Control systems and methods tuned to perception

Abstract

Disclosed is a system for controlling a vehicle (1), the vehicle implementing at least one control application using variables measured by at least one sensor (20) of a perception system (2) installed in the vehicle (1). The control system includes an adaptive controller (112) configured to dynamically activate one or more basic controllers of a set of basic controllers including at least two basic controllers, each basic controller configured to act on an actuator of the vehicle with a control function of a vehicle parameter, the control function being distinct based on an accuracy index of the perception system determined according to real-time values ​​of the variables.

Description

Technical Field

[0001] The present invention relates generally to control systems for controlling movable objects, and more particularly to control systems and methods for controlling vehicles.

[0002] Vehicle automation has experienced rapid expansion in recent years to improve safety and optimize driving for vehicles.

[0003] Automated vehicles, for example autonomous connected vehicles, have conventionally employed a perception system including a set of sensors arranged on the vehicle for detecting environmental information used by a control device to control the vehicle, such as an adaptive cruise control (ACC) distance adjustment radar system used, for example, to adjust the distance between vehicles.

[0004] The performance of a perception system is closely related to the distance of a detected obstacle. The farther away an obstacle is, the more noisy the information associated with the detected object is.

[0005] An automated vehicle (referred to as an "ego vehicle") may use a camera-type sensor, for example, to detect vehicles ahead of the ego vehicle on the road on which the ego vehicle is traveling. The control device may implement object detection and tracking algorithms to provide control over the vehicle.

[0006] In vehicles using such control devices, two parameters generally affect the performance of the perception system: - the type of object detection and tracking algorithm implemented in the vehicle, which results in more or less accurate measurements, and - the physical limitations of the sensors used for detection.

[0007] For example, with camera-type sensors, physical limitations can be related to the number of pixels, which limits image resolution so that detection accuracy decreases as objects are farther from the vehicle. The result is less accurate measurements that degrade the vehicle's performance. The vehicle's response to such measurements is further affected. This is because existing solutions do not allow for the existence of a single control device, such as a vehicle tracking system using cameras, that has the ability to manage all distances of vehicles detected in the vehicle's environment.

[0008] Figure 1 is a conventional graph showing the speed of the vehicle ahead 3, where the speed is measured in ground reality (line C1 shows the actual speed of the vehicle ahead) and by a perception system installed in the vehicle itself (curve C2 shows the estimated speed of the vehicle ahead by the perception system 2). In this example, the speed of the vehicle ahead is controlled to obtain a given response from the vehicle itself. Four continuous speed changes can be observed in the graph in Figure 1. When the vehicle itself, equipped with the installed perception system, follows the vehicle ahead at a predetermined distance, this distance is determined by the speed such that the distance between vehicles increases as the speed increases. Furthermore, the inaccuracy of the perception system increases with increasing speed, which worsens the measurement.

[0009] Solutions exist to correct the inaccuracies of perceptual systems. Known solutions include filtering perceptual signals at the output of the perceptual system. However, such solutions cannot be used to guarantee stability with respect to control responses. Moreover, while these solutions can be used to smooth the output of the perceptual system, it is not possible to establish a link between the overall system performance and filtering.

[0010] A solution to minimize errors in inter-vehicle distance is proposed in CN101417655B. CN101417655B describes adaptive cruise control (ACC), where vehicles are coordinated by the ACC system's multi-target tracking capabilities. Inter-vehicle distance errors are minimized by taking into account the vehicles' fuel consumption, as well as the maximum and minimum acceleration / deceleration parameters acceptable to the driver. However, this solution does not take noise reduction capabilities into account, and thus reduces the operating range in which sensors can detect their environment in an optimal manner.

[0011] U.S. Patent No. 9,266,533,B2 describes the use of a learning system to simulate human responses in an ACC system. However, this solution does not take into account noise reduction capabilities, and therefore its performance is contingent on the perception system having little to no noise.

[0012] Known solutions to address noise problems in perceptual systems include the use of communication devices, as described in U.S. Patent No. 8352112B2. However, such devices are expensive, involve signal loss, and do not take noise into account in the controller design.

[0013] U.S. Patent No. 10310509B1 describes a system capable of detecting degradation of a lidar-type sensor by evaluating newly acquired laser spots and comparing them to previously stored spots. A signal is provided indicating whether the sensor has been degraded based on the comparison. Such information, however, is not used for real-time adaptation of the behavior of an automated vehicle in relation to potential noise in the measurement.

[0014] U.S. Patent Application Publication 2019 / 0189104A1 discloses an ACC system for reducing noise caused by a vehicle's braking system by using a braking model and possible noise levels that the actuator may generate with each speed change. This noise is eliminated to improve the ACC system's response to braking. However, this solution only takes into account the noise identified in the braking actuator and does not take into account noise related to the perceptual system, which is larger and affects the ACC system to a greater extent. Therefore, such a solution cannot be adapted to deal with disturbances originating from the perceptual system.

[0015] U.S. Patent Application Publication 2013 / 0197736A1 describes a behavioral system that can adapt to perceptual uncertainty. Such a solution takes into account sensor noise in various measurements taken to classify objects detected on a road. Uncertainty calculation is performed to indicate various vehicle maneuvers, such as vehicle movement, in order to avoid obstacles, to better determine the dimensions of those obstacles, to lead to better classification and estimation of state variables. However, in this solution, noise does not change the vehicle's performance (control response), and noise is not taken into account in lateral or longitudinal action.

[0016] Thus, there is a need for control methods and devices that have the ability to adapt the vehicle's response according to the level of accuracy of the perceptual system. [Overview of the project]

[0017] The present invention improves the situation by proposing a control system for controlling a vehicle, wherein the vehicle implements at least one control application using magnitude measured by at least one sensor of a perception system installed on the vehicle. Advantageously, the system includes an adaptive controller configured to dynamically activate one or more basic controllers from a set of basic controllers, each basic controller being configured to use a control function for controlling vehicle parameters by acting on the vehicle's actuators, the control functions being distinct based on an accuracy index of the perception system determined according to real-time values ​​of magnitude.

[0018] In one embodiment, the control system further includes a compensator configured to generate a permutation signal, the value of which varies according to an accuracy index.

[0019] The accuracy signal value can vary between a first value representing the optimal level of accuracy of the perceptual system and a second value representing the minimum level of accuracy of the perceptual system.

[0020] In particular, the first value may be equal to 1, and the second value may be equal to 0.

[0021] In one embodiment, the basic controller may include a first basic controller and a second basic controller, wherein the substitution signal can be updated to a first signal value if the accuracy index has a first value, and the first basic controller is fully activated in response to the update of the substitution signal.

[0022] In one embodiment, if the accuracy index of the perceptual system has a second value, the substitution signal can be updated to a second signal value, and the second basic controller is fully activated in response to the update of the substitution signal.

[0023] The magnitude may be a magnitude selected from among the vehicle speed, the inter-vehicle distance between the vehicle and a head vehicle, and the yaw rate of the vehicle.

[0024] In one embodiment, the first controller may be a vehicle tracking controller, and the second controller may be a denoising controller.

[0025] In a certain embodiment, the control application may use a plurality of magnitudes measured by at least two sensors of the perception system, the control device includes a switch for selecting a magnitude from among the magnitudes, and the adaptive controller is applied to the selected magnitude by using an activated basic controller.

[0026] A control method for controlling a vehicle, which is executed in the vehicle, the vehicle implements at least one control application that uses a magnitude measured by at least one sensor of a perception system installed on the vehicle, is further proposed. Advantageously, the method comprises: - selecting at least one basic controller from a set of basic controllers comprising at least two basic controllers, each basic controller being configured to use a control function for controlling a parameter of the vehicle by acting on an actuator of the vehicle, the control functions being distinct based on an accuracy indicator of the perception system determined according to a real-time value of the magnitude. - activating the at least one selected basic controller The method comprises a step including the above steps.

[0027] An embodiment of the present invention thus provides a control device, which has the capability to adapt the response of the control device to the capability of a perception system and change the response of a vehicle in accordance with the level of accuracy of the perception system. The capability of a control system in an advanced driving assistance system for automobiles (e.g., ADAS) or in an autonomous vehicle can thus be improved.

[0028] By controlling performance limitations that may arise between the perception system and the control system of a vehicle, the performance of the vehicle can thus be improved.

[0029] Embodiments of the present invention can be used in particular to adapt the response of a vehicle to perception performance with respect to the total distance between the vehicle and a detected object, so as to optimize the performance of the vehicle by using the maximum capability of the vehicle whatever the situation of the vehicle is.

[0030] Other features, details and advantages of the present invention will become apparent upon reading the description provided with reference to the accompanying drawings, which are given by way of example, and the respective contents of the drawings are as follows. Brief Description of the Drawings

[0031] [Figure 1] It is a diagram showing a graph illustrating the speeds of two successive vehicles according to the prior art. [Figure 2] It is a diagram showing a control system installed in a vehicle according to some embodiments. [Figure 3] It is a diagram showing an example of a traffic environment adapted for executing a control application for controlling a vehicle according to a certain embodiment. [Figure 4] It is a diagram showing the architecture of a control device implemented in a vehicle for controlling the driving of the vehicle according to a certain embodiment. [Figure 5] It is a diagram showing a control system according to a certain embodiment of the present invention. [Figure 6] This figure shows an example of the realization of an adaptive controller in the form of a circuit that performs closed-loop control according to a specific embodiment. [Figure 7] This figure shows an example of the realization of an adaptive controller in the form of a circuit that performs closed-loop control according to a specific embodiment. [Figure 8] This figure illustrates the performance of a control device according to a specific embodiment of the present invention. [Figure 9] This figure illustrates the transition of the substitution signal corresponding to the embodiment. [Modes for carrying out the invention]

[0032] Figure 2 shows a control system 100 installed in a vehicle 1 (referred to as "the vehicle"). The control system 100 includes a control device 11 and a perception system 2, which are installed on the vehicle. The control system 11 may be a driver assistance system (such as an ADAS or AD system). The perception system 2 includes one or more sensors 20 arranged and configured in the vehicle 1 to measure the size of the vehicle and / or the vehicle's environment. The control system 100 uses the information provided by the perception system 2 of the vehicle 1 to control the operation of the vehicle 1.

[0033] The control device 11 may be configured to assist the driver in performing complex driving or steering maneuvers, detecting and avoiding dangerous situations, and / or limiting the impact of such situations on the vehicle 1. The vehicle 1 can detect the environment outside the vehicle thanks to data received from the sensors 20 of the perception system 2, and based on that data, the vehicle 1 can build and update an internal model of the environment's configuration.

[0034] The control device 11 implements at least one control application 110 that uses magnitude G measured by at least one sensor 20 of the perception system 2. The control system 100 is advantageously configured to adapt the response of the vehicle 1 according to the level of accuracy of the perception system 2.

[0035] The control application 110 may be a vehicle tracking application configured to control the speed of vehicle 1 according to, for example, the speed of the vehicle in front, and / or the speed of the vehicle behind.

[0036] The control application 110 may be, for example, an ACC (Adaptive Cruise Control) application. An ACC type application may use a radar system that uses a radar sensor 20 to estimate the speed difference between two successive vehicles, and the radar is configured to be positioned, for example, in front of the vehicle 1.

[0037] The control device 11 advantageously includes an adaptive controller 112 configured to dynamically activate one or more basic controllers from a set of basic controllers 1120, which include at least two basic controllers. Each basic controller 1120 is configured to use a control function to control parameters by acting on the vehicle's actuators (not shown). The control functions of the various basic controllers 1120 may be distinct. The control function is determined according to an accuracy index of the perception system, which is used by the control application 110 and is determined according to a real-time value of the measured magnitude G.

[0038] The sensors 20 of the perception system 2 may include, for example and without limitation, one or more LiDAR (laser detection and ranging) sensors, one or more radar systems, one or more cameras which may be cameras operating in the visible range and / or infrared range, one or more ultrasonic sensors, one or more steering wheel angle sensors, one or more wheel speed sensors, one or more braking pressure sensors, one or more yaw speed and lateral acceleration sensors, and others.

[0039] The perception system 2 may be configured to detect and / or identify objects in the environment of the vehicle 1, as well as pedestrians, vehicles, and / or road infrastructure, based on information measured by the sensor 20. Objects in the environment of the vehicle may include fixed or movable objects. Examples of objects in the environment of the vehicle include, without limitation, vertical objects (e.g., traffic lights, road signs, etc.) and / or horizontal objects (e.g., road markings, stop or yield lines, and other signal markings).

[0040] The perception system 20 performs vehicle perception processing to establish spatial and temporal relationships between the vehicle and static and movable obstacles in the environment. This perception processing may include, in particular, a simultaneous localization and mapping (SLAM) method, which includes modeling static portions, and / or a detection and tracking of moving objects (DATMO) method, which includes modeling movable portions in the environment.

[0041] The perception system 2 can further implement a fusion algorithm to process information from various sensors 20, enabling one or more perceptual operations, such as tracking and predicting the changes in the environment of the vehicle 1 over time, generating a map on which the vehicle 1 is positioned, locating the vehicle on the map, and so on.

[0042] For example, in order to perform perceptual methods such as detecting and / or tracking obstacles and / or determining the global localized position of the vehicle 1 using the integration of data from positioning systems such as GNSS (an acronym for "Global Navigation Satellite System") satellite positioning systems, the perceptual system 2 may combine information from various sensors 20 to determine one or more of the sizes to be used.

[0043] The perception system 2 can be associated with perception parameters that can be defined offline by calibrating the performance of the perception system 2 according to the installed sensors 20.

[0044] Figure 3 shows an example of a traffic environment 300 that can be adapted to run a control application 110 for controlling a vehicle, which implements a vehicle detection and tracking algorithm.

[0045] Vehicle 1 travels on a road 5 that includes three traffic lanes 50. Vehicle 1 is equipped with a set of sensors 20 belonging to a perception system 2.

[0046] In the environment 300 shown in Figure 3, vehicle 1 is followed by vehicle 4 (referred to as the "following" vehicle) in lane 50. Vehicle 1 is further preceded by another vehicle 3 (referred to as the "leading vehicle" or "vehicle ahead") in lane 50.

[0047] The vehicle itself (vehicle 1), the following vehicle (vehicle 4), and / or the lead vehicle (vehicle 3) may, advantageously, be autonomous connected vehicles.

[0048] The control application 110 can thus control the speed v of vehicle 1 according to the speed of the vehicle in front 3 and / or the speed of the vehicle behind 4 and / or the gap d between vehicle 1 and the vehicle in front 3 and / or the gap between vehicle 1 and the vehicle behind 4.

[0049] A vehicle tracking application type control application 110 can control the driving of vehicle 1 by, for example, comparing a gap d to a minimum vehicle-to-vehicle distance (safety gap) and / or comparing a vehicle-to-vehicle time, defined as equal to the ratio between gap d and the speed v of vehicle 1, to a minimum vehicle-to-vehicle time (also called the "reference time"). The control application 110 may further adjust the relative speed between vehicle 1 and vehicle 3 ahead to keep that relative speed substantially zero. A control application that implements vehicle tracking may further apply a control law to keep the vehicle-to-vehicle distance (self / forward) to a target value. The gap may conform to a reference model defined in relation to a predetermined collision zone.

[0050] The distance between vehicles can be measured, for example, by means of a lidar-type sensor 20 (laser telemetry) or by using two cameras installed on either side of the vehicle's windshield.

[0051] The speed of vehicle 1 can be determined, for example, by using an odometer-type sensor.

[0052] The longitudinal and lateral accelerations of vehicle 1 can be determined, for example, by using an inertial measurement unit.

[0053] In some embodiments, the yaw, roll, and pitch speeds of vehicle 1 may be determined using a gyroscope.

[0054] In one embodiment, the magnitude G used by the control application 110 may be a magnitude selected from among the vehicle speed, the distance between the vehicle 1 and the leading vehicle 3, and the yaw speed of the vehicle 1.

[0055] Figure 4 shows the architecture of a control device 110 implemented in vehicle 1 to control the operation of the vehicle, according to a specific embodiment.

[0056] In one embodiment, the control device 11 further includes a compensator 114 configured to generate a replacement signal γ, the value of which varies according to an accuracy index.

[0057] In one embodiment, the selected base controller includes a first base controller 1120-A and a second base controller 1120-B, wherein at least one of the base controllers is activated according to a value of the accuracy index.

[0058] In one embodiment, if the accuracy index has a first value representing optimal perceptual accuracy, the substitution signal is updated to the first signal value by the controller 11, and the first basic controller 1120-A is fully activated in response to the update of the substitution signal.

[0059] Otherwise, if the accuracy index of the perceptual system has a second value representing an inaccurate perception, the substitution signal is updated to the second signal value by controller 114, and the second basic controller 1120-B is fully activated in response to the update of the substitution signal.

[0060] In one embodiment, the first controller 1120-A (also represented by "CD1") is a vehicle tracking controller, and the second controller 1120-B is a noise suppression controller (also represented by "CD2"). The noise suppression controller is configured to suppress any noise having frequencies exceeding the bandwidth specified for the control system, including perceptual noise suppression.

[0061] In one embodiment, the control application 110 controls a plurality of N magnitudes G1, G2, ..., G measured by at least two sensors 20 of the perception system 2. N The control device 11 then uses N sizes G1, G2, ..., G NIt may include a switch 115 for selecting a size from among them, and the adaptive controller 112 is applied to the selected size by using the activated base controller 1120.

[0062] Figure 5 shows a control system 100 according to a specific embodiment of the present invention. In the example in Figure 5, two basic controllers 1120-A and 1120-B are used.

[0063] In such embodiments, the control device 11 may include a calibration unit 113 for calibrating the perception parameters. The calibration unit 113 may use a lookup table that associates each object distance with the speed of the vehicle 3 ahead and an inaccuracy parameter representing the level of inaccuracy of a given measured magnitude G. The magnitudes associated with the inaccuracy parameter may be, for example, vehicle-related dynamic magnitudes such as vehicle speed, distance, yaw velocity, etc. The lookup table (also called the “calibration table”) may be, for example, a three-dimensional (3D) table.

[0064] Offline calibration of the vehicle may be used to define a set of design parameters specific to the vehicle's performance. Thus, the control system 100 may use a level of inaccuracy that is specified at the time of design of the vehicle's performance.

[0065] In the example shown in Figure 5, the first basic controller 1120-A is a standard production system basic controller 1120-A, and the second basic controller 1120-B is a basic controller that can be adapted to the control application 110.

[0066] Thus, in this exemplary embodiment, the adaptive controller 112 can activate a standard production system basic controller 1120-A and / or a basic controller 1120-B that can be adapted to a control application, according to an accuracy index.

[0067] For example, it is assumed that the control device 11 includes a control application 110 that controls a speed change between two values, for example, between 0 m / s and 10 m / s, and that the inaccuracy of the perception system 2 is similar to the inaccuracy shown in Figure 1. The vehicle 1 equipped with the installed perception system 20 follows the vehicle 3 in front of it at a distance determined by its speed. As the speed increases, the distance between the vehicles increases further. Moreover, the higher the speed, the greater the inaccuracy of the perception system may be. By using an adaptive controller 112 according to the embodiment, which has the ability to activate a first or second basic controller 1120-B according to an accuracy index, the influence of such inaccuracy of the perception system on the performance of the vehicle can be advantageously minimized.

[0068] The control device 11 uses, at its input, a perceived magnitude G obtained in real time from the perception system 20 on which the control application 110 depends. For example, in an example where the control application 110 controls a change in speed between two values, the input perceived magnitude G may be the real-time speed of the vehicle ahead. The perceived magnitude G may be transmitted to a performance degradation unit 116 configured to determine the state of the installed performance system. The output of the performance degradation unit 116 and the output of the basic controller 1120 may then be supplied to a real-time vehicle performance adapter 19, which is configured to apply adaptive control in real time.

[0069] The real-time vehicle performance adaptor 19 may include a perceptually adaptable processing unit 112 configured to fuse two output data items from basic controllers 1120-A and 1120-B, which yield fused control data, into a single stable structure containing the fused control data. The real-time vehicle performance adaptor 19 may further include a response corrector 114 configured to adapt the performance of vehicle 1 according to the fused control data.

[0070] A control device 11 according to an embodiment of the present invention can thus be used to dynamically adapt the vehicle's response to a deterioration in perception by maximizing the vehicle's performance with respect to perceptual accuracy while maintaining stability.

[0071] A control device 11 according to an embodiment of the present invention is configured to constantly switch between different basic controllers 1120, including, in an example, two basic controllers 1120-A(C1) and 1120-B(C2), each basic controller defined according to a specific control design criterion.

[0072] In particular, in one exemplary embodiment, the first basic controller C1 (1120-A) may be configured to satisfy maximum tracking performance CD1, while the second basic controller C2 (1120-B) may be configured to satisfy maximum perceived noise rejection CD2.

[0073] In one embodiment, the response corrector 114 may be configured to generate a substitution signal γ, the value of which may vary between zero (0) and one (1) according to an accuracy index representing the level of perceptual accuracy in real time, which is determined by the adaptive controller 112. In one embodiment, the accuracy index may take a value between two extremes, including a first extreme value indicating an optimal level of accuracy (accurate perception) and a second extreme value indicating a degraded level of accuracy (inaccurate perception).

[0074] In one embodiment, when it is determined that the perceived magnitude G (which may be, for example, the vehicle's speed, the distance between vehicles, the vehicle's yaw speed, etc.) has a low-noise value, and that the accuracy index of the perception system 2 has a first extreme value indicating optimal perceived accuracy, the substitution signal γ can be set to a value of 0 (γ=0), and the first basic controller C1(1120-A) is fully activated.

[0075] In one embodiment, the perceived magnitude can be determined to be noisy or low in noise according to the vehicle speed and gap, with higher speeds resulting in noisier signals. Alternatively, the perceived magnitude can be determined to be noisy or low in noise by using a precision index related to perception.

[0076] In other variations, it is still possible to use methods that allow the signal to be analyzed in terms of noise estimation.

[0077] The accuracy index may take the form of a value that changes continuously over time. These values ​​may be represented by quality values. Thus, the evaluation performed by the performance degradation unit 116 may include determining the level of noise present in the signal in real time and continuously. The possible values ​​of the accuracy index may be determined according to the level of noise present in such a signal.

[0078] When it is determined that the accuracy index of perceptual system 2 has a second extreme value indicating an inaccurate perception, the substitution signal γ can be set to a value of 1 (γ=1), and the second basic controller C2(1120-B) can be fully activated.

[0079] The values ​​of the precision index may be binary or non-binary. In one exemplary embodiment, the two extreme values ​​of the precision index, including a first value representing an accurate perception and a second value representing an inaccurate perception, may be binary, and the two values ​​are then linearly interpolated to obtain an intermediate value.

[0080] The substitution signal γ can gradually transition between two extremes (γ=0 and γ=1) according to the value of the perceptual precision index. Thus, the substitution signal γ can be set to a value between 0 and 1 according to the value of the precision index between the two extremes. The first basic controller C1 (1120-A) and the second basic controller C2 (1120-B) can then be partially activated according to the value of the perceptual signal.

[0081] In this way, the vehicle's controller 112 can adapt to the actual operating conditions in real time.

[0082] It is worth noting that the basic controllers C1 and C2 (1120-A and 1120-B) can be configured more generally according to predefined control criteria. In the example considered above, the two basic controllers C1 and C2 are selected so that the desired accuracy performance ensures accurate tracking with respect to noise rejection. These control criteria may even be contradictory, which is not possible with respect to a single controller in the past.

[0083] Figures 6 and 7 show examples of the realization of the adaptive controller 112 in the form of a circuit that performs closed-loop control, according to a specific embodiment.

[0084] Figure 6 shows an example of an implementation of the adaptive controller 112 according to one embodiment. In the example in Figure 6, the adaptive controller 112 is implemented based on a structure parameterized by a filter Q that enables a stable switching from basic controller C1 to basic controller C2. Such parameterization enables stable switching between basic controllers 1120 (1120-A and 1120-B in the example) performed by switch 11 using a substitution signal γ representing a scalar factor.

[0085] In the embodiment shown in Figure 6, the basic controller 1120 includes two basic controllers C1(1120-A) and C2(1120-B). The two basic controllers C1(1120-A) and C2(1120-B) are advantageously configured as follows:

[0086] - Each basic controller C1 or C2 has parameter P veh The vehicle model (120) represented by this can stabilize the closed loop independently.

[0087] - The closed-loop response for each basic controller includes some form of unstable pole cancellation (the control structure used is implemented to avoid unstable configurations, even in the case of switching in the controller).

[0088] In this embodiment, the two basic controllers C1 or C2 have the same number of inputs and outputs. More specifically, each controller has an input corresponding to the output of the vehicle model 120 and an output that can be provided to the input of the vehicle model 120 if the controller is selected by the switch 115.

[0089] In Figures 6 and 7, the basic controllers C1 and C2, and the vehicle model P are shown. veh This is shown using a state representation that models the adaptive controller 112 in a dynamic form that uses state variables related to the system.

[0090] In the parameterized control structure shown in Figure 6, the adaptive controller 112 controls the switch U(1500) and the vehicle model P veh This is represented by (120).

[0091] Referring to Figure 7, we see the basic controllers C1(102) and C2(103), and the vehicle model P veh This is factorized to produce a matrix of stable transfer functions.

[0092] The notation used below is conventional. The matrix format may vary depending on the controller involved.

[0093] In the following explanation, X2 and Y2 represent matrices corresponding to C2, where C2 is represented according to A, B, X1, Y1, and Q, which is a controller parameter calculated according to X2 and Y2.

[0094] Thus, the vehicle model is represented in matrix form defined by relation (1). TIFF0007918173000001.tif10170

[0095] The basic controller C1 is represented in matrix form by relation (2). TIFF0007918173000002.tif10170

[0096] The basic controller C1 is represented in matrix form by relation (3). TIFF0007918173000003.tif10170

[0097] The parameterized adaptive controller 112 can then be represented by C(γ) defined by equation (4). C(γ)=(Y1+γAQ) -1 (X1 + γBQ) (4)

[0098] In equation (4), matrix Q represents a stable adaptive matrix that can be used to switch from basic controller C1 to basic controller C2, and vice versa, enabling stable interpolation between two controllers by activating a portion of each basic controller according to the value of the substitution signal γ.

[0099] A stable adaptation matrix Q can be calculated using equation (5) below. TIFF0007918173000004.tif10170

[0100] It is worth noting that the present invention is not limited to the use of two basic controllers, but rather applies to any number N basic controllers Ci(1120) such that [i=1;N].

[0101] Depending on the embodiment, stable interpolation (i.e., stable transition) between basic controllers Ci is possible. Interpolation can be performed using arbitrary signals that can be adapted to select the appropriate controller in response to real-time operating conditions. In this way, the adaptive controller 112 allows the vehicle to adapt to operating conditions in real time while maintaining system stability (i.e., the system does not increase the output if the input received is the same or less).

[0102] Figure 8 illustrates the performance of a control device 11 according to a specific embodiment of the present invention, using a first basic controller 1120-A of the vehicle tracking controller type and a second controller 1120-B of the perceptual noise reduction controller type. The performance is illustrated in comparison to a vehicle controller from the prior art that uses only the vehicle tracking controller 1120-A.

[0103] Figure 8 shows the speed of the lead vehicle (curve A1), the response of the adaptive control system ACC to the change in the lead vehicle's speed, corresponding to controller C1 implemented in the vehicle according to the prior art (curve A2), and the response of a perception-adaptive controller 112 implemented in the vehicle according to several embodiments (curve A3), the perception-adaptive controller 112 having the ability to select at least one of two basic controllers C1 and C2 according to a substitution signal γ (also called a "substitution factor") and to activate the selected basic actuator. The perception-adaptive controller 112 can thus be used to share the performance of the basic controllers. Figure 9 illustrates the transition of substitution signal γ corresponding to an embodiment, associated with curve A3.

[0104] Figure 8 can be seen to confirm that the lead vehicle 3 initially has a speed of 0 m / s (curve A1). The speed of the lead vehicle then increases to 5 m / s, and finally reaches 8 m / s. Two basic controllers C1(1120-A) and C2(1120-B) used by the adaptive controller 112, according to several embodiments (curve A3), each have different control objectives: one controls the tracking speed error (C1(1120-A)), and the other controls the noise level of the perception system (C2(1120-B)).

[0105] As shown in Figures 8 and 9, the adaptive controller 112 selects at least one of controllers C1 and C2 according to factor γ (also called the "substitution factor").

[0106] - When factor γ is equal to 0 (γ=0), the vehicle 1 determines that the perception system 2 is working perfectly and without any inaccuracies (i.e., the perception system 2 has optimal accuracy). The adaptive controller 112 then activates the first basic controller C1, which acts on the vehicle's actuators to minimize the vehicle tracking speed error as much as possible.

[0107] - When factor γ is equal to 1 (γ=1), the adaptive controller 112 activates the second basic controller C2, which acts on the vehicle's actuators to reduce the noise potentially introduced by the first basic controller C1 by only 50%.

[0108] This type of adaptive control sacrifices only a very slight reduction in tracking performance in terms of design, resulting in a completely different, but stable, response.

[0109] Figures 8 and 9 show that when the velocity is equal to 0, there is no noise in perception, and factor γ is equal to 0. When the velocity increases to 5 m / s, factor γ increases to 0.2, which leads to a small difference in the response of controller C1 and the perception-adapted controller 112. Furthermore, the next velocity change leads the perception system 2 to its maximum level of inaccuracy (leading vehicle velocity equal to 8 m / s). During the transition from 5 m / s to 8 m / s, factor γ changes from 0.2 to 1, which means that the contribution of the second basic controller C2 becomes larger, as clearly illustrated by the responses A2 and A3 of the two systems between 50 and 60 seconds. The closer the velocity is to 8 m / s, the greater the difference between response A2 (corresponding to controller C1 according to prior art) and the response of the perception-adapted controller 112. It should be noted that tracking ability begins to decrease around 58 seconds. However, the response is smoother and more robust. From 60 seconds onward, and until the end of the experimental measurement, factor γ remains close to a value of 1, although it fluctuates somewhat around 1 due to the speed fluctuations of the lead vehicle (curve A1).

[0110] The advantage of the perception-adaptable controller 112 is even more evident around 72 seconds, at which point it can be noted that the perception system 2 exhibits considerable inaccuracy when the speed of the lead vehicle is between 6 m / s and 10 m / s over a short period (in the test performed, the speed of the lead vehicle is 8 m / s). By using controller C1, such inaccuracy is reproduced in the performance of the vehicle itself, and however, this inaccuracy is completely absorbed by the perception-based control system according to the embodiment of the present invention.

[0111] Embodiments of the present invention can thus be used to adapt the vehicle's performance in real time in response to the inaccuracies of the perception system 2, enabling very good control capabilities, such as vehicle tracking ability, and good disturbance rejection when required. Therefore, it is possible to optimize the vehicle's performance in any situation.

[0112] Embodiments of the present invention can thus be used to provide an adaptive control system 112 that has the ability to adapt to various control objectives in real time.

[0113] The adaptive control system 1 can effectively adapt to the inaccuracies (noise reduction) of the perception system and adjust the vehicle's response to a given scenario or driver.

[0114] In one exemplary embodiment, the adaptive control device 10 may be multi-sensor by providing (more than two) basic controllers 1120, each corresponding to a predetermined type of target performance, so that the driver can select the target performance according to, for example, the driver's preference.

[0115] The multi-sensor functionality can thus be added to ADAS or AD-type control systems, and the present invention can be used to adapt, change, and parameterize the response of the control device 10, whatever the sensors may be. Multiple basic controllers 1120 can be enabled and activated / deactivated according to target criteria related to driving comfort or user preference.

[0116] The present invention is not limited to one specific type of vehicle, but applies to any type of vehicle (vehicle examples include, without limitation, automobiles, trucks, buses, and others). While not limited to such applications, embodiments of the present invention have particular advantages for realization in autonomous vehicles, where the autonomous vehicles are connected by a communication network that enables them to exchange V2X messages.

[0117] Those skilled in the art will understand that systems or several subsystems according to embodiments of the present invention can be realized in various forms, in different ways, as hardware, software, or a combination of hardware and software, particularly in the form of program code that can be distributed in the form of a program product. In particular, the program code may be distributed using computer-readable media, which may include computer-readable storage media and communication media. The methods described herein can be realized in particular in the form of computer program instructions that can be executed by one or more processors in a computer computing device. These computer program instructions may further be stored in computer-readable media.

[0118] Furthermore, the present invention is not limited to the embodiments described above as non-limiting examples. The present invention encompasses all actualized modifications that can be imagined by those skilled in the art. In particular, those skilled in the art will understand that the present invention is not limited to individual types, individual numbers, or individual types of basic controllers of the sensor system 20.

Claims

1. A control system for controlling a vehicle (1), wherein the vehicle implements at least one control application using a size measured by at least one sensor (20) of a perception system (2) installed on the vehicle (1), the control system includes an adaptive controller (112) configured to dynamically activate one or more basic controllers from a set of basic controllers, each basic controller, based on a substitution signal, each basic controller is configured to use a control function for controlling the parameters of the vehicle by acting on the actuators of the vehicle, the control functions are distinct based on an accuracy index of the perception system determined according to the real-time value of the size, and the value of the substitution signal varies according to the accuracy index. A control system characterized in that the adaptive controller is configured to dynamically activate one or more basic controllers by switching the at least two basic controllers based on the substitution signal while the vehicle is in motion.

2. The control system according to claim 1, characterized in that the value of the substitution signal fluctuates between a first value representing the optimal level of accuracy of the perception system and a second value representing the minimum level of accuracy of the perception system.

3. The control system according to claim 2, characterized in that the first value is equal to 0 and the second value is equal to 1.

4. The control system according to claim 2 or 3, characterized in that the plurality of basic controllers include a first basic controller and a second basic controller, and if the accuracy index has the first value, the substitution signal is updated to a first signal value, and the first basic controller is fully activated in response to the update of the substitution signal.

5. The control system according to any one of claims 2 to 4, wherein the plurality of basic controllers include a first basic controller and a second basic controller, and if the accuracy index of the perception system has the second value, the substitution signal is updated to the second signal value, and the second basic controller is fully activated in response to the update of the substitution signal.

6. The control system according to any one of claims 1 to 5, characterized in that the magnitude is one of the following: the speed of the vehicle, the distance between the vehicle and the leading vehicle, and the yaw speed of the vehicle.

7. The control system according to any one of claims 1 to 6, characterized in that a first basic controller included in the plurality of basic controllers is a vehicle tracking controller, and a second basic controller included in the plurality of basic controllers is a noise reduction controller.

8. The control system according to any one of claims 1 to 7, characterized in that the control application uses a plurality of sizes measured by at least two sensors of the perception system, the control device includes a switch for selecting a size from the sizes, and the adaptive controller is applied to the selected size by using the activated base controller.

9. A control method for controlling a vehicle (1), which is performed in the vehicle and which implements at least one control application using a size measured by at least one sensor (20) of a sensing system (2) installed on the vehicle (1), the method is a. Selecting at least one basic controller from a set of basic controllers, including at least two basic controllers, based on a substitution signal, wherein each basic controller is configured to use a control function for controlling the parameters of the vehicle by acting on the vehicle's actuators, the control functions being distinct based on an accuracy index of the perception system determined according to the real-time value of the magnitude, and the value of the substitution signal varies according to the accuracy index, and the selection of at least one basic controller. b. Activating the selected at least one basic controller, which includes switching the at least two basic controllers based on the substitution signal while the vehicle is in motion. A control method characterized by including a step that includes the following.

Images