FR3150494A1 - Method and device for autonomous driving of a vehicle - Google Patents

Abstract

The invention relates to a method and a device for autonomous driving of a vehicle, called an ego vehicle, said ego vehicle traveling on a road, said road comprising at least two lanes, a first lane and a second lane, said ego vehicle traveling on said first lane, a vehicle, called a target vehicle, traveling on said second lane, said method comprising the steps of: Receiving (201) a first probability indicator; Receiving (202) a location of said target vehicle; Receiving (203) an area focused by a gaze of a driver of said ego vehicle; Receiving (204) a modeling of said first and second lanes; Determining (205) whether the driver focuses on said first lane or, said second lane or the target vehicle; Generating (206) a second probability indicator; Sending (207) an autonomous driving instruction from said second probability indicator. Figure to be published for the abstract: Figure 2

Description

Method and device for autonomous driving of a vehicle Technical field of the invention

The invention relates to the field of autonomous vehicle driving assistance systems. In particular, the invention relates to a method and device for the autonomous driving of a vehicle.

State of the art

The term "vehicle" refers to any type of vehicle, such as a car, moped, motorcycle, warehouse robot, etc. "Autonomous driving" of a vehicle refers to any process capable of assisting the driving of the vehicle; the vehicle is then also called an "autonomous vehicle." This process may consist of partially or fully controlling the vehicle or providing any type of assistance to a person driving the vehicle. The process thus covers all autonomous driving, from level 0 to level 5 in the OICA (International Organization of Motor Vehicle Manufacturers) scale.

The systems designed to assist vehicle driving are also called ADAS (Advanced Driver Assistance Systems), ADAS systems, or driver assistance systems. Currently, a commercially available vehicle includes several well-known ADAS systems, such as at least adaptive cruise control (also called cruise control, system, function, or ACC, from the English acronym "Auto Cruise Control"), lane positioning assist (LPA, from the English acronym "Lane Positioning Assist"), and so on.

Some ADAS systems, such as ACC or LPA, require adapting their controls based on the behavior of adjacent vehicles. These systems determine a future trajectory for the vehicle, referred to as the ego vehicle, which is equivalent to determining the future locations of the ego vehicle, or corridor. To allow the ego vehicle to move safely, these systems determine a future trajectory based on a probability distribution of the presence of other vehicles along that trajectory.

We know of such a system through document FR1872874. However, there are still real-life situations where another vehicle is not taken into account in the same way a driver would be. For example, if the Ego vehicle is traveling on a two-lane highway, and the Ego vehicle is overtaking a truck, with another vehicle in front of the truck traveling in the same lane, this other vehicle, hidden by the truck, is not immediately detected by the sensors on board the Ego vehicle. Indeed, these ADAS systems use a radar sensor that is highly directional, only able to detect an object within a cone along the Ego vehicle's normal longitudinal axis of travel. These ADAS systems also use an onboard camera, with associated image processing, mounted at the top of the Ego vehicle's windshield. With the truck blocking the view of the other vehicle, once the ego vehicle has passed the truck, once the other vehicle becomes detectable by the ADAS system sensors, the ADAS system must temporally confirm (perform a number of successive coherent detections) before being certain of properly detecting the other vehicle.

This other vehicle may intend to change lanes and, therefore, before the aforementioned timing confirmation, find itself at least partially in the lane of the ego vehicle in front of it. The merging of another vehicle into the ego vehicle's lane, in front of it, is called a "cut-in." In this situation, an ADAS system must react early enough to manage it effectively, such as braking the ego vehicle comfortably and without causing anxiety. It must also avoid any braking that could be considered untimely when approaching the other vehicle in the adjacent lane. With current sensors, it is difficult to configure the ADAS system to achieve a good compromise.

It is also known from vehicle occupant monitoring devices, particularly eye-tracking and driver attention monitoring devices. These devices are capable of determining the direction of the driver's gaze and identifying an area focused on their gaze. This area can be a location within the driver's area, such as a predetermined area on a display. It can also be a location outside the vehicle, such as another vehicle (i.e., along the vehicle's normal longitudinal axis of travel), or another vehicle outside the longitudinal axis.

One object of the present invention is to remedy the aforementioned problem, in particular to select as early as possible a new vehicle, called the target vehicle, by an ADAS system during a cut-in of said target vehicle.

• Réception d’un premier indicateur de probabilité, ledit premier indicateur caractérisant une probabilité qu’un trajet futur dudit véhicule cible intercepte un trajet futur dudit véhicule égo ;
• Réception d’une localisation dudit véhicule cible ;
• Réception d’une zone focalisée par un regard d’un conducteur dudit véhicule égo, les zones qui peuvent être focalisées par le regard étant prédéterminées, ladite zone étant émise par un système d’analyse d’un regard d’un conducteur ;
• Réception d’une modélisation desdites première et deuxième voies ;
• Détermination, en fonction de ladite localisation du véhicule cible, de ladite zone focalisée et de ladite modélisation, si ledit conducteur focalise du regard la première voie ou, si le conducteur focalise ladite deuxième voie ou le véhicule cible ;
• Génération d’un deuxième indicateur de probabilité en fonction dudit premier indicateur de probabilité et de ladite détermination ;
• Emission d’une instruction de conduite autonome à partir dudit deuxième indicateur de probabilité.

To this end, a first aspect of the invention relates to a method for the autonomous driving of a vehicle, called the ego vehicle, said ego vehicle traveling on a road, said road comprising at least two lanes, a first lane and a second lane, said ego vehicle traveling on said first lane, a vehicle, called the target vehicle, traveling on said second lane, said method being implemented by a processor and comprising the steps of:

• Receipt of a first probability indicator, said first indicator characterizing a probability that a future route of said target vehicle intercepts a future route of said ego vehicle;
• Receipt of a location of said target vehicle;
• Reception of a focused area by a look from a driver of said vehicle, the areas which can be focused by the look being predetermined, said area being emitted by a driver's look analysis system;
• Reception of a model of said first and second channels;
• Determination, based on said location of the target vehicle, said focused area and said modeling, whether said driver focuses their gaze on the first lane or whether the driver focuses on said second lane or the target vehicle;
• Generation of a second probability indicator based on said first probability indicator and said determination;
• Issuance of an autonomous driving instruction based on said second probability indicator.

Thus, autonomous driving is contextualized by the driver's gaze. Depending on the area focused by the driver's gaze, the probability that the target vehicle's trajectory will intersect the ego vehicle's trajectory will be adjusted. The autonomous driving will then adapt accordingly.

Advantageously, when the driver focuses their gaze on the second lane or target vehicle, said second generated probability indicator is greater than said first probability indicator.

The driver's gaze, and therefore their attention, is directed towards the danger. Thus, when the driver's attention shifts to an adjacent lane or a target vehicle, the increased probability indicator will allow the autonomous driving system to more quickly take the target vehicle into account, as this target vehicle is more likely to interfere with the driver's trajectory. Therefore, the autonomous driving system will anticipate the target vehicle.

Advantageously, when the driver focuses their gaze on said first lane, said second probability indicator generated is smaller than said first probability indicator.

Thus, when the driver's attention is not focused on the target vehicle—in other words, when the driver's attention is focused on the trajectory of the ego vehicle—the probability of the target vehicle's trajectory being intercepted decreases, and the target vehicle is therefore less taken into account by the autonomous driving system. Consequently, a target vehicle that the driver deems safe will be less frequently considered by the autonomous driving system, thereby preventing false detections and/or inappropriate changes to the ego vehicle's trajectory.

Advantageously, the said process further includes a step of determining a duration of focus of the said target vehicle by the gaze of the said driver, the generation of the said second indicator being a function, moreover, of the said determined duration.

Advantageously, the said process further includes a step of determining a percentage, over a predetermined time, of the duration of focus of the said target vehicle by the gaze of the said driver, the generation of the said second indicator being a function, moreover, of the said percentage determined.

Thus, the longer, in absolute or relative terms, the driver's attention is focused on the target vehicle, the greater the potential danger that target vehicle poses to the ego vehicle. The second indicator then reflects the importance the driver places on the target vehicle.

Advantageously, said autonomous driving is an adaptive cruise control based on a set inter-vehicle time, the process further comprising a step of determining a set inter-vehicle time as a function of said second probability indicator.

Advantageously, said autonomous driving is an adaptive cruise control based on a set speed, the process further comprising a step of determining a set speed based on said second probability indicator.

Thus, in the case of autonomous driving based on adaptive cruise control, the speed of the ego vehicle is adjusted according to the focus of the driver's gaze.

A second aspect of the invention relates to a device comprising an autonomous driving device, a driver's gaze analysis system and a memory associated with at least one processor configured to implement the method according to the first aspect of the invention.

The invention also relates to a vehicle incorporating the device.

The invention also relates to a computer program comprising instructions which, when the program is executed by the device according to the second aspect of the invention, lead the latter to implement the process according to the first aspect of the invention.

Brief description of the figures

Other features and advantages of the invention will become apparent from the description of the non-limiting embodiments of the invention below, with reference to the attached figures, in which:

FIG. 1 schematically illustrates a device, according to a particular embodiment of the present invention.

FIG. 2 schematically illustrates a method of autonomous driving of a vehicle, according to a particular example of embodiment of the present invention.

Detailed description of the invention

The invention is described below in its non-limiting application to the case of an autonomous motor vehicle traveling on a road or traffic lane. Other applications, such as a robot in a warehouse or a motorcycle on a country road, are also conceivable.

There FIG. 1 represents an example of a device 101 included in the vehicle, in a network ("cloud") or in a server. This device 101 can be used as a centralized device responsible for at least some steps of the process described below with reference to the FIG. 2 In one embodiment, it corresponds to an autonomous driving computer.

In the present invention, the device 101 is included in the vehicle.

This 101 device can take the form of a case containing printed circuits, any type of computer or even a mobile phone (“smartphone”).

The device 101 includes a random access memory 102 for storing instructions for the implementation by a processor 103 of at least one step of the process as described above. The device also includes a mass storage 104 for storing data intended to be retained after the implementation of the process.

Device 101 may further include a digital signal processor (DSP) 105. This DSP 105 receives data to shape, demodulate and amplify, in a manner known per se, this data.

Device 101 also includes an input interface 106 for receiving data implemented by the method according to the invention and an output interface 107 for transmitting data implemented by the method according to the invention.

For example, input interface 106 can receive the following data: vehicle position or geographic location, vehicle speed and/or acceleration, target or predetermined positions/speeds/accelerations, engine speed, position and/or travel of the clutch, brake, and/or accelerator pedals, detection of other vehicles or objects, position or geographic location of other detected vehicles or objects, speed and/or acceleration of other detected vehicles or objects, sensor operating states, confidence level of data from or processed by sensors and/or devices similar to device 101. For example, sensors capable of providing data include: GPS (with or without mapping), tachometers, accelerometers, radar, lidar, lasers, ultrasound, cameras, eye-tracking devices, etc.

In one example implementation, the input interface receives a probability indicator, a vehicle location, an area focused by a driver's gaze (this area could be a lane on which the ego vehicle is traveling, an adjacent lane to the lane on which the ego vehicle is traveling, a target vehicle, etc.), a lane model of a road, a duration or percentage of focus time, a target inter-vehicle time, a target vehicle speed, etc.

For example, output interface 107 can transmit data similar to the data received by input interface 106. Output interface 107 can transmit a focus of the first lane, the second lane, or a target vehicle by a driver's gaze, a probability indicator, an autonomous driving instruction, a duration or percentage of focus time, a target inter-vehicle time, a target speed, etc.

FIG. 2 This schematically illustrates a method for the autonomous driving of a vehicle, according to a particular embodiment of the present invention. The vehicle is called the ego vehicle. The ego vehicle travels on a road, said road comprising at least two lanes, a first lane and a second lane. The ego vehicle travels in the first lane. A vehicle, called the target vehicle, travels in the second lane. The method is implemented by the processor 103.

It is known that an ADAS system, through a perception system combined with a geolocation system and mapping, is capable of determining a road, lanes on that road, and objects on those lanes, an object being a vehicle. This system is also capable of determining whether the target vehicle is traveling in the first lane and whether the target vehicle is traveling in the second lane.

Step 201, RxProb1, is a first probability indicator receiving step, said first indicator characterizing a probability that a future path of said target vehicle intercepts a future path of said ego vehicle.

It is known that an ADAS system, particularly an autonomous driving system like the one in FR1872874, is capable of determining the future trajectory of the ego vehicle and the target vehicle. This trajectory estimation is generally based on an extrapolation of the vehicle positions, measured or determined by sensors. This probability estimates the level of safety and the risk of an accident.

Generally, these trajectories are expressed in the same coordinate system shared between ADAS systems. Based on the estimates, these systems can provide a probability that a future path of the target vehicle will intersect a future path of the ego vehicle.

Step 202, RxLoc, is a step for receiving the location of the target vehicle. Typically, the location is expressed in a coordinate system shared between ADAS systems. If this is not the case, the location is expressed in a coordinate system whose definition (origins, axes) is known by the process, and the process is capable of changing the coordinate system to express the location in a coordinate system known and used by the process.

Step 203, RxZone, is a receiving step of a zone focused by a look from a driver of said ego vehicle, the zones which can be focused by the look being predetermined, said zone being emitted by a driver's look analysis system.

It is known that a gaze analysis system is capable of identifying an area focused by the driver's gaze. This area is identified by an intersection between measured gaze directions and/or a measured position, direction, and/or orientation of the driver's head, and a geographical representation of the regions a driver can see. This area is also expressed in a coordinate system known by the method. Otherwise, the method is capable of performing a change of coordinate system to express the area in a coordinate system known and used by the method.

Step 204, RxMdl, is a step for receiving a model of the first and second channels. Similarly, the process receives this information and is able to perform a change of coordinate system to represent the area in a coordinate system known and used by the process, if this information is not already represented in the same shared coordinate system.

Step 205, DetFocal, is a determination step, based on the target vehicle's location, the focal zone, and the modeling, to determine whether the driver's gaze is focused on the first lane, the second lane, or the target vehicle. Since the target vehicle's location, the focal zone, and the lane modeling are expressed in the same shared coordinate system (or, if not, the process is capable of performing the necessary coordinate system change), by comparison, it is determined whether the driver's attention is focused on the first lane (in which case the focal zone is on a region of the shared coordinate system corresponding to a location on the first lane), on the second lane (in which case the focal zone is on a region of the shared coordinate system corresponding to a location on the second lane), or on the target vehicle (in which case the focal zone is on a region of the shared coordinate system corresponding to a location on the located vehicle).

Step 206, Prob2, is a step that generates a second probability indicator based on the first probability indicator and the determination. Depending on the focused area, the region to which the driver's attention is directed, the second probability indicator is adjusted.

Advantageously, when the driver focuses their gaze on the second lane or the target vehicle, the second probability indicator generated is greater than the first probability indicator. If the driver directs their attention to the target vehicle or to a lane other than the one in which the autonomous vehicle is traveling, there is a potential hazard. Objects, such as the target vehicle, in the driver's line of sight must be taken into account for safe autonomous driving. Since the probability of an accident is higher, the autonomous driving system will adapt the trajectory and speed of the autonomous vehicle in anticipation of potential hazards, based on current best practices.

Advantageously, when the driver focuses their gaze on the first lane, the second generated probability indicator is smaller than the first probability indicator. If the driver does not direct their attention toward the target vehicle or to a lane other than the one in which the ego vehicle is traveling, it means the driver judges there to be little danger and wants to continue in their current direction. Objects, such as the target vehicle, should be less taken into account for safe autonomous driving. Since the probability of an accident is lower, autonomous driving will adapt less, taking less account of the target vehicle, the trajectory, and the speed of the ego vehicle compared to current best practices.

Step 207, TxIns, is a step involving the issuance of an autonomous driving instruction based on the aforementioned second probability indicator. This probability indicator has been updated according to the driver's attention. The vehicle's autonomous driving system will take this probability into account to adapt the trajectory and the vehicle's speed, thus providing the expected benefits.

Advantageously, said process further includes a step of determining a duration of focus of said target vehicle by the gaze of said driver, the generation of said second indicator being a function, moreover, of said determined duration.

Advantageously, the said process further includes a step of determining a percentage, over a predetermined time, of the duration of focus of the said target vehicle by the gaze of the said driver, the generation of the said second indicator being a function, moreover, of the said percentage determined.

Advantageously, said autonomous driving is an adaptive cruise control based on a set inter-vehicle time, the process further comprising a step of determining a set inter-vehicle time as a function of said second probability indicator.

Advantageously, said autonomous driving is an adaptive cruise control based on a set speed, the process further comprising a step of determining a set speed based on said second probability indicator.

The present invention is not limited to the embodiments described above by way of example: it extends to other variants.

For example, the process has been described as a sequence of steps. Some steps can be carried out in parallel or in a different sequence.

Claims (10)

A method for autonomously driving a vehicle, called the ego vehicle, said ego vehicle traveling on a road, said road comprising at least two lanes, a first lane and a second lane, said ego vehicle traveling on said first lane, a vehicle, called the target vehicle, traveling on said second lane, said method being implemented by a processor (103) and comprising the steps of:

• Reception (201) of a first probability indicator, said first indicator characterizing a probability that a future path of said target vehicle intercepts a future path of said ego vehicle;
• Receipt (202) of a location of said target vehicle;
• Reception (203) of a zone focused by a gaze of a driver of said vehicle ego, the zones which can be focused by the gaze being predetermined, said zone being emitted by a driver's gaze analysis system;
• Reception (204) of a model of the said first and second paths;
• Determination (205), based on said location of the target vehicle, said focused area and said modelling, whether said driver focuses his gaze on the first lane or, whether the driver focuses on said second lane or the target vehicle;
• Generation (206) of a second probability indicator as a function of said first probability indicator and said determination;
• Issuance (207) of an autonomous driving instruction from said second probability indicator.

Method according to claim 1, wherein, when the driver focuses his gaze on the second lane or the target vehicle, said second generated probability indicator (206) is greater than said first probability indicator. A method according to any one of the preceding claims, wherein, when the driver focuses his gaze on said first lane, said second generated probability indicator (206) is smaller than said first probability indicator. A method according to any one of the preceding claims, wherein said method further comprises a step of determining a focusing time of said target vehicle by the gaze of said driver, the generation of said second indicator being a function, furthermore, of said determined time. Method according to the preceding claim, wherein said method further comprises a step of determining a percentage, over a predetermined time, of the duration of focus of said target vehicle by the gaze of said driver, the generation of said second indicator being a function, furthermore, of said percentage determined. A method according to any one of the preceding claims, wherein said autonomous driving is an adaptive cruise control based on a setpoint inter-vehicle time, the method further comprising a step of determining a setpoint inter-vehicle time as a function of said second probability indicator. A method according to any one of the preceding claims, wherein said autonomous driving is an adaptive speed regulator based on a set speed, the method further comprising a step of determining a set speed as a function of said second probability indicator. Device (101) comprising an autonomous driving device, a driver's gaze analysis system and a memory (102) associated with at least one processor (103) configured to implement the method according to one of the preceding claims. Vehicle comprising the device according to the preceding claim. Computer program comprising instructions which, when the program is executed by the device (101) according to claim 8, cause the device to implement the method according to any one of claims 1 to 7.