicates to the driver when the maneuver must be executed. For example, such a cue might be expressed as: "now," "immediately," "as soon as possible," "at the roundabout," or "at the traffic light."

[0086] The concept of an "attribute" in visual cues or targets refers to the characteristics of that visual cue or target that allow it to be distinguished from other indicators. It could be color, text, direction, road type, lane width, road surface type, etc.

[0087] Two types of attributes can be considered:

[0088] - Self-centered attributes, that is, attributes related to the position of motor vehicle 1 (target's location relative to the vehicle's position, the positions of various elements around the vehicle, the number of roundabout exits considering the vehicle's position, its position on the lane, etc.).

[0089] - Non-egocentric attributes, i.e., attributes of absolute positioning (the positioning of visual cues relative to one or more fixed indicators in the environment, and therefore does not change with the vehicle's position - such as the number of lanes on a multi-lane highway).

[0090] Goals are typically egocentric attributes, while visual cues are typically non-egocentric attributes.

[0091] The concept of "familiarity criterion" refers to language elements that allow for a more human touch in the formulation of guidance instructions. For example, it might include expressions such as "Stay in your lane now," "Go first," and "Then turn left." This concept also refers to language elements with safety functions (such as "Watch out for the curve!").

[0092] The concept of a “geographical indicator” will be used to refer to the location of a visual cue relative to a target, and may be expressed as: “at the end of…”, “right after…”, “in the middle”, “completely to the right”, “slightly to the right”, “opposite”, “directly in front”, “along the direction of…”, etc.

[0093] Step S2: Acquire video

[0094] Once the vehicle is started, camera 10 acquires images of the vehicle's environment and sends them to processing module 25. All or some of the images acquired by camera 10 can be processed by processing module 25 at a frequency higher than one hertz.

[0095] Processing module 25 is programmed to process these images to generate voice guidance instructions that communicate to the driver, especially regarding maneuvers he will soon need to perform. It is programmed to generate at least a portion of these instructions taking into account the environment visible to the driver.

[0096] As the vehicle travels along a route defined by the driver, the processing module 25 tracks the vehicle's progress with the help of the navigation and geolocation system 22. This allows the processing module to predict before each intersection whether the vehicle must perform a maneuver, and if so, which maneuver it must perform.

[0097] Each maneuver that the vehicle must perform is then broken down into five consecutive stages by the processing module 25. In each stage, voice guidance commands may be generated and sent if necessary.

[0098] The first stage is the anticipation stage, which allows drivers to be informed how to position their vehicles on the road in order to approach the next event (intersection, roundabout, etc.) in the best possible way.

[0099] The second stage is the description stage, which allows the driver to be informed of the maneuver he will have to perform so that he can understand it before he has to perform it.

[0100] The third stage is the accompanying stage, which is specifically designed to assist the driver during maneuvers to enhance his confidence.

[0101] The fourth stage is the dispelling of doubts stage, which allows the driver to be assured that he has performed the correct maneuver.

[0102] The fifth stage is the doubt-relief stage, which occurs when the distance or time between two maneuvers exceeds a predefined threshold. It allows the driver to be shown that the processing module 25 is still active (after a long period without guidance instructions).

[0103] Then boot instructions may be generated in each of these five stages.

[0104] These guidance instructions are generated sequentially in multiple consecutive steps, which generally include detecting visible static and moving objects in the images, identifying roads and routes in each acquired image, locating the detected objects on a map obtained from map database 20, associating the detected static objects with known static objects in map database 20, and applying statistical and / or continuity reasoning to these objects to generate voice guidance instructions.

[0105] This specification will first detail how to generate guidance instructions. The second part will explain in detail how to enhance these guidance instructions using environmental objects detected in the acquired images. Finally, the third part will provide instructions on how to accurately generate these instructions at the current stage of the vehicle's journey (prediction, description, accompaniment, etc.).

[0106] Step S3: Detect the object

[0107] The processing module 25 uses existing artificial intelligence and machine learning techniques to perform the steps of detecting moving and static objects in the image acquired by the camera 10 and characterizing the attributes of each of these objects.

[0108] It should be noted that the object is defined here as an environmental element that is potentially or actually visible to the vehicle driver.

[0109] The list of attributes depends on the type of object detected.

[0110] Therefore, as an illustrative example, if the object is a "4-wheeled motor vehicle", the properties would be as follows:

[0111] - Categories (automobiles, light trucks, medium trucks, heavy trucks, buses, etc.)

[0112] -The main color of the car body

[0113] -Direction of travel

[0114] - The parts seen by the driver (front, rear, left, right),

[0115] - Status (parked, stopped, running)

[0116] - The orientation of the object relative to the vehicle.

[0117] -Percentage visible.

[0118] Detected objects and each attribute are assigned a confidence index expressed as a percentage. The closer the confidence index is to 100%, the easier it is to use the object and attribute. Attributes or objects below a predetermined threshold are no longer considered. The higher the number of attributes a given object has, the easier it is for that object to be identified in each iteration of the scene analysis (i.e., in each image processed).

[0119] Step S4: Locate the object

[0120] In order to identify one or more objects to be used in the formulation of voice guidance instructions, the processing module 25 matches the objects detected in the processed image with data stored in the map database 20, which allows understanding of the relationship of each object relative to the environment and the driver's perception.

[0121] To this end, the processing module 25 uses data stored in the map database 20, especially the objects referenced therein (points of interest, road signs, traffic lights, bus stops, etc.) and their descriptions (name, type, color, etc.).

[0122] Next, the processing module 25 identifies roads and routes in the processed image.

[0123] Next, processing module 25 locates each detected object on the map by matching the referenced objects with the detected objects. It also locates both detected and unreferenced objects (vehicles, pedestrians, etc.) on the map.

[0124] Step S6: Classification

[0125] At this stage, processing module 25 classifies all detected objects.

[0126] The first classification technique includes using existing artificial intelligence and machine learning techniques.

[0127] The technology may include training artificial intelligence to automatically identify at least ten detected objects, focusing on the right side of the image if the vehicle must turn right, the central area if it must continue straight, and the left side if it must turn left, and then classifying these ten objects in order of importance.

[0128] The second classification technique involves performing statistical calculations to assign weights to each detected object.

[0129] The weight can be calculated based on the criteria selected from the following list:

[0130] - In a series of images, the persistence of an object within the scene (the longer an object remains visible, the greater its weight increases; if an object alternates between being visible and invisible, its weight decreases),

[0131] - The area occupied by objects in the scene and their visible percentage (trucks are more visible than cars).

[0132] - The salience of an object (the greater the difference in its color from the rest of its environment, the greater its weight; for example, a red car is more conspicuous than a gray car).

[0133] - The uniqueness of the object (using a blue vehicle that is only one is clearer than using a red vehicle among other red or orange vehicles).

[0134] - The position of the object relative to the maneuver to be performed (the closer the object, the greater its weight).

[0135] - The current stage of the vehicle (as will be further detailed in this description; depending on the stage, objects of some classes are better suited to provide guidance than others).

[0136] The third technology involves a strategy based on the continuity of visual cues along the route.

[0137] If an object's attribute has been identified by processing module 25 as an attribute common to multiple objects or as a hint that has been used to generate previous guidance instructions for the current route and / or could potentially be used to generate subsequent guidance instructions on the same route, then the object is given priority for instruction generation in step S7 as described below.

[0138] Advantageously, when the processing module 25 determines route S1, the common attributes will have already been identified by the processing module.

[0139] For example, in determining route S1, processing module 25 identifies multiple road signs bearing the name "Long Island". The common attribute identified is the name "Long Island" written on them. In video acquisition step S2, camera 10 acquires an image including the signs. In object detection step S3, processing module 25 identifies the signs in the image and detects whether they possess features of the previously identified common attributes. This attribute can be detected by deciphering the name written on the sign in the image or by obtaining a match via geolocation with signs included in map database 20, which are associated with attributes also included in map database 20. In classification step S5, processing module 25 prioritizes the symbol to ensure that it will be used when generating the next guidance instruction in step S7.

[0140] Therefore, if signs with the same name are found at multiple consecutive intersections along a route, the guidance instructions associated with these intersections will prioritize the given visual cues, thus ensuring the continuity of visual cues. This guarantees consistency and helps guide users while reducing their cognitive load.

[0141] For example, on a route that includes a first roundabout, then an intersection, and a road where you are not allowed to branch off in the middle, the following sequential guidance instructions are generated: "Turn right at the roundabout toward Long Island", "Continue straight following the Long Island sign", "Keep going toward Long Island".

[0142] Step S7: Generate instructions

[0143] Then, bootstrap instructions are generated, possibly using hints that are appropriate to the context and the maneuver to be executed and that depend on the properties of the selected object.

[0144] One objective is to describe when and under what circumstances objects must be used to generate voice guidance instructions in order to best help drivers find their way.

[0145] The context depends in particular on the complexity of the maneuver to be performed. This complexity is estimated based on at least one of the following criteria, and preferably on each of them:

[0146] - The vehicle's position relative to the moving vehicle (the situation is considered more complex if a lane change is required).

[0147] - The configuration of road infrastructure (the more lanes an intersection has, the more complex it is judged to be),

[0148] - Driving context (driver stress, i.e., pressure; congested traffic, etc.)

[0149] - There are many objects in the driver's field of vision.

[0150] - Driver's understanding of road infrastructure,

[0151] -The current stage of the vehicle, and

[0152] - Statistical weights of various detected visual indicators.

[0153] Here, preferably, visual indicators are used in voice-guided instructions only when proven necessary.

[0154] If the situation is simple, i.e., if the complexity is below a first complexity threshold, it is preferable not to use visual indicators to avoid increasing the cognitive burden on the driver. Therefore, it is preferable to say "turn left at this intersection" rather than "turn left after the bus stop," because the latter instruction will require the driver to locate the bus stop and the intersection, not just the intersection.

[0155] Conversely, if the complexity exceeds this first complexity threshold, instructions are generated using time-based or visual cues (or "indicators"), which improves the understanding of the maneuver to be performed. This is known as "enhanced voice-guided instructions."

[0156] Depending on various parameters, this enhancement can be carried out in a variety of ways.

[0157] One way to enhance voice-guided instructions is to incorporate a time concept into the instructions to facilitate understanding of the action to be performed. This can be illustrated with several examples of using contextual prompts.

[0158] In the first example, if traffic is congested or if an object is blocking or obstructing the target of the maneuver, the processing module 25 is programmed to integrate contextual cues such as "as soon as possible" into the guidance instruction, allowing the driver to freely reposition his vehicle at the optimal moment.

[0159] In the second example, if the maneuver is possible and must be performed immediately, the processing module 25 is programmed to include another contextual prompt, such as "now" or "immediately," so that the driver can reposition his vehicle without delay.

[0160] In the third example, if a maneuver can be performed but does not need to be performed immediately, the processing module 25 is programmed to include contextual cues such as "as soon as possible," allowing the driver to freely reposition his vehicle at the optimal moment.

[0161] Each bootstrap instruction is then formed by sequentially arranging at least one action, a contextual cue, and an attribute.

[0162] In cases where a boot instruction is cascaded with another boot instruction, the processing module 25 is programmed to replace the contextual prompt "as soon as possible" with "then" or "after".

[0163] In this context, a complete set of guiding instructions is formed by sequentially arranging actions and egocentric attributes and cascading them with actions, contextual cues, and another attribute.

[0164] For example, this complete guidance instruction can be expressed as "turn right first, then turn left", or "turn right first, then turn left", or "turn right, then proceed along Long Island".

[0165] A second way to enhance voice-guided instructions is to use motion indicators to facilitate understanding of the action to be performed. This method is used when the motion indicator is the object with the highest weight detected in the acquired image.

[0166] It can be assumed that processing module 25 has succeeded:

[0167] - Detected various types of movement indicators (cars, trucks, buses, vans, bicycles, motorcycles).

[0168] - Position the movement indicator on the lane of the map obtained from map database 20.

[0169] - Detect the properties of these movement indicators in order to distinguish them from each other.

[0170] - A unique movement indicator is identified, which is clearly visible to the driver by being characterized at least by its type, its color, and its location, and

[0171] - The unique motion indicator is identified as stable (i.e., it appears in various consecutive processed images).

[0172] In this case, the processing module 25 is programmed to use this unique movement indicator to help the driver orient his vehicle correctly.

[0173] Then, complete guidance instructions are formed by arranging actions, geographic indicators, visual cues, and attributes in sequence.

[0174] For example, the guidance instruction could be expressed as "keep in the same lane as the red car".

[0175] A third way to enhance voice-guided instructions is to use static indicators (such as elements of road infrastructure) to facilitate understanding of the action to be performed. This approach is used when the object with the highest weight detected in the acquired image is a static indicator.

[0176] Therefore, it can be considered that processing module 25 has succeeded:

[0177] - Various types of static indicators (bridges, traffic lights, etc.) were detected.

[0178] - Associate the detected static indicators with objects stored in the map database 20.

[0179] - Detect the properties of these static indicators in order to distinguish them from each other.

[0180] - A unique static indicator has been identified, which is clearly visible to the driver by being characterized at least by its type, color, and location, and has been successfully identified.

[0181] - The indicator was identified as stable.

[0182] In this case, the processing module 25 is programmed to use this unique static indicator to help the driver correctly orient his vehicle.

[0183] Then, complete guidance instructions are formed by arranging actions, targets, geographical indicators, and visual cues in sequence.

[0184] For example, this complete guidance instruction can be expressed in the form of "exit after entering the bridge".

[0185] A fourth way to enhance voice-guided instructions is to use directional markers to facilitate understanding of the action to be performed. This method is used when the marker is the object with the highest weight detected in the acquired image.

[0186] Therefore, it can be considered that processing module 25 has succeeded:

[0187] - Various types of signs and their attributes (traffic lights, stop signs, etc.) were detected.

[0188] - The detected signs are associated with those contained in the map database 20, specifically to determine the attributes of each sign (color, text, indicated direction, etc.).

[0189] - This determines the attributes of each flag in order to distinguish them.

[0190] - Identify unique signs that are clearly visible to the driver (by their color, the text written on them, etc.).

[0191] - The logo is recognized as stable (and in particular, the shape of the text or display does not change in the short term).

[0192] In this case, the processing module 25 is programmed to use this unique sign and its attributes to help the driver correctly locate his vehicle.

[0193] Then, complete guidance instructions are formed by arranging actions, geographical indicators, self-centered attributes, and purposes in sequence.

[0194] For example, this complete guidance instruction can be expressed as "Enter the left lane and drive in the direction of the green sign".

[0195] A fifth approach to enhancing voice-guided commands involves using important locations (restaurants, pharmacies, etc.) to facilitate understanding of the action to be performed. This approach is used when the most weighted object detected in the acquired image is an important location.

[0196] Therefore, it can be considered that processing module 25 has succeeded:

[0197] - Various types of static indicators that form important locations were detected.

[0198] - Associate the detected static indicators with the indicators contained in the map database 20.

[0199] - Detect the attributes of these indicators in order to distinguish them from each other.

[0200] - A unique indicator is identified that is most easily seen by the driver due to its attributes (indicator type, color, and position), and

[0201] - The indicator was identified as stable.

[0202] In this case, the processing module 25 is programmed to use this important location and its properties to help the driver correctly locate his vehicle.

[0203] Then, complete guidance instructions are constructed by arranging actions, self-centered attributes, geographic indicators, and visual cues in sequence.

[0204] For example, the instruction could be expressed as "turn right after the restaurant".

[0205] A sixth approach to enhancing voice-guided commands involves using salient locations (known important buildings whose function is unlikely to change in the next few years, such as historical sites or cathedrals) to facilitate understanding of the action to be performed. This approach is used when the object with the highest weight detected in the acquired image is a salient location.

[0206] In this case, to help the driver correctly locate his vehicle, the processing module 25 is programmed to use only the name of the prominent location without adding any attributes.

[0207] Then, complete guidance instructions are constructed by arranging actions, self-centered attributes, geographic indicators, and visual cues in sequence.

[0208] For example, the complete guiding instruction can be expressed as "turn right after the Empire State Building".

[0209] A seventh approach to enhancing voice-guided instructions involves using two important locations and / or salient locations to facilitate understanding of the action to be performed. This approach is used when the object with the highest weight detected in the acquired image is a salient location or important location, and when one of the other detected objects is also a salient location or important location.

[0210] Therefore, it can be considered that processing module 25 has succeeded:

[0211] - Detected various types of static indicators that form important or prominent locations.

[0212] - Associate the detected static indicators with the indicators contained in the map database 20.

[0213] - Detect the attributes of these indicators in order to distinguish them from each other.

[0214] - A moving target was identified between two detected static indicators that were clearly visible to the driver.

[0215] - These indicators have been identified as stable.

[0216] In this case, the processing module 25 is programmed to use these prominent locations or important places and their attributes to help the driver correctly locate his vehicle.

[0217] Then, complete guidance instructions are formed by arranging actions, self-centered attributes, geographical indicators, first visual cues, and second visual cues in sequence.

[0218] For example, this complete guiding instruction can be expressed as "turn right between the church and the pharmacy".

[0219] An eighth way to enhance voice-guided instructions is to use road features to facilitate understanding of the action to be performed.

[0220] Specifically, map database 20 stores information about road characteristics that can be used in the context of this invention.

[0221] Then it can be assumed that the processing module 25 has successfully detected the property that distinguishes the road to be taken from other roads visible to the driver.

[0222] In this case, the processing module 25 is programmed to use this property to help the driver correctly orient his vehicle.

[0223] Then, complete guiding instructions are formed by arranging the actions, self-centered attributes and goals, and even the attributes of the goals in sequence.

[0224] For example, this complete guidance instruction can be expressed in the form of "turn right onto the side road", "turn right onto the side street", or "turn right onto the paved road".

[0225] The ninth method for enhancing voice guidance instructions is used in the following situations: when guidance instructions are recommended during the prediction or description phase, but no visual indicators are available at that time, while visual indicators are available during the accompanying phase.

[0226] In this case, the processing module 25 is programmed to use the visual indicator to generate enhanced voice guidance instructions as early as possible.

[0227] The visual indicator can be used before the driver can see it (during the description phase or even the anticipation phase) or from the moment he can see it. In particular, it is conceivable that the processing module 25 commands the voice guidance instruction to be repeated, using the visual cue to supplement the voice guidance instruction during repetition.

[0228] Depending on the complexity of the maneuver to be performed and the degree of hesitation detected, the tenth method of enhanced voice guidance instructions is used.

[0229] Therefore, when it is determined that the complexity of the maneuver is very high (above the second predetermined complexity threshold), or when the driver's hesitation exceeds the predetermined threshold, it is stipulated that multiple types of reinforcements will be combined.

[0230] The degree of driver hesitation can be assessed in various known ways, such as depending on the path taken (whether it deviates from the expected path), the movement of the steering wheel, the way the driver holds the steering wheel, and the driver's position.

[0231] The first example of combining these elements includes using time-based reinforcement and any other type of reinforcement in voice-guided instructions.

[0232] A second example of combining these approaches includes using reinforcements that utilize important locations or prominent sites, and reinforcements that utilize the characteristics of the road to be traveled.

[0233] In this example, the complete guidance instruction is then formed by sequentially arranging the action, self-centered attributes, geographical indicators, visual cues, target, and target attributes. This complete guidance instruction can then be expressed as "Turn right onto the path after the Empire State Building."

[0234] A third example of combining these approaches includes using reinforcement with directional signs and reinforcement with important or prominent locations.

[0235] A fourth example of combining these elements includes using two reinforcements that utilize important or prominent locations.

[0236] In this example, the complete onboarding instructions are then formed by sequentially arranging the action, the self-centered attribute, the visual cue, the geographical indicator, and another visual cue. This complete onboarding instruction can be expressed as "Turn right after the Empire State Building between the church and the pharmacy."

[0237] A fourth example of combining these approaches includes using reinforcements that utilize mobile signs and reinforcements that utilize fixed signs or road infrastructure elements.

[0238] In this example, the complete guiding instruction is then formed by sequentially arranging the action, the self-centered attribute, the first visual cue, the second visual cue, and the cue's attribute. This complete guiding instruction can be expressed as "turn left at a traffic light like a red car."

[0239] As taught above, voice-guided instructions should not be amplified in cases of low complexity.

[0240] It also teaches that the complexity of a maneuver depends on multiple variables.

[0241] Preferably, the complexity may also depend on another parameter, namely, the cost in terms of time or distance that a driver’s mistake (if he takes the wrong exit) will result in.

[0242] Therefore, when the processing module 25 predicts that if the driver makes an error during maneuvering, the error will result in a significant increase in the remaining travel time, the processing module is programmed to increase the complexity value, which leads to enhanced voice guidance instructions.

[0243] Therefore, it should be understood that, in this particular case, the regulations may more frequently or even systematically employ a combination of various enhancements, such as those detailed above.

[0244] Specifically, if failing to perform a maneuver would significantly impact the estimated time to reach the destination, the processing module 25 must provide the driver with assurances of the maneuver's success. This is particularly relevant on highways, roads with few entrances and exits, or one-way roads.

[0245] In practice, processing module 25 is programmed to calculate the additional time (or additional distance) required to reach the destination if the driver takes the wrong turn during the maneuver to be performed. Therefore, when the time is above a predetermined threshold (expressed as time, distance, or percentage of time remaining before reaching the destination), processing module 25 will generate more enhanced voice guidance instructions compared to voice guidance instructions when the calculated time is below that threshold.

[0246] Preferably, the processing module 25 may use a firmer tone than usual when giving voice guidance commands to make the driver understand the importance of performing the correct maneuver.

[0247] The cost of not performing the maneuver (and therefore the choice regarding generating more enhanced voice guidance instructions) can also be calculated based on other parameters (route type, road type, etc.), and may increase in cases of explicit recognition, such as in the following situations:

[0248] - The vehicle is not in the correct position and cannot be maneuvered.

[0249] - The vehicle is correctly positioned relative to the maneuver to be performed, but the relevant turn signals are not yet activated.

[0250] - The driver did not act as advised before the end of the anticipation phase or the end of the description phase.

[0251] In these cases, the processing module 25 may be programmed to repeat the voice guidance instruction but with added emphasis, such as “Now enter the… lane”, or to add another emphasis that allows for the execution of the maneuver, such as “Now enter the… lane, in the direction of…”.

[0252] As the vehicle moves forward, the context may change, making guidance instructions ambiguous at best and outdated at worst.

[0253] Therefore, two scenarios can be considered.

[0254] The first scenario occurs when an element enters between the vehicle and the object selected to generate guidance instructions, thus preventing the driver from seeing the object.

[0255] The second scenario occurs when another object, similar to the one selected to generate the guidance instructions, comes into the driver's field of vision. This could happen, for example, when a bus leaves its stop and a second bus stop is displayed.

[0256] Both of these situations will lead to increased driver stress and cognitive burden, and create a risk of errors or maneuvers being executed too late.

[0257] In both cases, the processing module 25 will then be programmed to generate new instructions suitable for the new context.

[0258] As mentioned above, the voice guidance instructions to be spoken vary depending on the current stage of the vehicle. Therefore, it will now be described how to formulate the guidance instructions in each of these stages.

[0259] Prediction phase

[0260] Recall that the anticipation phase occurs before the maneuver to be performed. It begins and ends at a certain distance from the area where the maneuver will be carried out. It allows the driver to position the vehicle and / or prepare for the maneuver as optimally as possible (especially when the target of the maneuver is only visible in the last minute).

[0261] During this predictive phase, it is recommended to strengthen voice guidance instructions, but avoid using prominent locations or important places.

[0262] The processing module 25 is programmed to issue so-called predictive voice guidance commands when the route being traveled includes multiple lanes.

[0263] The way to strengthen this predictive guidance may vary depending on the circumstances.

[0264] Normally (i.e., in cases other than those considered below), it is strengthened in the following manner.

[0265] If the vehicle is on a multi-lane highway and in the lane most suitable for the next maneuver, the processing module 25 is programmed to generate predictive voice guidance instructions to dispel any doubts the driver may have.

[0266] Then, voice-guided instructions are formed by sequentially arranging actions, familiarity criteria, geographical indicators, and egocentric attributes.

[0267] For example, this voice guidance instruction can be expressed as "stay in the lane".

[0268] If a vehicle is on a multi-lane road and is in a lane suitable for the next maneuver (typically if there is no solid white line between the vehicle's lane and the point of the next maneuver) but needs to change lanes, the processing module 25 is programmed to generate a predictive voice guidance instruction to instruct the driver to change lanes quickly.

[0269] Then, voice guidance instructions are formed by arranging actions, geographical indicators, and contextual cues in sequence.

[0270] For example, the voice guidance instruction can be expressed as "Enter lane x now (or as soon as possible)".

[0271] If a vehicle must enter a multi-lane road and is located in a limited-length entrance lane, the processing module 25 is programmed to generate predictive voice guidance instructions to instruct the driver to change lanes.

[0272] Then, voice guidance instructions are formed by sequentially arranging at least one action and a geographical indicator.

[0273] For example, the voice guidance instruction can be expressed as "enter the lane".

[0274] If a vehicle must enter a multi-lane road and the vehicle is in an entrance lane of unlimited length (i.e., a lane merged with a road lane) and that lane is best suited for the next maneuver, then the processing module 25 is programmed to generate a predictive voice guidance instruction instructing the driver to remain in that lane.

[0275] Then, voice-guided instructions are formed by sequentially arranging at least one action, a geographic indicator, and an egocentric attribute.

[0276] For example, this guidance instruction can be expressed as "stay in the lane".

[0277] If a vehicle must enter a multi-lane road and the vehicle is in an infinitely long entrance lane that is not the most suitable lane for the next maneuver, the processing module 25 is programmed to count the number of lane changes required to reach the most suitable lane and then generate a predictive voice guidance command to instruct the driver to change lanes.

[0278] Then, voice-guided instructions are formed by sequentially arranging at least one action and attribute.

[0279] For example, the voice guidance instruction can be expressed as "enter lane n".

[0280] In this situation, the way to express which lane a vehicle must move to can be determined in the following way.

[0281] In the absence of lane restrictions on multiple lanes of the highway, the processing module 25 is programmed to generate voice guidance instructions for the driver to enter.

[0282] Then, a complete guiding instruction is formed by sequentially arranging at least one action and one target.

[0283] For example, the guidance instruction can be expressed in the form of "Enter...". Then, the road name or its type can be specified.

[0284] If the highway has only two lanes, the processing module 25 is programmed to generate voice guidance instructions that indicate whether the vehicle must be in the "right lane" or the "left lane".

[0285] The guiding instructions are then formed by sequentially arranging at least one action, a geographical indicator, and a self-centered attribute.

[0286] If the highway has exactly three lanes, the processing module 25 is programmed to generate voice guidance instructions that require the vehicle to be in the "right lane", "left lane" or "middle lane".

[0287] The guiding instructions are then formed by sequentially arranging at least one action, a geographical indicator, and a self-centered attribute.

[0288] If the highway has exactly four lanes and only one lane is suitable for the maneuver to be performed, the processing module 25 is programmed to generate a voice guidance instruction that the vehicle must be in the "rightmost lane", "leftmost lane", "second lane", or "third lane".

[0289] The guiding instructions are then formed by sequentially arranging at least one action, a geographical indicator, and a non-egocentric attribute.

[0290] If the highway has exactly four lanes and multiple of them are suitable for the maneuver to be performed, the processing module 25 is programmed to generate voice guidance instructions that the vehicle must be in the "rightmost lane", "leftmost lane", or "middle lane".

[0291] The guiding instructions are then formed by sequentially arranging at least one action, a geographical indicator, and a non-egocentric attribute.

[0292] If the number of lanes on the highway is an odd number greater than four, and only the middle lane is suitable for the maneuver to be performed, then the processing module 25 is programmed to generate a voice guidance instruction that instructs the vehicle to be in the "middle lane".

[0293] The guiding instructions are then formed by sequentially arranging at least one action, a geographical indicator, and a non-egocentric attribute.

[0294] If a highway has multiple lanes and at least one of those lanes is closed to the vehicle, the processing module must integrate it into the lane count.

[0295] When the vehicle is outside the building area, the enhanced voice guidance instructions during the anticipation phase should ideally indicate to the driver the reason (i.e., the purpose) for positioning himself by describing the next maneuver to be performed.

[0296] Therefore, the purpose is to indicate the distance between the vehicle and the motor, the type of motor to be performed (highway exit, overpass, roundabout, etc.), and the road infrastructure visible to the driver (directional signs, etc.).

[0297] On highways, the processing module 25 must use visual indicators, which are preferably static (of the type of directional sign), or (in the absence of such a sign) moving indicators, or (in the absence of both signs and moving indicators) elements of the road infrastructure.

[0298] The following example illustrates this rule.

[0299] If the processing module 25 detects a sign and is able to read the text marked on it, it is programmed to generate predictive voice guidance instructions that instruct the driver to follow the directions indicated on it.

[0300] Then, complete guidance instructions are formed by sequentially arranging at least one action, a geographic indicator, an egocentric attribute, a purpose, a geographic indicator, a visual cue, and an attribute.

[0301] For example, the voice guidance instruction can be expressed as "Enter the left lane and prepare to take the exit in the direction of...".

[0302] If the processing module 25 detects a marker but cannot read the text on it, yet the map database 20 contains that information, it operates in the same manner as described above.

[0303] If the processing module 25 detects a sign with a unique color among other signs, it is programmed to generate voice guidance instructions that direct the driver to follow the direction of the sign with the identified color.

[0304] Then, voice-guided instructions are formed by sequentially arranging at least one action, a geographic indicator, an egocentric attribute, a purpose, and an attribute.

[0305] For example, the guidance instruction could be expressed as "Enter the left lane and prepare to follow the blue sign." If the text on the sign is readable or known, the instruction can be supplemented by indicating the direction it points to.

[0306] When vehicles are in built-up areas, signs are more frequently hidden or covered by vegetation. Therefore, processing module 25 must identify the most relevant visual cues among other indicators, including:

[0307] - Movement indicator,

[0308] - Elements of road infrastructure

[0309] - Static indicators other than flags, and

[0310] -Important locations or prominent places.

[0311] The following four examples illustrate this situation.

[0312] In the first example, processing module 25 generates the following guidance instruction: "Keep in the same lane as the red car." In this example, the guidance instruction is formed by sequentially arranging at least one action, a geographic indicator, a visual cue, and an attribute.

[0313] In the second example, processing module 25 generates the following guidance instruction: "Enter the left lane and drive in the direction of the green sign." In this example, the guidance instruction is formed by sequentially arranging at least one action, a geographic indicator, an egocentric attribute, a geographic indicator, a visual cue, and an attribute.

[0314] In the third example, processing module 25 generates the following guidance instruction: "Enter the left lane, then take the third exit at the roundabout." In this example, the guidance instruction is formed by sequentially arranging at least one action, a geographic indicator, a visual cue, a cascading element, another visual cue, the action, and a non-egocentric attribute.

[0315] In the fourth example, processing module 25 generates the following guidance instruction: "Enter the right lane, prepare to turn right at the intersection." In this example, the guidance instruction is formed by sequentially arranging at least one action, a geographic indicator, a visual cue, a destination, and another visual cue.

[0316] Maneuver description phase

[0317] The second stage is the description stage. In this second stage, the voice guidance instructions, called the description voice guidance instructions, represent the actions that must be performed in order to achieve the goal.

[0318] In this instruction, the processing module 25 transmits the information needed by the driver to determine his location and to locate the target relative to his environment (and especially relative to the detected visual indicators).

[0319] When a vehicle arrives at an intersection, the processing module 25 calculates the angle of each road from the intersection relative to the road where the vehicle is located. This angle is expressed in degrees and includes a range of -180° to 180°.

[0320] The guidance instructions can then be enhanced by geographic attributes suitable for the intersection, such as “completely to the right” or “facing” (or “straight ahead”).

[0321] The methods for enhancing guidance instructions may vary depending on the type of intersection.

[0322] For a T-shaped intersection, where a vehicle reaches the road forming the T via the perpendicular line of the T, the processing module 25 may be programmed to use the term "intersection".

[0323] If the angle between the vertical and horizontal of T is substantially perpendicular (e.g., within 10, 20, or 30 degrees), the processing module 25 will use the verb "turn" or "take".

[0324] If the angle between the vertical and horizontal axes of T is not essentially perpendicular, then two cases must be considered.

[0325] The first scenario is when a vehicle must deviate from its initial trajectory at an angle far less than 90 degrees at an intersection. In this case, processing module 25 will supplement the verb used with the expression "slightly". It will further favor using the verb "go" instead of the verb "turn".

[0326] The second scenario is when a vehicle must deviate from its initial trajectory by an angle greater than 90 degrees at an intersection. In this case, processing module 25 will supplement the verb used with the expression "completely". It will further favor using the verb "turn" instead of the verb "go".

[0327] A "forked" intersection is an intersection consisting of one entrance lane and two exit lanes located on an extension of the entrance lane.

[0328] Then, the method of developing voice guidance instructions at such intersections will depend on the configuration of the relevant site.

[0329] When both exit lanes are to the right of the entrance lane, if a vehicle must use the rightmost lane, instructions such as "use the rightmost lane" should be used. Conversely, if a vehicle must use the other exit lane, stronger guidance is needed to instruct the driver which exit lane to take.

[0330] Conversely, when both exit lanes are to the left of the entrance lane, if a vehicle must use the leftmost lane, instructions such as "use the leftmost lane" are used. Conversely, if a vehicle must use the other exit lane, stronger guidance is needed to instruct the driver which exit lane to take.

[0331] When the two exit lanes are located on either side of the entrance lane, instructions will be given using expressions such as "take the left lane" or "take the right lane".

[0332] At roundabouts, voice guidance instructions will be defined in particular based on the number of exits at the roundabout.

[0333] A roundabout consists of a circular road and roads leading to this central road, which are referred to below as "adjacent roads". A roundabout has entrances and exits, which may be located on the same adjacent road.

[0334] It is worth noting that the complexity of a roundabout depends on the following criteria:

[0335] - The angle between the road the vehicle arrives at and the road it must leave.

[0336] - The angle between the exit you want to take and the nearest exit.

[0337] -Number of adjacent roads

[0338] - The difference between the number of imports and the number of exports

[0339] - The number of lanes on a ring road,

[0340] - The outer diameter of the roundabout.

[0341] A roundabout is considered simple if it has as many entrances as exits, paired on adjacent roads, if the number of adjacent roads is four or less, if these adjacent roads have significantly different directions, and if the roundabout prioritizes use by vehicles traveling on the circular road. Otherwise, a roundabout is considered complex.

[0342] The first example of a simple roundabout is one with four adjacent roads spaced approximately 90 degrees apart. This is called an "orthogonal roundabout". In this example, the processing module 25 could be programmed to give voice guidance instructions, such as "Exit at the right (or left or continue straight) exit at the roundabout".

[0343] A second example of a simple roundabout is one with three or four adjacent roads separated by angles, at least one of which is significantly different from 90 degrees. In this example, processing module 25 could be programmed to provide a voice-guided instruction reinforced with the lane number of the exit to be taken. This instruction might be expressed as "Take the first exit on the right at the roundabout."

[0344] The third example of a simple roundabout is a small-diameter roundabout. In this example, processing module 25 will treat this roundabout as a simple intersection (see below).

[0345] The description emphasizes that voice guidance instructions, which allow navigation through complex roundabouts, will always be reinforced.

[0346] Various examples of instructions that will be enhanced can be given.

[0347] In the first example, the roundabout does not have four pairs of orthogonal exits.

[0348] Therefore, the first case to consider is when the exit is orthogonal to the entrance lane where the vehicle arrives (within 10 degrees) and corresponds to the first or second exit. In this case, the processing module 25 is programmed to preferably provide a voice guidance instruction reinforced with the number of the exit to be taken. This instruction might be expressed as "Take the first exit on the right at the roundabout."

[0349] The second scenario to consider is where the exit is not orthogonal to the entrance lane where the vehicle arrives and corresponds to the first or second exit. In this case, the processing module 25 is programmed to forcibly issue a voice guidance instruction reinforced with the number of the exit to be taken. This instruction might be expressed in the form of "take the first exit on the right at the roundabout".

[0350] The third case to consider is when the exit corresponds to a third or subsequent exit. In this case, the processing module 25 is programmed to issue a voice-guided instruction reinforced by a non-egocentric attribute. This instruction could be expressed as "Take the exit to the left before entering the restaurant."

[0351] A second example of a complex roundabout is a roundabout with a different number of entrances and exits.

[0352] One particular scenario to consider is an entrance that has more entrances than exits, located on an adjacent one-way street that is different from the lanes where vehicles arrive.

[0353] In this example, processing module 25 applies the rules defined in the previous example, while sometimes avoiding using their numbers to indicate the exit to take.

[0354] Therefore, if the exit to be taken is located after a one-way entrance, the processing module 25 is programmed to issue a voice-guided instruction reinforced by non-egocentric attributes, rather than indicating an exit number. This instruction might be expressed as "Go to the exit on the right next to..."

[0355] Conversely, if the exit to be taken is located before a one-way entrance, the processing module 25 is programmed to speak the voice guidance instructions in the same manner as in the first example of the complex roundabout given above.

[0356] A third example of a complex roundabout is one where the distances or angles between the exits are very close to each other. In particular, a roundabout is considered complex when the angle between two exits is less than or equal to 30°.

[0357] In this example, processing module 25 applies the same rules as those applied in the first example of the complex roundabout given above, while also forcibly amplifying the voice guidance instructions.

[0358] A fourth example of a complex roundabout is one in which vehicles arriving at the roundabout have priority over vehicles already in the roundabout (the roundabout at Place de l'Étoile in Paris, France, is the most famous example of this type of roundabout).

[0359] In this example, the processing module 25 applies the same rules as those applied in the first example of the complex roundabout given above, while forcibly reinforcing the voice guidance instructions with warnings such as “Caution! Vehicles from the right have the right of way in the roundabout.”

[0360] At an intersection with at least four lanes extending from it, processing module 25 applies the same rules as those applied in the various examples of roundabouts given above, except that the term “roundabout” is replaced with “intersection” or “crossroads”.

[0361] The processing module 25 can count the exits of vehicle lanes, this time counting all lanes, even those that are one-way or have no entrance.

[0362] If the exit lane to be taken is offset relative to other exit lanes, the processing module 25 is programmed to add commands such as "Proceed through the intersection and take the second exit on the left in the direction of..." to the instruction.

[0363] Maneuvering Accompanying Phase

[0364] The accompanying phase is primarily designed to indicate to the driver the exact moment when he can begin actual maneuvering (i.e., changing direction). During this phase, the processing module 25 only sends so-called accompanying voice guidance instructions when it determines that the situation at the intersection is complex.

[0365] In this case, whether to strengthen voice guidance instructions and how to strengthen them depends on the type of intersection.

[0366] Therefore, when multiple T-junctions appear one after another on the road and there are no visual indicators (infrastructure, signs, landmarks, etc.) nearby to distinguish them, the processing module 25 will be programmed to issue accompanying voice guidance instructions. These instructions will include contextual prompts and can be formulated as follows: “Turn right now…”.

[0367] At forked intersections, no accompanying voice guidance instructions will be sent.

[0368] At roundabouts, if the roundabout's diameter is less than a predetermined threshold and the exits are not paired orthogonal, processing module 25 will be programmed to generate accompanying voice guidance instructions during maneuvers by repeating instructions already sent during the maneuver description phase or by using new, previously invisible features or salient locations or road infrastructure elements. The features or salient locations or road infrastructure elements used must, of course, be visible, unique, and immediately identifiable by the driver.

[0369] The instruction might be worded as: "Leave the roundabout now, before Hotel X."

[0370] If the processing module 25 detects that the driver has circled the roundabout once, it can be programmed to generate accompanying voice guidance instructions that repeat the voice guidance instructions sent during the description phase.

[0371] If the voice guidance instruction sent during the description phase does not involve any visual indicators because there are no clearly visible indicators at the time, and the vehicle speed has decreased significantly (due to driver hesitation or traffic congestion), then the processing module Figure 25 is programmed to generate accompanying voice guidance instructions using features or prominent location or road infrastructure elements that have become visible.

[0372] The decrease here must be greater than the predetermined threshold.

[0373] Eliminating doubts stage

[0374] Once the maneuver has been performed, the processing module 25 assesses whether it is necessary to reassure the driver about the quality of the maneuver based on its complexity and any driver hesitation.

[0375] Therefore, if the processing module 25 detects hesitation during or upon exiting a maneuver, it is programmed to generate an enhanced voice guidance instruction. The decision to generate this instruction may also depend on other factors, such as the complexity of the maneuver being performed or the immediate presence of another maneuver to be performed.

[0376] The instruction is preferably reinforced by a visual cue located near the end of the maneuver or visible only at the completion of the maneuver, which is preferably different from the visual cue used to date.

[0377] continuation phase

[0378] When no maneuver is required within a distance or time greater than a threshold (e.g., more than one minute or more than one kilometer), the processing module 25 preferably indicates to the driver that it remains active.

[0379] Then, voice guidance instructions can be tailored to the specific road conditions and the driver's surroundings.

[0380] Consider a scenario where at least one intersection is detected between the vehicle's position and the next maneuver to be performed, there is no plan to change roads, and lane changing is impossible. In this case, processing module 25 is programmed to generate voice guidance instructions that depend on the road curvature.

[0381] Therefore, if the radius of curvature of the road is greater than a predetermined threshold, a voice guidance instruction is formed by sequentially arranging at least one action, a geographic indicator, an egocentric attribute, a purpose, and a geographic indicator. It can be expressed in the form of "Continue straight through the intersection..."

[0382] Conversely, if the radius of curvature of the road is less than this threshold, the voice guidance instruction is formed by sequentially arranging at least one action, a geographic indicator, an egocentric attribute, a purpose, another geographic indicator, and another egocentric attribute. It can be expressed in the form of "follow the main road through the next intersection..."

[0383] The following scenario is also considered: at least one intersection is detected between the vehicle's position and the next maneuver to be performed, there is no plan to change roads, and lane changing is not desired due to the upcoming maneuver. In this case, the processing module 25 is again programmed to generate voice guidance instructions that depend on the road curvature.

[0384] Therefore, if the radius of curvature of the road is greater than a predetermined threshold, a voice guidance instruction is formed by sequentially arranging at least one action, a geographic indicator, an egocentric attribute, a purpose, and another geographic indicator, and then cascading them with another action, a familiarity criterion, another geographic indicator, and another egocentric attribute. This can be expressed in the form of "Continue straight through the intersection and stay in your lane..."

[0385] Conversely, if the radius of curvature of the road is less than this threshold, a voice guidance instruction is formed by sequentially arranging at least one action, a geographic indicator, an attribute, a purpose, and another geographic indicator, and then cascading them with another action, a familiarity criterion, another geographic indicator, and another egocentric attribute. This could be expressed as "Go along the main road through the next intersection and stay in your lane."

[0386] Finally, consider the following scenario: only one intersection is detected between the vehicle's position and the next maneuver to be performed, and there is no planned road change but a lane change is required due to the next maneuver. In this case, the processing module 25 is again programmed to generate voice guidance instructions that depend on the road curvature.

[0387] Therefore, if the radius of curvature of the road is greater than a predetermined threshold, a voice guidance instruction is formed by sequentially arranging at least one action, a geographic indicator, an attribute, a purpose, and another geographic indicator, and cascading them with another action, a familiarity criterion, and a target. This can be expressed as "Continue straight at the intersection, then enter lane X."

[0388] Conversely, if the radius of curvature of the road is less than this threshold, a voice guidance instruction is formed by sequentially arranging at least one action, a geographic indicator, an attribute, a purpose, and another geographic indicator, and cascading them with another action, a familiarity criterion, and a target. This can be expressed as "Drive along the main road at the next intersection, then move to lane X."

[0389] When the distance between the vehicle's location and the next maneuver to be performed (or the time required to reach the maneuver area) exceeds a predetermined threshold (e.g., 2 km), the processing module 25 is programmed to integrate the concept of distance into the voice guidance command. This can be expressed in the form of "Continue straight for x km..."

[0390] In the phase known as the continuation phase, as the maneuver to be performed approaches, voice guidance instructions may also refer to visible (and distinctive) road infrastructure elements (roundabouts, bridges, tunnels, intersections) or distinctive or prominent locations (stations, gas stations, neon lights). Here, we again consider both scenarios.

[0391] If the radius of curvature of the road is greater than a predetermined threshold, the voice guidance instruction can be expressed as "Continue straight to the gas station".

[0392] Conversely, if the radius of curvature of the road is less than this threshold, the voice guidance instruction can be expressed as "drive along the main road to the next bridge".

[0393] If the same maneuver can be performed in multiple locations, the voice guidance instructions can be expressed in the following forms: "Continue straight until the next instruction" or "Continue straight to the end of the street" or "Continue straight to the high-rise building...".

Claims

1. A method for generating voice-guided instructions for an individual, the method comprising the following steps: - Using computers to determine the route to take. - The image acquisition device (10) acquires at least one image of the individual's environment. - Process the image to detect at least one object therein and characterize the object, and - Establish and send voice guidance instructions, thereby informing the individual how to maneuver to follow the stated route. The feature is that it provides a step of determining the complexity of the maneuver based on the vehicle's position relative to the area in which the maneuver will be performed, in order to consider whether a lane change is necessary to reach the area, and The feature is that, in the establishment step, the voice guidance instruction is formulated only when the complexity of the maneuver is higher than a first threshold, using a prompt inferred from the representation of the object; and if the complexity of the maneuver is higher than a second threshold that is strictly higher than the first threshold, then two prompts are used to generate the voice guidance instruction.

2. The method as described in claim 1, wherein, The calculation of this complexity depends on at least one of the following criteria: - The location of the vehicle relative to the area in which the maneuver must be performed. - The type of road infrastructure must exist in the area where the aforementioned maneuver is performed. - The tension felt by the individual, - Road traffic density, - The number of objects detected in at least a portion of the acquired image. - The individual's knowledge of the road infrastructure, - The current stage of the vehicle Furthermore, if the road infrastructure is a roundabout, it also depends on at least one of the following criteria: - The angle between the axis from which the person arrives at the roundabout and the axis from which the person leaves the roundabout. - The angle between at least two roads leading to the roundabout. - Number of roads leading to the roundabout - The difference between the number of entrances to the roundabout and the number of exits to the roundabout. - The number of lanes in the roundabout, and - The outer diameter of the roundabout.

3. The method as described in any one of the preceding claims, wherein, The cue is a visual cue visible to the individual. If the visual cue disappears from the environment visible to the individual between the moment the voice guidance instruction is sent in audible form and the moment the maneuver must be performed, then steps are specified for generating a new voice guidance instruction without using the visual cue.

4. The method as described in claim 1 or 2, wherein, The prompt is a visual cue visible to the individual. If, between the moment the voice guidance instruction is sent in audible form and the moment the maneuver must be performed, another visual cue, indistinguishable from the visual cue used in the voice guidance instruction, appears in the environment visible to the individual, then the steps for generating a new voice guidance instruction different from the one sent in audible form are specified.

5. The method as described in claim 1 or 2, wherein, The regulations define the costs associated with additional time or distance incurred due to errors in following the route, and the complexity is determined based on these costs.

6. The method as claimed in claim 1 or 2, wherein, The specification stipulates that at least two distinct and consecutive phases must be distinguished before the execution of the maneuver: a maneuver prediction phase in which voice guidance instructions can be sent to instruct the individual on how to position themselves to approach the maneuver, and a description phase in which voice guidance instructions are sent to instruct the individual on how to execute the maneuver. The voice guidance instructions sent in these phases use different prompts. The specification further stipulates at least one additional phase following the description phase in which voice guidance instructions can be generated, said additional phase being: - The accompanying phase that occurs simultaneously with the aforementioned maneuver, or - Following the maneuver and allowing the individual to be instructed on whether he or she performed the maneuver correctly, a dispelling doubts phase, or - Allow the individual to be informed of the ongoing phases of the activity characteristics of the generation method.

7. The method of claim 6, wherein, If the voice guidance instructions sent during the description phase do not involve any visual indicators and the vehicle decelerates, then it is stipulated that a voice guidance instruction with a prompt be sent during the accompanying phase.

8. The method of claim 6, wherein, During this dispelling doubts phase, it is stipulated that the voice guidance instructions be sent using visual indicators that are different from one or more visual cues used in one or more voice guidance instructions sent during the previous phase.

9. The method of claim 6, wherein, In this continuation phase, different voice guidance instructions are generated depending on whether the radius of curvature of the road being traveled is below or above a threshold.

10. The method as claimed in claim 1 or 2, wherein, The steps of determining the route by computer include: retrieving static visual indicators located on the route and attributes associated with the static visual indicators from a map database (20); defining the attributes common to multiple static visual indicators, and using a prompt that includes the common attributes to formulate the voice guidance instruction in the establishment step when the object is characterized as including the common attributes in the step of processing the image.

11. The method of claim 10, wherein, The static visual indicator is a road sign, and its common attributes include the name written on the road sign or the background color of the road sign. The object is a road sign.

12. The method of claim 10, wherein, The steps of processing the image include detecting multiple objects, and the generation method further includes a step of classifying the detected objects, wherein objects characterized as having the common attribute are given priority, and the objects selected for the generation step are the objects given priority.

13. An apparatus for generating route guidance instructions suitable for an environment, the apparatus comprising: - Image acquisition device (10) for acquiring images of the environment. - Map database (20) - Navigation and geolocation systems (22). - Processing module (25), which is adapted to receive and combine elements from the image acquisition device (10), the map database (20), and the navigation and geolocation system (22), and - A voice transmitter, suitable for sending instructions to the individual. The characteristic feature is that the processing module (25) is programmed to implement the method as described in any one of the preceding claims.

Images