Method for determining a speed profile of a motor vehicle using non-predetermined acceleration - Patent Application 20070122997

Abstract

The present invention relates to a method for determining a velocity profile to be followed by a vehicle, comprising the steps of taking event data including a distance from an event and a target velocity (V3) for the vehicle at the event, and determining a velocity profile between an initial velocity (V0) and a target velocity to be followed as a function of time in three successive distinct phases, a first phase (Phase_1) in which the jerk is set constant with a predetermined maximum jerk value so as to reach an optimal target acceleration value, a second phase (Phase_2) in which the optimal target acceleration value is kept constant, and a third phase (Phase_3) in which the jerk is set constant again so as to reach a zero acceleration value at the end of the third phase, the optimal target acceleration value being such that the distance required to execute the three phases of the profile is equal to the distance to the event.

Description

[Technical Field]

[0001] The present invention relates to the field of motor vehicles, and in particular to the field of systems and devices for assisting in driving such vehicles, and more particularly to a method for determining a speed profile to be followed by a motor vehicle. [Background technology]

[0002] Adaptive speed control systems for motor vehicles (known by the prefix ACC, standing for "adaptive cruise control") are known, and ACC is designed to continuously control the speed of a "host" vehicle according to a longitudinal control law for the vehicle that can control the speed of the "host" vehicle according to a speed command requested by the driver and the presence of a vehicle ahead in the same driving lane, referred to as a "target vehicle." This longitudinal control law may be used in a variety of scenarios. In this way, the equipped vehicle can automatically adjust its speed to maintain a predetermined safe distance relative to the target vehicle. As soon as a lane becomes clear in front of the equipped vehicle, the vehicle's acceleration automatically increases until it reaches the commanded speed selected by the driver, thus assisting the driver in the driving task. If a slower vehicle is detected ahead in the equipped vehicle's lane, the speed is automatically adjusted to adjust the safe distance. This control law may also simultaneously check comfort criteria (limited acceleration, deceleration, and jerk (i.e., acceleration derivative)).

[0003] The control law thus makes it possible to control the vehicle's longitudinal speed according to a speed command set by the driver and, if applicable, to reduce this speed if a target vehicle is detected in front of the host vehicle. However, this control does not take into account contextual and semantic information that may be extracted from the road scene observed by the multi-sensor detection system installed in the host vehicle, such as arriving at a stop sign, approaching a detour, a corner, traffic congestion, or a change in speed limit, in order to predict deceleration and acceleration. As a result, the vehicle's behavior may not always adapt to its environment. However, taking contextual information into account is desirable to achieve a more robust control system, which is essential for autonomous driving.

[0004] U.S. Patent Application No. 2019106108 describes a control law that determines a speed profile for a motor vehicle along the planned travel route in response, at least in part, to (i) detecting objects along a planned travel route, (ii) determining speed limit changes along the planned travel route, and (iii) a speed selection by a user. The control law then enables managing the speed of the vehicle along the planned travel route in accordance with the determined speed profile.

[0005] However, this publication remains very general in practice. In particular, it shows that a speed profile is used that is based on an acceleration profile that is calculated differently depending on certain conditions, but it does not explain how this profile is calculated to ensure driver comfort. Another problem raised is that of having enough data to allow the system to be calibrated.

[0006] More generally, another problem remains in the prior art, that of generating a speed profile for the speed controller to follow that allows for energy savings over the entire stroke. Summary of the Invention

[0007] In order to overcome the problems mentioned above, one subject of the present invention is a method for determining a speed profile to be followed by a motor vehicle, comprising the steps of capturing contextual information about the road environment of said vehicle via a multi-sensor system of said vehicle, extracting from said captured contextual information event data comprising at least one distance from an event for said vehicle and a target speed for said vehicle at this event, providing a measured initial speed of said vehicle, and determining three successive distinct phases to be followed as a function of time, each with a jerk predetermined to reach an optimal target acceleration value at the end of the first phase. and determining a velocity profile between the measured initial velocity and the target velocity in a first phase in which the optimal target acceleration value is set constant at a determined maximum jerk value, a second phase in which the optimal target acceleration value is held constant throughout the duration of the second phase, and a third phase in which the jerk is again set constant so that the jerk reaches a zero acceleration value at the end of the third phase, the method including determining the optimal target acceleration value during the second phase such that the distance required to execute the three phases of the profile by applying the determined optimal target acceleration value is equal to the distance from the event.

[0008] Advantageously, the distances required to perform the three phases of the profile are calculated from a set of mathematical formulas implemented for the calculation of the speed profile, the calculation of the distances being performed with respect to a set of fixed parameters including the initial speed and acceleration of the vehicle when starting the speed profile, the target speed at the event, and the predetermined maximum jerk value, and with respect to non-fixed parameters including the optimal target acceleration to be reached in the second phase of the profile. - calculating the duration of said phases and the start and end times that define the boundaries of said phases; - calculation of the transit speed at the start and end times that define the extent of the second phase; - calculating said velocity as a function of time for each of said phases; - Calculating the distance traveled at the start and end times that define the phase.

[0009] Advantageously, said determination of said optimum target acceleration value is performed through an iterative and bisecting method from a predetermined range bounded by a minimum acceleration value and a maximum acceleration value.

[0010] Advantageously, at each iteration, the distance required to generate the profile is calculated using an intermediate acceleration value that is the centroid of the two minimum and maximum acceleration values.

[0011] Advantageously, in said third phase, said jerk is set constant at said predetermined maximum jerk value of said first phase.

[0012] Alternatively, in the third phase, the jerk is set constant at a predetermined maximum jerk value different from the predetermined maximum jerk value of the first phase.

[0013] Advantageously, the predetermined maximum jerk value differs depending on whether the speed profile relates to a positive or negative acceleration of the vehicle.

[0014] Advantageously, the method includes the step of transmitting said speed profile as instructions to an adaptive speed control system suitable for said vehicle.

[0015] The invention also relates to a device intended to be installed on board a vehicle for implementing a method as described above, characterized in that it comprises a multi-sensor system capable of capturing contextual information about the road environment of the motor vehicle, a vehicle speed sensor, means for calculating said speed profile as a function of event data extracted from said captured contextual information, and control means for applying said calculated speed profile to a vehicle adaptive speed control system.

[0016] The invention also relates to a motor vehicle equipped with a device such as that described above.

[0017] Other characteristics and advantages of the invention will become clearly apparent from the description given below, given entirely as non-limiting indications, with reference to the attached drawings, in which: [Brief explanation of the drawings]

[0018] [Figure 1] 10 is a graph illustrating an example of a velocity profile according to the present invention as a function of time that allows a vehicle to go from an initial velocity to a target velocity to be reached at an event, where the target velocity is slower than the initial velocity, with acceleration depending on the distance from the event. [Figure 2] 10 is a flowchart illustrating a bisection algorithm implemented to determine the acceleration required in the second phase of the velocity profile. [Figure 3] 10 is a graph illustrating various distance values required to generate a velocity profile as a function of selected jerk and acceleration values for set initial and target velocity values. [Figure 4] 1 is a set of graphs illustrating the speed of a vehicle according to the speed profile, the distance from an event, and the acceleration of the vehicle. DETAILED DESCRIPTION OF THE INVENTION

[0019] The present invention is applied to a motor vehicle equipped with an adaptive speed control system and a multi-sensor recognition system, and can distribute context information related to events in the road scene in front of the vehicle, such as the approach of a bypass, a bend, traffic congestion or a change in speed limit. The data collected by the vehicle sensors are sent to an electronic computer, and thanks to an environmental recognition algorithm that analyzes these data, a description of the environment near the vehicle and a construction of the road scene are built. From this environment, the system can provide a (distance, speed) pair related to the detected event, and the distance D event from this event and the speed V3 (referred to as the target speed) to be reached at this event are included. The detected event may be, for example, a speed limit sign.

[0020] The vehicle also has an on-board speed sensor and sends information about the vehicle's speed and its acceleration by processing that speed.

[0021] The speed profile described below is a speed profile determined from the measured speed of the vehicle and is intended to be sent as an instruction to follow to the vehicle's adaptive speed control system in order to anticipate deceleration and acceleration when approaching an event. More specifically, the speed profile must allow the vehicle to go from its initial speed to the target speed while observing the constraints in vehicle dynamics at points of acceleration that may be positive or negative (deceleration) depending on the type and jerk of the profile, i.e., the derivative of the acceleration. These last constraints will make it possible to optimize the performance when following this profile by the vehicle's adaptive speed control system.

[0022] The principle of the speed profile of the present invention is explained. Starting with the assumption of zero initial acceleration and a constant initial vehicle speed (denoted as V0), and taking as the end assumption a constant target vehicle speed V3 with V3 < V0 since the speed profile here is deceleration, begin with an example of a deceleration profile as shown in FIG. 1.

[0023] The velocity profile presented in FIG. 1 is defined according to the invention in three successive phases: The first phase (denoted Phase_1) spans between a start time t0 and an end time t1 that defines this phase, where the jerk value is A target J to reach the optimum target acceleration value at time t1, whose absolute value is denoted by maxi is set to a predetermined maximum jerk value, the absolute value of which is denoted by a second phase (denoted Phase_2) extending between a start time t1 and an end time t2 delimiting this phase, in which the target acceleration value reached at t1 is kept constant throughout the duration of the second phase, and the jerk value is then zero; The third phase (denoted Phase_3) extends between a start time t2 and an end time t3 that defines this phase, in which the jerk value is increased by a predetermined maximum jerk value J in order to return to a zero acceleration value at time t3. maxi is again preferably set at

[0024] The division of the velocity profile according to the invention into three distinct and consecutive phases is necessary in order to observe the constraints in the vehicle dynamics with regard to the jerk values in phases Phase_1 and Phase_3.

[0025] Therefore, we have the following set of fixed parameters: V0: initial velocity when starting the velocity profile; V3: target velocity to be reached in the event; A init : is the acceleration of the vehicle at the start of the speed profile, which may not be zero; J maxi : The maximum jerk value defined for the profile.

[0026] The speed profile is represented by two calculation functions actually used by the vehicle speed control system: The commanded speed to be followed as a function of time (t), V profil : V profil (t)=f(t,V0V3,A init ,J maxi )

[0027] The distance required to achieve the velocity profile, D profil It is written as:

[0028] Following the profile is done by using the vehicle's multi-sensor perception system as a starting point to predict the distance, D, from the target event that is desired to be predicted. event is the D for a given set of parameters. profil It will be triggered as soon as it is equal to

[0029] Distance D from the event event It is desirable to start the profile when the velocity V3 to be reached at the event is known, and following the profile must ensure behavior that matches the desired target velocity at this event. For this type of profile, the acceleration to be used is not a constraint. This involves generating a velocity profile using an acceleration that is not predetermined. Therefore, it is necessary to determine the acceleration before calculating the velocity profile to be followed. The acceleration is determined by a bisection method.

[0030] The bisection method finds the optimal target acceleration value A target which allows you to find: TIFF0007727630000001.tif17170

[0031] Speed Profile D profil The computational steps for calculating this distance required to achieve ≡ ...

[0032] As defined in FIG. 1, the phase durations and start and end times t0, t1, t2 and t3 that delimit phases Phase_1, Phase_2 and Phase_3 are calculated.

[0033] Passing speeds V1 and V2 are calculated which correspond to the respective passing speeds at the start and end times t1 and t2 that delimit the second phase Phase_2.

[0034] The behavior of the velocity as a function of time is calculated for each of the phases Phase_1, Phase_2 and Phase_3.

[0035] The distance traveled as a function of time at each of the times t0, t1, t2 and t3, which delimit the different phases Phase_1, Phase_2 and Phase_3, denoted X0, X1, X2 and X3, respectively, is calculated. Thus, the distance required to achieve the profile is D profil =X3.

[0036] The results of these velocity profile calculation steps are explained in detail below and subsequently demonstrated.

[0037] For the calculation of the start and end times t0, t1, t2 and t3 that delimit the phases Phase_1, Phase_2 and Phase_3, the following notation is used: t0=0 (by assumption) t1=T 01 t2=T 01 +T 12 t3=T 01 +T 12 +T 23

[0038] T 01 , T 12 and T 23By integrating the velocity between various points using TIFF0007727630000002.tif16170TIFF0007727630000003.tif17170TIFF0007727630000004.tif16170

[0039] Regarding the various passage rates of the phases, i.e. the rates at times t0, t1, t2 and t3, denoted V0, V1, V2 and V3 respectively, these rates can be expressed as follows: V0, the initial velocity of the profile, this is a forced value and corresponds to the currently measured velocity of the vehicle when the profile is calculated; TIFF0007727630000005.tif14170TIFF0007727630000006.tif15170V3 is the target velocity that is desired to be reached by following the velocity profile and is also constrained in the same way as the initial velocity.

[0040] The mathematical expression for the velocity profile v(t) for each of the phases is determined as a function of elapsed time and is as follows: For Phase_1, 0≦t≦t1: TIFF0007727630000007.tif13170Phase_2, for t1≦t≦t2: V(t)=V1-A target ×(t-t1) For Phase_3, t2≦t≦t3: TIFF0007727630000008.tif13170

[0041] The distance traveled as a function of time is calculated as follows: X0=0 (by assumption) TIFF0007727630000009.tif15170TIFF0007727630000010.tif15170TIFF0007727630000011.tif15170

[0042] The distance required to achieve the profile is D profil =X3. Thus, when the distance from the event is equal to X3, following the velocity profile will be triggered by sending a velocity command V(t) as defined above.

[0043] These results are demonstrated here.

[0044] For phase Phase_1, for t such that 0≦t≦t1: Jerk J in profile throughout this first phase 01 (t) is set constant, and J 01 (t)=-J maxi It has a value of

[0045] Initial acceleration A init is zero, so the acceleration as a function of time in this first phase, A 01 Denoted by (t), it has the following values: TIFF0007727630000012.tif17170A 01 (t)=-J maxi ×t

[0046] Then, what can be inferred is the change in velocity as a function of time in this first phase, V 01 Denoted by (t), it is: TIFF0007727630000013.tif17170TIFF0007727630000014.tif15170

[0047] Therefore, the distance traveled as a function of time in the first phase, Phase_1, is 01 (t) is: TIFF0007727630000015.tif17170TIFF0007727630000016.tif15170

[0048] For the second phase Phase_2, for t such that t1≦t≦t2: Acceleration as a function of time (A in this second phase) 12 (denoted as (t)) is constant during this phase, i.e.: A 12 (t)=-A target

[0049] Therefore, what can be inferred is the change in velocity as a function of time in this second phase, V 12 Marked with (t): TIFF0007727630000017.tif19170V 12 (t)=-A target ×(t-t1)

[0050] Therefore, the distance traveled as a function of time in this second phase is X 12 (t) is: TIFF0007727630000018.tif17170TIFF0007727630000019.tif15170

[0051] Finally, for the third phase, Phase_3, for t such that t2≦t≦t3: Jerk J in profile throughout this third phase 23 (t) is set constant and has the following value: J 23 (t)=J maxi

[0052] In the third phase, the final acceleration A final is zero, so in this phase, A 23 The acceleration as a function of time, denoted (t), has the following value: TIFF0007727630000020.tif19170A 23 (t)=(t-t3)×J maxi

[0053] What can then be inferred is the change in velocity as a function of time in this third phase, V 23 Marked with (t): TIFF0007727630000021.tif18170TIFF0007727630000022.tif17170

[0054] Therefore, the distance traveled as a function of time in this phase is X 23 (t) is: TIFF0007727630000023.tif18170TIFF0007727630000024.tif15170

[0055] The duration of each phase, T 01 , T 12 and T 23 It is recalled that for the expressions of and the start and end times that delimit the phases, the following notation is used: t0=0 (by assumption) t1=T 01 t2=T 01 +T 12 t3=T 01 +T 12 +T 23

[0056] The period T of the first phase, Phase_1 01 has the following values: T 01 =t1-t0=t1

[0057] However, the acceleration is continuous between the first phase Phase_1 and the second phase Phase_2, i.e.: A 01 (t1)=A 12 (t1)

[0058] This is equivalent to: -J maxi ×t1=-A target Therefore: TIFF0007727630000025.tif16170

[0059] Thus, the velocity V1 reached at time t1 (corresponding to the end of the first phase and the beginning of the second phase) has the value: V1=V 01 (t1)=V 01 (T 01 ) That is, TIFF0007727630000026.tif14170

[0060] Here, the period T of the third phase Phase_3 23 Starting from period T 23 has the following values: T 23 =t3-t2

[0061] However, just like between the first and second phases, the acceleration is continuous between the second phase Phase_2 and the third phase Phase_3, i.e.: A 23 (t2)=A 12 (t2)

[0062] This is equivalent to: (t2-t3)×J maxi =-T 23 ×J maxi =-A target Therefore: TIFF0007727630000027.tif16170

[0063] Thus, the velocity V2 reached at time t2 (corresponding to the end of the second phase and the beginning of the third phase) has the value: V2=V 23 (t2)=V 23 (T 23 ) Therefore: TIFF0007727630000028.tif17170

[0064] Here, the period T of the second phase Phase_2 12 Starting from period T 12 has the following values: T 12 =t2-t1

[0065] However, the velocity at the end of the second phase at time t2 is denoted V2 and may be written as: V 12 (t2)=V2

[0066] This is equivalent to: V1-A target ×(t2-t1)=V1-A target ×T 12 =V2 therefore: TIFF0007727630000029.tif17170

[0067] Previously deployed V1, V2, T 01 and T 23 By substituting the formula, the following can be deduced therefrom: TIFF0007727630000030.tif16170

[0068] The above is clear from the demonstration of the results of the velocity profile calculation, and its implementation remains simple since the results of the calculation require only simple mathematical operations (addition, multiplication, division) and simple logic checks, even though the results of the calculation are defined in three phases as a function of time. The required computing power is therefore limited.

[0069] We now refer to more general velocity profiles, that is to say velocity profiles that may be accelerated or decelerated, with initial acceleration values that may not be zero.

[0070] The starting assumption used here is therefore an initial acceleration A, which may not be zero. initand a constant initial vehicle velocity V0, with zero acceleration as the end assumption and V3 depending on whether the velocity profile is decelerating or accelerating.<V0またはV3> V0 is a constant target vehicle velocity V3.

[0071] As in the previous example, the speed profile is always defined according to the invention by three successive phases: - The first phase, Phase_1, where the jerk value is calculated as A, depending on the distance from the detected event. target To reach the optimum target acceleration value, denoted by the absolute value, the maximum jerk value J maxi is set to a second phase, Phase_2, in which the acceleration is maintained at the optimal target acceleration value; - The third phase is the maximum jerk value, to return to zero acceleration value.

[0072] In the same way as above, the distance required to achieve the velocity profile is D profil Thus, the profile is written as profil will be triggered when

[0073] Parameter A init and J. maxi Since is an absolute value, the following variables s and s1 are introduced to reflect the relative acceleration and jerk values.

[0074] The variable s is defined as follows: TIFF0007727630000031.tif18170

[0075] Thus, if the profile is an acceleration, i.e., V3>V0, then s will take the value of 1, otherwise (if it is a deceleration), s will take the value of -1. In addition, in the first phase of the velocity profile, Phase_1, the acceleration is A init From s×A targetIn the second phase, Phase_2, the acceleration is s×A target In the third phase, Phase_3, the acceleration is s×A target to 0 m / s 2 It will change to.

[0076] Additionally, the variable s1 is defined as follows: TIFF0007727630000032.tif19170

[0077] This variable is A like this init From s×A target This represents the direction of change of acceleration in the first phase, Phase_1, which changes to s1×J. Thus, the jerk in this phase is maxi takes the value of

[0078] As explained with reference to the previous example, the steps for calculating the velocity profile are as follows: The durations of the phases and the start and end times t0, t1, t2 and t3 that delimit the phases Phase_1, Phase_2 and Phase_3 are calculated.

[0079] Passing speeds V1 and V2 are calculated which correspond to the respective passing speeds at the start and end times t1 and t2 that delimit the second phase Phase_2.

[0080] The behavior of the velocity as a function of time is calculated for each of the phases Phase_1, Phase_2 and Phase_3.

[0081] The distance traveled as a function of time at each of the times t0, t1, t2 and t3 that delimit the different phases Phase_1, Phase_2 and Phase_3, denoted X0, X1, X2 and X3 respectively, is calculated.

[0082] The table below summarizes the changes in the variables jerk, acceleration, velocity and distance traveled during the various phases of a typical velocity profile. TIFF0007727630000033.tif78170

[0083] Following the same procedure as in the previous example with a decelerating velocity profile and zero initial acceleration, the profile parameters and velocity are defined as follows: For the start and end times of the first phase, Phase_1, the following notation is used: t0=0 (by assumption) t1=T 01

[0084] With respect to the respective durations of the first and third phases, Phase_1 and Phase_3, these can be stated as follows: TIFF0007727630000034.tif18170TIFF0007727630000035.tif16170

[0085] The phase passage rate is expressed as: V0 is the forced initial velocity of the profile, TIFF0007727630000036.tif16170TIFF0007727630000037.tif15170V3 is the forced target speed.

[0086] The period T of the second phase, Phase_2 12 is defined as follows: TIFF0007727630000038.tif17170

[0087] The times t2 and t3 that delimit the third phase, Phase_3, are defined as follows: t2=T 01 +T 12 t3=T 01 +T 12 +T 23

[0088] The mathematical expression for the velocity profile v(t) for each of the phases is determined as a function of elapsed time as follows: For Phase_1, 0≦t≦t1: TIFF0007727630000039.tif14170Phase_2, for t1≦t≦t2: V(t)=V1-A target ×s×(t-t1) For Phase_3, t2≦t≦t3: TIFF0007727630000040.tif12170

[0089] The distance traveled as a function of time is calculated as follows: X0=0 (by assumption) TIFF0007727630000041.tif16170TIFF0007727630000042.tif16170TIFF0007727630000043.tif15170

[0090] The distance required to achieve the profile is D profil =X3.

[0091] As a variable, the jerk values set in the first and third phases of the velocity profile may be different.

[0092] A further variable may be defined to take on different jerk values depending on whether the profile involves acceleration or deceleration.

[0093] The acceleration A required in the second phase of the profile, Phase_2 target is determined by bisection according to the bisection algorithm presented in Figure 2. This bisection determines the distance D required to achieve the profile. profil Taking as the standard, D profil is calculated using the formula above and should be equal to the distance from the event.

[0094] Advantageously, the speed profile is therefore generated from the following set of fixed parameters: an initial speed V of the vehicle when starting the speed profile; a speed V to be reached at the event; an initial acceleration A of the vehicle when starting the speed profile; init , a predetermined maximum jerk value and a non-predetermined parameter, in this case the target acceleration A to be reached in the second phase of the profile. target According to one particular feature of the invention, this last parameter takes an optimal value determined upon completion of a bisection search, said optimal acceleration value being selected within a limited range, such as the distance required to achieve a profile for a fixed set of parameters, and said optimal acceleration value being determined within a limited range, such as the distance required to achieve a profile for a fixed set of parameters, and said optimal acceleration value being selected within a limited range, such as the distance D from the event. event is equal to.

[0095] The bisection algorithm will now be described in more detail with reference to Figure 2. The objective is to ensure that the distance required to achieve the profile is within the distance D from the event. event The minimum acceleration value α is set to be equal to min and maximum acceleration value α max an optimal acceleration value a selected within a predetermined range ranging between optim Note that the distance required to achieve the profile decreases as the acceleration increases.

[0096] Thus, in a first initialization step E0, these minimum and maximum acceleration values that define the range are set, namely: α min =α min_init α max =α max_init

[0097] In step E1, the minimum acceleration value α is first determined as the acceleration value required in the second phase of the profile, using the formula explained above. minThe distance Dα corresponds to the distance required to achieve the velocity profile min is calculated, and second, the maximum acceleration value α is used as the acceleration value required in the second phase of the profile. max , and the distance Dα corresponding to the distance required to realize the velocity profile. max is calculated.

[0098] A test step E2 is then carried out, which aims to check whether it is possible to determine an optimal acceleration value, in other words the distance D at which the event is located. event are the maximum acceleration values α max and the minimum acceleration value α min It is checked whether the distances are between the distances required for the profile calculated using: TIFF0007727630000044.tif12170

[0099] If the test fails, the algorithm ends at step E2, meaning that no acceleration value was found to realize the profile such that the distance required to realize the profile is equal to the distance from the event.

[0100] Conversely, if the test is successful, then the algorithm continues to step E3, where the loop iteration index NB iteration is initialized to 0.

[0101] Then, in each iteration, in step E4, the first one is determined to be an intermediate value a within a range bounded by a predefined minimum and maximum acceleration value. bary and the intermediate values are assigned to the respective coefficients δ and (1-δ). min and a max is the centroid of two values containing: a bary =δ×a min +(1-δ)×a max

[0102] Next, in step E5, the distance D corresponding to the distance required to achieve the velocity profile is calculated, still based on the formula explained above. bary is calculated and the value a is used as the acceleration value required for the second phase of the profile. bary Take.

[0103] Distance value D bary and D event are then compared in step E6. event A bary If the distance is greater than or equal to the distance required for the profile calculated using the maximum acceleration value α max and the distance Dα required for the profile max is set and α is max is calculated using: α max =α bary TIFF0007727630000045.tif10170

[0104] Otherwise, in step E8, the minimum acceleration value α min and the distance Dα required for the profile min is set and α is min is calculated using: α min =α bary TIFF0007727630000046.tif10170

[0105] In step E9 it is then checked whether the maximum number of loop iterations has been reached. If not, then the loop iteration index NB iteration is increased in step E30, i.e.: NB iteration =NB iteration +1

[0106] Loop back to step E4.

[0107] If the maximum number of iterations is reached in step E9, the algorithm consequently determines the optimal acceleration value a in step E10. optim gives the result: α optim =α min

[0108] Range [α min_init ,α max_init The determination of the optimum acceleration value over the whole of [Delta] is therefore a bisection search process involving the calculation, at each iteration, of the distance required to achieve the profile using intermediate acceleration values, which are then assigned two values ​​a, respectively, to the coefficients δ and (1-δ). min and a max is the center of gravity.

[0109] The jerk is the value J maxi This method is set to the optimum acceleration A target =α optim ∈[α min_init ,α max_init ], resulting in: D profil (a optim )=D event

[0110] In other words, this optimal acceleration value a optim is calculated using this value determined by the bisection method, such that the distance required to achieve the profile is equal to the distance from the event.

[0111] This method for determining the optimal acceleration value in the second phase of the profile is particularly advantageous. First, the required computational load is predictable. Specifically, convergence is guaranteed for a fixed number of iterations. Second, the method ensures good convergence. Thus, for a fixed number of iterations N and a factor δ=0.5, the result α optim The accuracy of It is approximately TIFF0007727630000047.tif15170.

[0112] Taking a real-world example application of a velocity profile calculated according to the present invention, if: J maxi =0.6m / s 3 A init =0 m / s 2 V0=22.2m / s=80km / h V3=10m / s=36km / h α min_init =0.5ms -2 , α max_init =3.4ms -2

[0113] Below is D event = 180m, we obtain: a optim =2.1ms -2

[0114] Figure 3 shows the D as a function of the jerk and acceleration values ​​selected for a set value of V0 and V3. profil Here, the value of the jerk is fixed, so the sweep is performed by accelerating along the arrow F.

[0115] In this case the predetermined maximum jerk value is therefore 0.6 m / s 3 The value of J maxi However, the maximum jerk value may also be predetermined in another way using a sweep technique. Specifically, the bisection determination step should not use an acceleration sweep along a line with constant y as in FIG. 3, except for a sweep along an axis such as x=y, and the jerk value J maxi is then predetermined as a function of the dependence of the jerk value on the acceleration, and the two optimal accelerations A target and Jerk J maxiThe value is the distance D_profil required to achieve the three phases of the profile by applying the determined optimal target acceleration value and the jerk value, relative to the distance D_profil from the event. event is calculated in two steps in such a way that

[0116] The velocity profile P calculated according to the principles described above and related to these constraints is illustrated in the first graph of Figure 4. In the second graph of Figure 4, the distance D from the event event The curve of is shown as a function of time. The measured velocity V of the vehicle, resulting from following the profile via the longitudinal control law for the vehicle, is calculated when the event is reached, i.e., D event It can be seen in the first graph that when =0 the event does indeed reach the required target velocity, i.e., V3.

[0117] In the third graph of FIG. 4, the acceleration A measured for the vehicle mes The curve is shown as a function of time. If the vehicle acceleration is 1.6 m / s 2 The above acceleration is limited by the limit (α max_init =3.4ms -2 ) is found to be much lower than

[0118] Therefore, the speed profile calculated according to the invention and transmitted as instructions to the vehicle adaptive speed control system appears to enable the vehicle to automatically reduce its speed behavior in order to gradually reach the target speed (36 km / h according to the example), thus making it possible to adapt the speed of the vehicle's movement to the contextual factors detected by the vehicle's sensors, and in particular to reach the desired speed only at the desired distance, i.e. when the event is reached.

[0119] It can be seen from the above that implementing a velocity profile requires only limited computing power due to the simple mathematical operations required. In addition, it does not require preliminary measurements on a predetermined path. In other words, it may be generalized whenever an event ({distance, velocity}) is received, whatever its form.

Claims

1. 1. A method for determining a speed profile to be followed by a motor vehicle, comprising: capturing contextual information about a road environment of the vehicle via a multi-sensor system of the vehicle; A road scene ahead of the vehicle is constructed from the captured context information, an event relating to a road condition is detected from the road scene ahead of the vehicle, and a distance (D ｅｖｅｎｔ ), and a target velocity (V ３ extracting data collected by on-board sensors, including the vehicle's location, from the captured context information; The measured initial velocity of the vehicle (V ０ ) providing a The jerk reaches the optimal target acceleration value (A ｔａｒｇｅｔ determining a velocity profile (P) between the measured initial velocity and the target velocity in a first phase (Phase_1) in which the jerk is set constant at a predetermined maximum jerk value so as to reach a zero acceleration value; a second phase (Phase_2) in which the optimal target acceleration value is kept constant throughout the duration of the second phase; and a third phase (Phase_3) in which the jerk is again set constant so as to reach a zero acceleration value at the end of the third phase; The method determines the distance (D) required to achieve the velocity profile. ｐｒｏｆｉｌ ) is the distance (D ｅｖｅｎｔ ) during the second phase. ｔａｒｇｅｔ ) determining A method wherein said determining said optimal target acceleration value is performed through an iterative and bisection method from a predetermined range bounded by a minimum acceleration value and a maximum acceleration value.

2. The distance required to achieve the velocity profile is calculated from a set of mathematical formulas implemented for calculating the velocity profile, and the distance calculation step is performed by calculating the initial velocity (V) of the vehicle when starting the velocity profile. ０ ) and initial acceleration (A ｉｎｉｔ ) and the target velocity (V ３ ) and the predetermined maximum jerk value, and the optimal target acceleration value (A) to be reached in the second phase of the velocity profile. ｔａｒｇｅｔ ) for parameters that are not fixed, The duration of the first phase, the second phase, and the third phase (T ０１ , T １２ , T ２３ ) and calculation of start and end times that delimit the first phase, the second phase, and the third phase; the transit speed (V) at the start and end times that define the second phase; １ , V ２ ) and calculating a velocity for each of the first phase, the second phase, and the third phase; calculating the distance traveled at the start and end times defining the range for each of the first phase, the second phase, and the third phase; 2. The method of claim 1, comprising:

3. The method of claim 1 , wherein the road environment event is at least an approaching detour, a turning point, a traffic jam, or a change in speed limit.

4. 4. The method of claim 3, wherein in each iteration, the distance required to generate the velocity profile is calculated using an intermediate acceleration value that is the centroid of the two minimum and maximum acceleration values.

5. 5. The method according to claim 1, wherein in the third phase, the jerk is set constant at the predetermined maximum jerk value of the first phase.

6. 5. The method according to claim 1, wherein in the third phase the jerk is set constant at a predetermined maximum jerk value different from the predetermined maximum jerk value of the first phase.

7. 7. A method according to any one of claims 1 to 6, characterized in that the predetermined maximum jerk value differs depending on whether the speed profile relates to a positive or negative acceleration of the vehicle.

8. 8. A method according to any one of claims 1 to 7, characterized in that the method comprises the step of transmitting the speed profile as instructions to an adaptive speed control system.

9. 9. A device intended to be installed inside a motor vehicle for implementing the method according to any one of claims 1 to 8, characterized in that it comprises a multi-sensor system capable of capturing contextual information about the road environment of the motor vehicle, a vehicle speed sensor, means for calculating said speed profile as a function of data collected by the on-board sensors extracted from the captured contextual information, and control means for applying said calculated speed profile to an adaptive speed control system.

10. A motor vehicle comprising a device according to claim 9.

Images