Patent Application
20220194417
The invention relates to a method for the autonomous driving or autonomous control of a motor vehicle. The invention also relates to a system for the autonomous driving or autonomous control of a motor vehicle. The invention also relates to a motor vehicle comprising such an autonomous driving system or comprising hardware and/or software means implementing such an autonomous driving method. The invention further relates to a computer program product comprising program code instructions recorded on a support that can be read by an electronic control unit to implement the steps of the autonomous driving method. The invention also relates to a data recording support that can be read by an electronic control unit and on which the program product is recorded. The invention finally relates to a signal from a data support, bearing the computer program product.
In particular, the invention relates to an accurate method for autonomous driving in order to begin a journey in autonomous mode in a situation of low speed of travel of the vehicle and regardless of the situation surrounding the vehicle. This solution solves a problem of a lack of alignment of the vehicle at the time of starting or in a bend. The solution significantly increases the autonomous nature of the vehicle.
An autonomous motor vehicle is able by itself to manage the trajectory along which it is moving without action on the part of a user of the vehicle, except for indicating to the vehicle an arrival point or an end-of-journey point. To this end, an autonomous vehicle generally comprises an autonomous driving mode, managed by an autonomous driving system, able to guide it.
The autonomous driving mode manages the control of various actuators that allow the vehicle to move along a trajectory. These actuators are chiefly the power unit that drives the motor vehicle and a steering actuator able to adequately orient the steered wheels of the vehicle.
An autonomous vehicle generally comprises an autonomous cruising driving mode, managed by an autonomous cruising driving system, able to guide it at cruising speeds, for example speeds in excess of 30 km/h, or even in excess of 50 km/h. This mode is used in town on main roads and out of town on highways or freeways. The autonomous cruising driving mode is activated only once the autonomous vehicle has already reached a cruising speed. Thus, when the vehicle is starting or pulling away or the vehicle is pulling out (starting from near a kerb for example) or the vehicle is pulling in (parking near a kerb for example) and more generally when maneuvering at low speed, which is to say at speeds below 10 km/h or even below 5 km/h, it is a driver or a remote operator who drives the vehicle.
On-demand mobility services are designed to transport individuals door to door so as to offer the passengers the comfort of being picked up and set down wherever they wish. However, these services require autonomous vehicles capable of managing automated maneuvers involving extremely precise longitudinal and transverse commands, especially during the pulling-in or pulling-out maneuvers, in order to keep the vehicle close enough to the kerb while at the same time avoiding injury to the users or to third parties or material damage to the autonomous vehicle, to other surrounding vehicles or to surrounding infrastructures. These maneuvers are complex because they require precise longitudinal and lateral positioning data at very low speed. These requirements are not met by the autonomous driving systems currently available. The autonomous driving systems are in effect dedicated to extra-urban roadways and can be activated once the vehicle is in circulation. However, when the vehicle is stationary or running at a low speed and the autonomous driving system is activated, the trajectory is poorly managed and there are significant risks of injury or damage as a result of the movement of the vehicle.
For good management of low-speed maneuvers autonomously, precise low-speed control is of the utmost importance. Specifically, during these maneuvers, small steering radii are required whereas the dynamic data and current status of the vehicle are neglected.
Just after activating the autonomous mode for a vehicle circulating at zero or substantially zero speed, the vehicle may be impeded by unforeseen circumstances forcing a remote operator to take over control of the vehicle in order to solve the problem. This problem set is due to an (excessively) approximate estimate of the state of the vehicle in which, for example, the heading, the direction of the wheels or the lateral position error may be neglected. This leads to the calculation of a trajectory which may be unattainable, for example being outside of the capabilities of the vehicle.
In other words, starting from a zero or substantially zero speed and up to a cruising speed, a vehicle is not autonomous. This results in the need to be able to call upon a driver within the vehicle, or even a remote operator, particularly when the vehicle arrives at a platform or kerb (pulls in) to set down and/or pick up passengers, or when the vehicle is starting from a platform or kerb (pulling out) or else for effecting complex maneuvers.
Document U.S. Pat. No. 9,645,577 discloses a guidance system designed to autonomously guide the vehicle in a constrained environment. The system generates various spatiotemporal trajectory solutions as different movement strategies so that the vehicle chooses the one that appears to be optimal. Once the strategy has been determined, the necessary commands for following it are sent to the actuators of the vehicle. One disadvantage with this solution is that, at low speed, it may choose a trajectory that cannot be realized given the current state of certain actuators.
It is an object of the invention to provide an autonomous driving method that overcomes the above disadvantages and improves the driving methods known from the prior art. In particular, the invention makes it possible to produce an autonomous vehicle capable of autonomously managing complex maneuvers performed at low speeds of travel of the vehicle.
In order to achieve this objective, the invention relates to a method for the autonomous maneuvering driving of a motor vehicle, notably a method for the autonomous maneuvering driving of a motor vehicle at low speed or zero speed, comprising a step of modifying the orientation of the steered wheels of the vehicle while the speed of the vehicle is zero or substantially zero or below or equal to a threshold, notably a threshold equal to 1 km/h.
The step of modifying the orientation of the steered wheels of the vehicle can be implemented:
The invention also relates to a method for the autonomous driving of a motor vehicle, comprising:
The operating logic for the autonomous maneuvering driving mode and for the autonomous cruising driving mode may be different; and/or in autonomous maneuvering driving mode the vehicle may be driven or moved at a speed below 10 km/h, or even below 5 km/h.
The method may comprise:
The method may comprise a step of validating the maneuvering trajectory with respect to:
The method may comprise:
The invention further relates to a system for the autonomous driving of a motor vehicle, the system comprising hardware and/or software elements implementing the method as defined hereinabove, notably hardware elements and/or software elements designed to implement the method as defined hereinabove, and/or to a system comprising means for implementing the method as described hereinabove.
The invention further relates to a motor vehicle, the vehicle comprising hardware and/or software elements implementing the method as defined hereinabove, notably hardware elements and/or software elements designed to implement the method as defined hereinabove, and/or to a vehicle comprising means for implementing the method as defined hereinabove.
The invention further relates to a computer program product comprising program code instructions recorded on a support readable by an electronic control unit for implementing the steps of the autonomous driving method as defined hereinabove when said program is operating on an electronic control unit or to a computer program product that can be downloaded from a communications network and/or recorded on a data support that can be read by a computer and/or executed by a computer, comprising instructions which, when the program is executed by the computer, cause the latter to implement the method as defined hereinabove.
The invention further relates to a data recording support, that can be read by an electronic control unit and on which is recorded a computer program comprising program code instructions for implementing the method as defined hereinabove or to a recording support that can be read by a computer comprising instructions which, when executed by a computer, cause the latter to implement the method as defined hereinabove.
The method further relates to a signal from a data support bearing the computer program product as defined hereinabove.
These objects, features and advantages of the present invention will be set out in detail in the following description of one particular embodiment given without implying limitation and in connection with the attached figures among which:
The vehicle 300 comprises an autonomous driving system 200. The autonomous driving system is able to manage an autonomous maneuvering driving mode. This autonomous maneuvering driving mode is able to manage the autonomous movement of the motor vehicle at low speed, notably during complex maneuvers such as pulling-in maneuvers or pulling-out maneuvers. Advantageously, the autonomous driving system is also able to manage a second autonomous driving mode: an autonomous cruising driving mode. This autonomous cruising driving mode is able to manage the autonomous movement of the motor vehicle at high speed, notably during extra-urban travel on highways or freeways or on major traffic routes in an urban environment.
The autonomous driving system 200 comprises:
Advantageously, the autonomous driving system 200 may also comprise a man-machine interface 210. For example, the man-machine interface is able to inform a user, a passenger or a remote operator of the feasibility of autonomously executing a maneuver and/or of the way in which the commands of the actuators of the vehicle are corrected in order to adjust the trajectory.
The set of actuators notably comprises a steering actuator 221 or actuator of the orientation of the steered wheels of the vehicle 300, and a drive unit 222 that powers the vehicle 300. The vehicle drive unit may be of the combustion engine, hybrid drive unit or electric motor type.
The set of elements 220 supplying information regarding the state and surroundings of the vehicle may comprise one or more of the following elements:
One or more of the sensors may be replaced by an element providing equivalent information, for example replaced by an estimator.
The position sensor 14 or a position estimator advantageously provides information as to the location of the vehicle, such as GPS coordinates, and information of the orientation of the vehicle or the heading of the vehicle, such as for example the orientation of the longitudinal axis of the vehicle.
The computer 100 therefore uses as main inputs:
From all of this information it is possible to calculate:
The computer 100 comprises:
As illustrated in
As illustrated in
The trajectory following module 3 and more specifically the command regulator 30 receives at input the signals and/or information which follow:
The motor vehicle, particularly the autonomous driving system and/or the computer, comprise hardware and/or software means able or configured to implement the method that forms the subject matter of the invention.
The hardware and/or software means may comprise software modules.
One mode of execution of an autonomous driving method is described hereinafter with reference to
This mode of execution may also be seen as a mode of execution of a method of operation of the vehicle or of the autonomous driving system. This mode of execution is described here in detail as being a mode of execution of a method of operation of the autonomous driving system.
It is assumed that the vehicle is initially stationary after one or more passengers have met the vehicle and climbed aboard in order to be driven by the vehicle to a destination of their choice. It is also assumed that the vehicle is in an environment in which there are obstacles and which therefore requires a complex maneuver so that it can start or pull away or pull out.
It is also assumed that the vehicle or the autonomous driving system has a reference trajectory or has calculated a reference trajectory. The reference trajectory is a trajectory allowing the vehicle to go from a start point to an arrival point following or substantially following the lines of the traffic lanes. Thus, when a vehicle follows a reference trajectory, it is centered or substantially centered on the traffic lane. Such a trajectory is incompatible with a vehicle state in which the vehicle is parked near a kerb, in an interior courtyard or in a parking lot, which is to say that, in such situations, the position (location and heading) of the vehicle does not correspond to any of the positions defined by the reference trajectory.
In a first phase 110, the autonomous driving system operates in an autonomous maneuvering driving mode and simulates the state of the vehicle, vehicle actuator commands and a maneuvering trajectory.
The maneuvering trajectory is a vehicle trajectory that allows the vehicle to meet the reference trajectory from its initial state, for example a state of being parked near a kerb, in an interior courtyard or in a parking lot.
As illustrated in
Steps of a mode of execution of this first phase 110 are described in greater detail hereinafter.
In a first step, the actuator command simulation module 2, particularly the vehicle modelling module 4, receives as input all or part of the following information:
In a second step, the reference trajectory is used to virtually model the vehicle using the vehicle modelling module 4 by giving the vehicle a minimal speed but without causing the position of the vehicle to move virtually. What that means is that the modelled virtual vehicle is still in the same virtual position (for example having identical longitude, latitude and heading values).
In a third step, the command modelling module 5 corrects the steered wheels orientation error that there is between the current orientation of the steered wheels and the orientation necessary for following the reference trajectory. In consequence, the command modelling module 5 generates a command for controlling the vehicle steered wheels orientation actuator.
In a fourth step, this command generated in the third step is supplied to the steering modelling module 6 which, on this basis, is able to generate a future change to the position of the virtual vehicle in order to verify the future values of the positions (longitude, latitude and heading) of the vehicle with respect to the limits of the driving area available in the vicinity of the vehicle. A verification is then made as to whether the last command generated leads to a virtual orientation of the steered wheels that allows a desired point on the reference trajectory to be attained or whether the command generated exceeds the physical capabilities of the vehicle in terms of vehicle wheel orientation angle. If the check is not ok, the method loops back to the third step and the third and fourth steps are repeated iteratively until:
Thus, the method comprises a step of validating the maneuvering trajectory against the physical maneuvering capabilities of the vehicle.
In the first scenario, the method moves on to a second phase 120. The last value for the command of the actuator for orienting the wheels of the vehicle is retained.
In the second scenario, the autonomous driving system has not found an actuator command that allows the maneuver to be performed. In consequence, the autonomous driving system informs a user of the vehicle or a remote operator of this fact. This information is communicated for example via the man-machine interface 210.
Thus, in this first phase, the method comprises:
In the second phase 120, the autonomous driving system tests whether the complex maneuver can be realized in autonomous mode or whether the maneuver can be realized with a sufficient level of safety to the individuals and assets in the vehicle and in the external environment in the vicinity of the vehicle. If it can, the autonomous driving system implements a phase 130. If it cannot, the autonomous driving system implements a phase 140. Thus, in this second phase 120, the autonomous driving system evaluates the vehicle and the actuator commands determined previously (by the command simulation module 2) to generate the virtual trajectory that the vehicle is to cover. The virtual trajectory is verified by the pull-away or start validation module 1 to ensure a safe start (for example that the foreseen future positions of the vehicle will lie inside the driving area). To do this, a vehicle speed is used to simulate the future changes to the position of the virtual vehicle during the course of the maneuver. This speed can be parameterized. It is for example fixed at 2 m/s.
Advantageously, the autonomous driving system may command for certain ones of the iterations or all of the iterations (until a maneuver that can be realized is or is not found) to be displayed and/or illustrated on the man-machine interface. As a preference, the autonomous driving system may command for the man-machine interface to display information relating to the performance of the autonomous driving system, for example a simultaneous display of the last virtual trajectory adopted and of the reference trajectory supplied by the module 15 or a simultaneous display of the last calculated initial orientation of the steered wheels of the vehicle and of the current orientation of the steered wheels of the vehicle. As a preference, the autonomous driving system may command for the man-machine interface to display:
Thus, the method comprises a step of validating the maneuver trajectory with respect to the safety of the assets and/or individuals in the vehicle and/or in the vicinity of the vehicle.
In the phase 130, the autonomous driving system generates and executes the actual commands for controlling the actuators of the vehicle so as to execute the complex maneuver. When a validation indicator indicates that a safe start maneuver is possible, the autonomous driving system activates the trajectory following module 3. This module receives the virtual commands determined by the command simulation module 2 and allowing the maneuver to be executed according to the determined virtual trajectory. The validation indicator thus stops the iterative calculations of the vehicle actuator command simulation module 2, adopting the last simulated actuator command values. Next, a three-step procedure is implemented by the trajectory following module 3:
Such a procedure is well accepted by users and does not catch them by surprise.
Steps (already mentioned hereinabove) of a mode of execution of this third phase 130 are described in greater detail hereinafter.
In a first step, the trajectory following module 3, particularly the command regulator 30, receives at input all or some of the following information:
This information contains a value for the distance separating the reference trajectory from the center of gravity of the vehicle.
All of these inputs allow the regulator 30 to activate the various command actions according to the state of the vehicle.
Thus, in the first step of the third phase 130, the initial direction command correction module 37 has the task of minimizing the steering command model error value 34. It imposes a zero or substantially zero speed on the vehicle for as long as the orientation of the steered wheels of the vehicle are being corrected in order to achieve the orientation determined at the end of the first phase. The method therefore comprises a step of modifying the orientation of the steered wheels of the vehicle while the speed of the vehicle is:
This step of modifying the orientation of the steered wheels of the vehicle is implemented:
Advantageously, once the value of the error 34 is below a predetermined threshold, the regulator 30 deactivates the module 37.
Next, in the second step of the third phase 130, the regulator 30 activates the position correction module 38. This module 38 fixes a given speed (for example 2 m/s and/or for example identical to the speed used in the first phase 110). In consequence, the vehicle moves actually driven by its drive unit so as to correct the position errors, particularly so as to correct the heading error and the lateral position error that there are with respect to the point of the reference trajectory to be met. The maneuver is therefore executed in autonomous mode. In autonomous maneuvering driving mode, the vehicle is preferably driven or moved at a speed below 10 km/h, or even below 5 km/h. The vehicle position error values are preferably constantly measured and/or estimated and then compared against their predetermined respective thresholds (typically 30 cm in the case of the lateral position error and 0.1 radian in the case of the heading error). When all the position error values are below their predefined thresholds, the regulator 30 deactivates the module 38. The maneuver is completed.
Next, in the third step of the third phase 130, the regulator 30 activates the autonomous cruising driving mode module 39.
The method therefore comprises:
Thus, the autonomous driving method allows the automatic, and/or without user action, transition from an autonomous maneuvering driving mode to an autonomous cruising driving mode. The method therefore comprises:
As a preference,
In a phase 150, the autonomous driving system tests whether the vehicle is on a trajectory defined by an autonomous cruising driving mode, and if it is, the autonomous driving system exits the autonomous maneuvering driving mode and automatically switches over to this autonomous cruising driving mode.
In the phase 140, the autonomous driving system has not found an actuator command that allows the maneuver to be performed with satisfactory criteria. In consequence, the autonomous driving system informs the vehicle user or an operator of this fact. This information is communicated for example via the man-machine interface. The control of the actuators of the vehicle is then performed by the user of the vehicle or by the operator, notably a remote operator, in order to realize the maneuver.
In a phase 160, once the maneuver has been performed and with the vehicle in motion, the autonomous cruising driving mode can be activated. This autonomous cruising driving mode is activated by action on the part of the user or of the operator.
Thus, the autonomous driving system is capable of recognizing and taking into account misalignments between the position of the vehicle, the state of an actuator and the desired trajectory before beginning to execute any maneuver. The autonomous driving system commands and modifies the response of the vehicle (if need be) when the autonomous maneuvering mode is engaged and/or the speed is low so as to detect and correct dangerous situations when the vehicle is starting while its state is not consistent with the trajectory. The autonomous driving system continuously monitors the state of the vehicle and compares it against a reference trajectory in order to determine the feasibility thereof, modifying the actuator control commands accordingly in order to improve the following of the trajectory and reduce the risks.
The above-described embodiment of the autonomous driving system was coded and integrated into a vehicle belonging to the applicant (ZOE® robot taxi) and tests yielded the following results.
The system was tested on track with different initial states and different orientations of the vehicle. The results were compared against those obtained with the same vehicle without implementing the invention. These results demonstrate a significant improvement in the autonomy capabilities of the vehicle.
It may be noted that kerb B immediately leaves the lane. The trajectory remains out of the lane for a long moment before returning to follow the desired reference trajectory (to make a U-turn around the roundabout). By contrast, curve A remains in lane. Specifically, with the solution according to the invention, the steered wheels of the motor vehicle are oriented first of all before the motor vehicle is moved. This allows the vehicle to enter the roundabout correctly starting from the same initial configuration.
It is important to note that the good stability capabilities of the autonomous driving system never put the vehicle in unstable situations. The significant lateral position and heading errors are due to the autonomous driving mode being initialized with the vehicle in a non-optimal state and the associated actuator data.
The second test was obtained in the context of a bend. The vehicle starting conditions for the two tests are the same, with a heading error of around 40 degrees with respect to the desired reference trajectory and a steering wheel angle in full left lock (around 480 degrees).
In this case, although both trajectories are depicted, only trajectory C can be realized without exceeding the limits (notably without occupying the lane for the oncoming traffic). Trajectory C is executed by the vehicle equipped for implementing the invention. Here, trajectory C rapidly converges towards the desired reference trajectory, without straddling the oncoming lane and bringing the vehicle into normal behavior after around 11 seconds. By contrast, trajectory D represents a vehicle that is not implementing the invention and which stabilizes after around 15 seconds following the start and having clearly straddled the other lane, in other words having departed from the fixed limits.
Finally,
A general illustration of an example of a man-machine interface 210 has been depicted in
On the interface 210, the trajectory that the vehicle will follow with the initial position and the initial steering wheel angle (trajectory A′ depicted on a map with the desired trajectory TR′ and the driving area) is compared against the trajectory B′ that the vehicle would follow without implementing the invention. As can be seen, the virtual vehicle corrects the orientation of the steered wheels of the vehicle and then the vehicle is moved. The angular position of the virtual steering wheel can be indicated on the interface 210, as can the current angular position of the steering wheel.
If the trajectory verified by the start validation module 1 is feasible, an indication 211, such as a lamp 211, may be displayed on the interface.
If the trajectory verified by the start validation module 1 is not feasible, an indication 212, such as a lamp 212, may be displayed on the interface.
If several tests are executed without any possible trajectory being found, an indication 213, such as a lamp 213, may be displayed on the interface.
The interface is therefore able to indicate how the vehicle is going to manage a complex maneuver situation autonomously.
The advantage of the solutions described hereinabove is that they increase the capabilities of the vehicles, notably of the robot taxi type. Such vehicles have difficulties starting in curved zones or where there is an appreciable lateral position or orientation error with respect to a reference trajectory supplied by an autonomous cruising driving system. The solutions described above also enable management of a pulling-in, pulling-out or remote operated maneuver which represent maneuvers that are complex for an autonomous vehicle to manage.
| Number | Date | Country | Kind |
|---|---|---|---|
| FR1909974 | Sep 2019 | FR | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2020/073484 | 8/21/2020 | WO | 00 |
