FR3067683A1 - ARCHITECTURE PROPOSAL FOR THE SAFETY OF VEHICLE FUNCTIONS ADAPTED TO AUTONOMOUS VEHICLE. - Google Patents

Abstract

The invention relates to a method for securing an autonomous vehicle (2) traveling on a taxiway, said method comprising a step of detecting at least one obstacle (1) on the taxiway, a step of determining , from the at least one detected obstacle (1), a free space (3) on the taxiway comprising no detected obstacle (1), a step of determining the longitudinal and lateral controllability of the vehicle (2) on the taxiway, and a step of controlling the longitudinal and lateral trajectory of the vehicle (2), according to the determined longitudinal and lateral controllability, so as to maintain said vehicle (2) in space free determined.

Description

METHOD FOR SECURING A SELF-CONTAINED VEHICLE [001] The invention relates, in general, to autonomous driving, in particular of a motor vehicle.

The invention relates more particularly to a method for securing an autonomous vehicle.

In a known manner, a motor vehicle includes driving aids, such as ESP, for Electronic Stability Program in English, which is a trajectory corrector of the vehicle. Certain vehicle components are also designed to help the driver, in particular the GMP, for MotoPropulseur Group, allowing in particular a traction and shifting aid, or the DAE, for electric power steering, helping the driver to maneuver the vehicle.

[004] Such aids thus make it possible to assist the driver when using the vehicle. However, these aids are not suitable in the case of autonomous driving of the vehicle. Autonomous vehicles must comply with standards in order to guarantee the safety of passengers in the autonomous vehicle. Such standards aim in particular to impose a safety level, designated ASIL for Automotive Safety Integrity Level in English, for each function of the vehicle. Driving aids do not offer a sufficient level of safety.

Also, it is known from document WO 2016162624 to integrate into a vehicle a function, such as advanced driving assistance, designated ADAS for Advanced Driving Assistance System in English, which requires a high level of security. The function is fulfilled by two redundant computers performing the same calculations in order to allow the proper functioning of ADAS. Indeed, in the event of a problem with one of the computers, the calculations are always carried out by the other computer, which allows the use of ADAS. Such computers have a lower level of security. Thus, it is possible to obtain a high level of security with computers whose security level is lower by doubling the calculations.

However, such a solution has drawbacks. Indeed, since the ADAS requires carrying out important calculations, the computer must be powerful and is therefore expensive. Also, the use of two redundant computers further increases costs.

In addition, the use of two computers each performing the same calculations increases the complexity of the vehicle and therefore its cost.

The invention therefore aims to solve these drawbacks by proposing an easy-to-install, reliable and inexpensive system allowing the management of the operation of an autonomous vehicle requiring a high level of security.

To achieve this result, the present invention relates to a method for securing an autonomous vehicle traveling on a traffic lane, said method comprising a step of detecting at least one obstacle on the lane of traffic, a step determination, from the at least one detected obstacle, of a free space on the taxiway not comprising any detected obstacle, a step of determining the longitudinal and lateral controllability of the vehicle on the taxiway, and a step of controlling the longitudinal and lateral trajectory of the vehicle, as a function of the determined longitudinal and lateral controllability, so as to maintain said vehicle in the determined free space.

Thanks to the method according to the invention, a motor vehicle can be controlled in autonomous mode by elements having a low level of safety, in particular of level ASIL B or level ASIL C, which makes it easy to integrate of such an autonomous mode in a motor vehicle while limiting the cost of such integration. In addition, the autonomous operation of the vehicle has a high level of security thanks to the determination of a free space and the control of the vehicle in this free space, which makes it possible to limit the control of the vehicle to the "perception" part the vehicle environment achieved by sensors. The vehicle is thus kept away from obstacles, which limits the use of the "decision" part requiring a higher level of security.

Advantageously, the vehicle comprises at least one sensor from the following list of sensors: camera, radar, lidar system, said detection step being carried out by said at least one sensor.

Advantageously, the longitudinal and lateral controllability of the vehicle is determined from a set of initial conditions and a set of capacities of the vehicle to move longitudinally and laterally.

Advantageously, said initial conditions include the initial position of the vehicle on the traffic lane and / or the initial speed of the vehicle.

Advantageously, said vehicle displacement capacities include the acceleration, braking and / or lateral displacement capacities of the vehicle.

Advantageously, the trajectory control step comprises a sub-step of limiting the lateral movement of the vehicle, preferably of the order of 4 m, when changing the lane of the vehicle.

The invention also relates to a motor vehicle comprising at least one sensor adapted to detect at least one obstacle and a computer adapted to control the trajectory of the motor vehicle in autonomous mode, said computer being configured to implement the method as previously described.

Other characteristics and advantages of the invention will appear on reading the detailed description of the embodiments of the invention, given by way of example only, and with reference to the drawing which shows:

• Figure 1, a diagram of a step in the method according to the invention of securing an autonomous vehicle.

In what follows, the embodiments described relate more particularly to an implementation of the method according to the invention in a motor vehicle. However, any implementation in a different context, in particular in any type of vehicle, is also covered by the present invention.

A motor vehicle 2 is, conventionally, adapted to travel on a traffic lane. To do this, a driver controls the longitudinal movement of vehicle 2, in other words the speed of vehicle 2, using brake and accelerator controls. The driver also controls the lateral movement of vehicle 2, in other words when vehicle 2 turns to the left or to the right, using the steering.

The motor vehicle 2 shown in Figure 1 also includes an autonomous mode of operation.

The autonomous operation of a motor vehicle 2 is regulated by standards aimed in particular at imposing a minimum level of security on vehicle functions. The safety levels are designated ASIL for Automotive Safety Integrity Level in English. The security levels include ASIL A, ASIL B, ASIL C, ASIL D, in ascending order of security.

The security level is assigned on the basis of various elements, such as, for example, the occurrence of a dangerous situation, its frequency, the consequences of damage linked to such a situation, the ability to manage the dangerous situation, etc. Also, the more a function has a high level of security, for example an ASIL D level, the more complex and expensive the function will be to implement in order to minimize the risks. In general, the autonomous operation of a vehicle requires an ASIL D security level to guarantee maximum safety for the occupants of the vehicle in the event of dangerous situations.

In practice, the autonomous operation of the motor vehicle 2 is generally controlled by a system comprising two parts: a part called "perception" and a part called "decision". The perception part allows the system to identify various elements present around the vehicle, such as obstacles to avoid or the traffic lane to follow. The decision part allows the system to control the movement of the vehicle from the perceived elements in order to follow the traffic lane while avoiding obstacles. The perception part can be carried out by physical elements, such as sensors, while the decision part is carried out by a computer program, in particular using artificial intelligence. The decision part notably involves the implementation of complex "deep learning" algorithms, also known as "deep learning" in English.

The invention aims to limit the use of the decision part which requires a higher level of ASIL security than the perception part and is therefore more complex and costly to implement.

To do this, the vehicle 2 comprises means for controlling the trajectory of the vehicle 2 adapted to control the longitudinal and lateral movement of the vehicle 2 independently, in other words, without driver intervention.

The control means are adapted to determine the controllability of the motor vehicle 2, in other words the position that the vehicle 2 can reach after a determined time from a set of initial conditions and a set of vehicle capacities. 2.

In particular, from an initial position and an initial speed of the vehicle 2 and from the acceleration, braking and lateral displacement capacities of the vehicle 2, the control means are adapted to determine the different positions that the motor vehicle 2 can reach after a predetermined time, in particular of the order of a few seconds.

The determination of the controllability of the vehicle 2 has a lower level of security, for example of level ASIL C, preferably of level ASIL B, because it presents lower risks for the occupants of vehicle 2. Also, the control means are easy to implement in an autonomous vehicle 2 and have a limited cost.

The motor vehicle 2 also includes various sensors adapted to detect obstacles 1 in the environment of the motor vehicle 2 in order to assist the driver while driving.

Such sensors may in particular comprise cameras placed at the front and / or at the rear of the vehicle 2 in order to capture images of the traffic lane respectively in front of and / or behind the vehicle 2 in order to detect therein obstacles 1 using image processing software.

The sensors can also include radars adapted to emit radio waves around the vehicle 2 in order to detect obstacles 1 in the manner of a sonar.

Finally, the sensors can further comprise a lidar system, for Light Detection And Ranging in the English language, designating a remote sensing system by laser emitting light around the vehicle 2 in order to detect obstacles 1.

Such sensors are known and their operation will not be described in more detail.

The vehicle 2 includes a computer adapted to receive data from the various sensors.

Referring to Figure 1, the computer according to the invention is adapted to detect the various obstacles 1 present around the motor vehicle 2 from the computer data and to determine an area 3 on the traffic lane not comprising of obstacles 1. Such a zone 3 is designated “free space”.

Advantageously, the determination of the free space 3 has a security level of ASIL C level, preferably of ASIL B level, since it presents lower risks for the occupants of vehicle 2. This allows a computer to motor vehicle 2 to easily implement and for a limited cost such determination of a free space 3.

In autonomous mode, the control means are adapted to control the movement of the motor vehicle 2 so as to maintain the vehicle 2 in the free space 3 from the determined controllability of the vehicle 2. In other words, among the different positions that vehicle 2 can reach thanks to its controllability, the control means will control the trajectory of vehicle 2 to a position included in free space 3.

Thus, the motor vehicle 2 is kept away from obstacles 1, which avoids risky situations requiring a high level of security, ASIL D level for example. Indeed, such an order for the vehicle 2 is carried out from the perception part without taking into account the decision part, which makes it possible to limit the complexity of such an order.

In the following description, the method for securing an autonomous vehicle 2 will be presented.

A motor vehicle 2 travels on a traffic lane.

When using the motor vehicle 2 in autonomous mode, the sensors detect obstacles 1 present on the traffic lane around the motor vehicle 2.

Then, the motor vehicle computer 2 receives the data from the sensors and merges them in order to determine a free space 3 in which the motor vehicle 2 is located. Such a fusion of the data of several sensors is designated “heterogeneous fusion” and optimizes obstacle detection 1.

The controllability of the motor vehicle 2, in other words the ability or capacity of the motor vehicle 2 to move on the lane, is also determined.

Finally, the control means control the longitudinal and / or 5 lateral movement of the motor vehicle 2 from the controllability determined so as to maintain the vehicle 2 in the determined free space 3. The control means control the acceleration of the engine, brakes and steering of the vehicle to control the path of the vehicle 2.

The motor vehicle 2 can move on a road comprising a plurality 10 of traffic lanes separated from each other by white lines. The sensors are adapted to detect white lines in order to allow the control means to control a lateral movement of the motor vehicle 2 initially traveling in a lane to a target lane located next to it.

Advantageously, the traffic lanes have a minimum width of 15 3.5 m. Also, the control means are adapted to limit the lateral movement of the motor vehicle to 2 to 4 m, in a predetermined time of the order of 1 s for example, during a change of lane. Thus, in the event of an error in detecting the lines of the target taxiway in which the motor vehicle 2 is moving, the vehicle 2 is kept in the target taxiway by limiting its lateral movement corresponding to the width d 'a traffic lane.

Such limitation of lateral movement in the event of erroneous detection of the lines has a level of security at ASIL B level, which allows easy integration and for limited costs of such a function.

Claims (7)

1. Method for securing an autonomous vehicle (2) traveling on a traffic lane, said method comprising: - a step of detecting at least one obstacle (1) on the taxiway, a step of determining, from at least one detected obstacle (1), of a free space (3) on the traffic lane comprising no detected obstacle (1), a step of determining the longitudinal and lateral controllability of the vehicle (2) on the taxiway, and - a step of controlling the longitudinal and lateral trajectory of the vehicle (2), as a function of the determined longitudinal and lateral controllability, so as to maintain said vehicle in the determined free space (3). 2. A security method according to claim 1, in which the vehicle (2) comprising at least one sensor from the following list of sensors: camera, radar, lidar system, said detection step being performed by said at least one sensor. 3. Method according to one of the preceding claims, in which the longitudinal and lateral controllability of the vehicle (2) is determined from a set of initial conditions and a set of capacities of the vehicle (2) to move longitudinally. and laterally. 4. Method according to the preceding claim, wherein said initial conditions include the initial position of the vehicle (2) on the traffic lane and / or the initial speed of the vehicle (2). 5. Method according to one of claims 3 to 4, wherein said vehicle displacement capacities (2) include the acceleration, braking and / or lateral displacement capacities of the vehicle (2). 6. Method according to one of the preceding claims, in which the step of controlling the trajectory comprises a sub-step of limiting the lateral movement of the vehicle (2), preferably of the order of 4 m, when a change in the lane of the vehicle (2). 7. Motor vehicle (2) comprising at least one sensor adapted to detect at least one obstacle (1) and a computer adapted to control the trajectory of the motor vehicle (2) in autonomous mode, said computer being configured to implement the method according to one of the preceding claims.