IMPROVED CONTROL SYSTEM OF ACTUATOR IN A VEHICLE

Description

“IMPROVED CONTROL SYSTEM OF ACTUATORS IN A VEHICLE" *** **** *** TECHNICAL SECTOR 5 This solution is related to a system of refined control of actuators in a vehicle, in particular in a motor vehicle. EARLY ART As is known, recently produced motor vehicles are 10 equipped with numerous electronic control systems, such as for example traction control systems (TCS, Traction Control System), stability control systems (ESP / ESC, Electronic Stability Program / Electronic Stability Control) or anti-lock braking systems (ABS, Anti-lock Braking System), 15 just to name a few. These electronic control systems base their operation on real-time detection of a certain number of quantities associated with the operation of the motor vehicle, such as speed, acceleration, 20 steering angles and so on. These quantities are detected by a plurality of sensors on board the vehicle, which provide raw signals which are subsequently subjected to digital processing operations, such as 25 example of amplification and filtering, in order to make them available for one or more control units of the motor vehicle (typically at least one control unit main, usually known as ECU – Electronic Control Unit). 5 These control units are then configured in in order to implement a plurality of logics for the real-time control while the vehicle is moving, of various actuators or similar devices on board the vehicle same. 10 The control logics are closed in feedback on the quantities detected by on-board sensors or, alternatively, on quantities estimated from the same quantities detected (in case it is not possible, or it is not advantageous, the use of dedicated sensors for the 15 detection of these quantities). In particular, in commonly used solutions, each quantity (or state) is estimated with an algorithm of dedicated estimate (for example, using an estimator of the vehicle speed, a slip estimator 20 lateral, etc.); in the aforementioned control logics, the control loop is then closed on the provided measurements in feedback from the sensors and / or on the estimates provided by the estimators present in the vehicle. Although efficient, control solutions 25 currently used are quite complex and resource-intensive with regards to the related tests required for validation and testing (e.g. for example, because it is necessary to test a plurality of different estimation algorithms and 5 control). PURPOSE OF THE INVENTION The aim of this solution is in general to provide an improved solution for the control of actuators on board a vehicle, which can allow for 10 overcome or at least limit the problems previously highlighted. In accordance with the above objective, according to the This solution provides a control system as defined in the appended claims. 15 BRIEF DESCRIPTION OF THE FIGURES The present invention will now be described with refer to the attached drawings, which illustrate an example of non-limiting implementation, in which: - Figure 1 shows a schematic representation of a motor vehicle 20 equipped with a control system implemented in a related electronic control unit; - Figure 2 is a block diagram of the system control in the motor vehicle of Figure 1, in a first form of realization; 25 - Figure 3 is a block diagram of the system control, in a second embodiment; - Figure 4 is a schematic representation of the motor vehicle, in which quantities are represented main features of a related dynamic model used as 5 “digital twins”; - Figure 5 shows a block diagram of a possible implementation of an estimator in the system check; - Figure 6 shows a possible implementation example 10 of the control system, for a “steer-by-wire” system; and - Figures 7A-7C and 8A-8C show graphs relating to the control performance, respectively for a solution of known type and for the control system according to the present solution. 15 DETAILED DESCRIPTION OF THE FORMS OF IMPLEMENTATION As will be described in detail later, one aspect of this solution plans to implement, inside a control unit on board a vehicle (e.g., within a related ECU), a model of 20 digital twin of the vehicle, so-called “digital” model twin”, capable of providing a plurality of output states, or of quantities associated with the vehicle's behavior; and to close one or more control loops in feedback for the generation of commands to be given to actuators of the 25 vehicle on the outputs of this digital twin model, rather than on measurements provided by real sensors on board the vehicle or on estimates generated on the basis of the same measurements provided from real sensors. Advantageously, this solution allows you to have 5 better control performance, thanks to the closure of the control ring on a digital model (that is, for its nature, free from uncertainties and measurement noise); Furthermore, this solution makes it faster and less expensive in terms of resources the calibration of the system 10 control (since the simulation calibration can be the same one used in the vehicle). Figure 1 shows a vehicle, in particular a motor vehicle 1, equipped with a passenger compartment 2, with two wheels front 3 (therefore belonging to the same axis, 15 front) and two rear wheels 4 (therefore belonging to the same axle, the rear one); at least one of these front or rear axle receives a driving torque from a powertrain 5 (shown in the example in position rear) of motor vehicle 1, traditional type, hybrid 20 or electric. Motor vehicle 1 comprises at least one control unit electronics or “control unit” (so-called “ECU”) 6 which, among the other functions, manages and controls behavior general of the motor vehicle 1 and its systems 25 main ones, including for example the powertrain 5, by intervening on the torque delivered, the transmission, the brakes, possibly in collaboration with other existing implementations on board the motor vehicle 1. Physically, the electronic control unit 6 can 5 be composed of a single processing device digital (microprocessor, microcontroller or similar), or from multiple devices separated and paired in communication, for example via the CAN network of the motor vehicle 1. 10 The motor vehicle 1 may also comprise further units control, such as one or more VCUs (Vehicle Control Units) Unit), dedicated to the control and monitoring of devices / sensors present on board the vehicle. The control unit 6 (or an additional control unit 15 in the motor vehicle 1) is configured to implement in real-time a control system, generally referred to as with 10, for the control of one or more actuator devices, generally indicated with 7, present on board the same motor vehicle 1 (such as, for example, actuation devices 20 of the steering, brakes, accelerator, or similar, in so-called “by-wire” drive systems, i.e. using electrical or electromechanical connections rather than purely mechanical connections). As shown in Figure 2, the control system 10 25 comprises a first control ring 10a, which closes on a digital twin model (“Digital Twin”) of the motor vehicle 1; and a second control ring 10b, which closes directly on the vehicle 1. The first control ring 10a comprises a model of 5 digital twin 12 configured to implement a virtual representation (or modeling) that replicates faithfully the behavior, especially the behavior dynamic, of the motor vehicle 1. This digital twin model 12 is configured in 10 way to output a vector of states, denoted by ̃ 𝑟 , including current (real-time updated) values ​​of a plurality of quantities associated with the aforementioned behavior of the motor vehicle 1, including for example speed, acceleration, steering angles or wheel angles, 15 pitch and roll angles, heave motion, ground forces or grip. In other words, the digital model 12 provides in output a plurality of status signals, each associated to one of the aforementioned quantities. 20 The aforementioned first control ring 10a comprises a digital controller 14, e.g. a push-button controller proportional, integrative and derivative (PID), a predictive controller (MPC – Model Predictive Control), a “sliding mode” controller or an H-infinity controller, 25 configured to receive an input signal error S , generated by a subtractor stage 15, on the basis And of the difference between a target signal St and a corresponding state signal 𝑆 provided at the output of the 𝑟 ̃ the aforementioned digital twin model 12 and included in the 5 above-mentioned state vector ̃ 𝑟 . The digital controller 14 is configured so as to generate an output command signal S to be supplied to CMD an actuator device 7 of the motor vehicle 1 in operation of the above mentioned error signal S, therefore on the basis of a And 10 control loop that is closed in feedback on the digital twin model 12 (rather than physically on the motor vehicle 1). This control signal S is supplied at the input to the CMD same digital twin model 12, so as to replicate 15 the command action implemented on vehicle 1 (real) and modify accordingly the state vector ̃ 𝑟 given in exit. The aforementioned second control ring 10b comprises a compensator 16, configured to generate output 20 a compensation signal S , based on a comp difference between the actual behavior of the vehicle 1 and the behavior replicated by the twin model digital 12 (this difference being therefore indicative of the goodness of the same digital twin model 12 and the 25 of its performances in faithfully replicating the behavior of the motor vehicle 1). In particular, the second control ring 10b comprises a first combination stage 17, which receives in input the above mentioned Scomp compensation signal and also 5 the above control signal S and is configured so CMD to generate an actual or real command signal SCMD_real which is physically supplied to motor vehicle 1, in particular to a related actuator device 7 for implement effective control action. 10 In the illustrated implementation, this first stage of combination 17 is a summing stage and is configured for generate such a real command signal S as a sum CMD_real between the above compensation signal S and signal comp S command. CMD 15 The second control ring 10b further comprises a second stage of combination 18, in particular a stage subtractor, which receives as input the state vector ̃ 𝑟 provided by the digital twin model 12 and furthermore a vector of measures r provided by a plurality of 20 sensors physically located on board vehicle 1, for example inertial sensors (accelerometers, gyroscopes, sensors force sensors, angle sensors, etc.). The second stage of combination 18 provides output a difference vector ∆ 𝑟 , based on the combination, in 25 detail of the difference between the above mentioned vector states ̃ 𝑟 and vector of measures r; this difference vector ∆ 𝑟 is supplied at the input to the compensator 16, thus being indicative of the aforementioned difference between the behavior actual performance of motor vehicle 1 and the replicated behavior 5 from the digital twin model 12. It should be noted that the aforementioned compensator 16 can be a controller that is easy to build and implement, for example a proportional action controller, Integrative and Derivative (PID controller), since they are 10 generally expected minimal differences (with dynamics from compensate lens) between the actual behavior of motor vehicle 1 and the behavior defined by the model of digital twin 12 (which is in fact configured for faithfully replicate the behavior of the same 15 motor vehicle 1). The second control ring 10b, comparing the were measured from vehicle 1 and those supplied exiting from the digital twin model 12 therefore allows to generate a compensation of the control action, 20 before sending the actual command to vehicle 1 (the the above mentioned real command signal S ) which determines CMD_real the actual control action on board the same vehicle 1. As shown in Figure 3, in one possible form of 25 implementation, the aforementioned control system 10 comprises also a 10c estimation stage, including an estimator of state 20, configured to implement, in particular in real time during the operation of the motor vehicle 1, estimation algorithms (e.g. based on a digital model 5 which replicates the behavior of car 1 to estimate quantities that cannot be measured on the real vehicle), to generate an output vector of estimates ̂ 𝑦 , including a plurality of estimated quantities associated with behavior of the same vehicle 1. 10 In a possible implementation, as it will be discussed in detail also below, the state estimator 20 is based on a Kalman filter implementation, which allows to improve the signal estimates based on a correction of the internal states of the model, taking into account 15 of the measurement noise of the sensors in the real vehicle. In this embodiment, the aforementioned second control loop 10b includes a concatenator stage 19, for example including a multiplexer, which receives in input is the vector of measurements r provided by the sensors 20 physically placed on board the vehicle 1, both the vector of estimates ̂ 𝑦 provided by the state estimator 20. In particular, in this case, this vector of estimates ̂ 𝑦 can be used to complete the vector of measures r for get the set of states provided in output from the 25 digital twin model 12 (state vector ̃ 𝑟 ). In other words, in this case, some of the quantities that define behavior, in particular the dynamic behavior of motor vehicle 1, are measured through physical sensors, while others are estimated by 5 digital estimation algorithms. The above mentioned second control ring 10b is therefore in this case configured to compare both the measured states both the estimated states of the real vehicle with corresponding were provided as output from the digital twin model for 10 generate control action compensation and generate the actual commands for motor vehicle 1 and determine the actual control action. In other words, the the aforementioned concatenator stage 19 is configured so as to provide the difference input of the second stage of 15 combination 18, both the vector of measures r and the vector of estimates ̂ 𝑦 . The state estimator 20 is also configured so to receive as input a correction factor, which is function of a difference or deviation between at least some 20 of the estimates in the estimate vector ̂ 𝑦 and at least some corresponding measures, indicated here with y, provided by respective sensors on board the vehicle 1. In detail, the above mentioned 10c estimation stage of the system control stage 10 also includes a subtractor stage 21 25 which receives at an input of the above mentioned measures ye ad a difference input the corresponding estimates ̂ 𝑦 and provides a relative difference Δ at the output; and a stage of correction 22, configured to apply a factor corrective K to the above difference Δ and provide as input 5 to the state estimator 20 a relative correction factor K∙Δ, according to which the same state estimator 20 is in able to correct the estimation algorithms implemented for the generation of the estimate vector ̂ 𝑦 (in general, given that There are many states to correct, Δ can be a 10 vector and K can be a matrix). In more detail, in a possible implementation, the dynamic model implemented by the twin model digital 12 of the first control ring 10a is a model of the dynamic behavior of motor vehicle 1, of type 15 so-called “double track”, which is performed in real time in the above-mentioned electronic control unit 6 of the motor vehicle 1. In particular, this dynamic model can be described by the following equations: ̂ 𝑣 ̇ = 𝑎 ̂ + 𝜔 ⋅ 𝑣 ̂ 𝑥 𝑥 𝑧 𝑦 ̂ 𝑣 ̇ = 𝑎 ̂ − 𝜔 ⋅ 𝑣 ̂ 𝑦 𝑦 𝑧 𝑥 Where: 𝑎 ̂ = ⋅ [ ( 𝐹 + 𝐹 ) cos ( 𝛿 ) − ( 𝐹 + 𝐹 ) sin ( 𝛿 ) + 𝐹 + 𝐹 ] ; 𝑥 𝑥 , 𝑥 , 𝑓𝑟 𝑓 𝑦 , 𝑓𝑙 ​​𝑦 , 𝑓𝑟 𝑓 𝑥 , 𝑟𝑙 𝑥 , 𝑟𝑟 𝑚 𝑎 ̂ = ⋅ [ ( 𝐹 + 𝐹 ) sin ( 𝛿 ) + ( 𝐹 + 𝐹 ) cos ( 𝛿 ) + 𝐹 + 𝐹 ]. 𝑦 𝑥 , 𝑓𝑙 ​​𝑥 , 𝑓 𝑦 , 𝑦 , 𝑓 𝑦 , 𝑟𝑙 𝑦 , 𝑟𝑟 𝑚 𝑓𝑟 𝑓𝑙 𝑓𝑟 𝑓𝑙 Figure 4 shows a schematic representation of motor vehicle 1, which features, among other things, the main quantities of the aforementioned dynamic model, which intervene in the above equations. In particular, these 5 sizes are as follows: 𝐹 and 𝐹 , which respectively represent the force 𝑥 𝑦 longitudinal and lateral force at the wheels of motor vehicle 1, with the suffixes 'fl', 'fr', 'rl' and 'rr' which respectively indicate the left front wheel, the 10 right front wheel, left rear wheel and the right rear wheel; m, or the mass of motor vehicle 1; CG, or the center of gravity of the vehicle itself; 𝜔 , which is the angular velocity around an axis 𝑧 15 vertical passing through the aforementioned center of gravity CG; 𝛿 , which is the steering angle; O 𝑣 and 𝑣 , which are the speed respectively 𝑥 𝑦 longitudinal and lateral speed of motor vehicle 1; and 𝑎 and 𝑎 , which are the acceleration respectively 𝑥 𝑦 20 longitudinal and lateral acceleration of the motor vehicle 1. In particular, it is highlighted that the aforementioned forces longitudinal and lateral to the wheels of the motor vehicle 1 can be estimated using the well-known “Pacejka formula”, applied to tire dynamics; as is known, the 25 Pacejka model (see for example Hans Pacejka, “Tire and vehicle dynamics”, Elsevier, 2005) is widely used in studies of vehicle dynamics and is bases a semi-empirical model of the tire to calculate forces and moments acting on the same tire. 5 With reference to Figure 5, it is now described in more detail a possible embodiment of the estimation stage 10c of the control system 10, in which the aforementioned state estimator 20 is implemented. In this embodiment, the state estimator 20 10 is based on a Kalman filter, which implements a model two-degree-of-freedom nonlinear mathematical model with correction. In particular, the state estimator 20 is configured so as to receive an input vector u, which includes a plurality of quantities associated with behavior 15 of motor vehicle 1, including for example: angular velocities (yaw rate), steering angles, tire forces, speed at the wheels, suspension shakes measured with potentiometers, etc. This input vector u is also shown in input 20 to motor vehicle 1, determining its dynamic behavior and in particular the aforementioned measures y which are provided in entrance to the subtractor stage 21, in this case through a processing stage 30 (which can perform, if required, appropriate processing, for example 25 amplification and filtering, of the same measures y). The above correction factor K∙Δ is in this case applied to internal states of the model, here indicated by ̃ 𝑥 , which, following the correction (applied in one stage adder 31), are reported at the input of the same 5 state estimator 20 (to be used at the instant of subsequent time as an initial condition of integration). In more detail, in a possible implementation, the above model can be described by the following equations: 10 𝑣 ̇ = 𝑎 ̂ + 𝜔 ⋅ 𝑣 + 𝐾 ( 𝑎 ̃ − 𝑎 ̂ ) 𝑥 𝑥 𝑧 𝑦 𝑎 𝑥 𝑥 𝑥 𝑣 ̇ = 𝑎 ̂ − 𝜔 ⋅ 𝑣 + 𝐾 ( 𝑎 ̃ − 𝑎 ̂ ) 𝑦 𝑦 𝑧 𝑥 𝑎 𝑦 𝑦 Y where 𝑣 , 𝑣 are the states of the double track model, i.e. the 𝑥 𝑦 longitudinal and lateral speed of motor vehicle 1; and 𝐾 𝑎 𝑥 15 and 𝐾 are multiplying factors that define the above 𝑎 Y correction factor K. With reference to Figure 6, a possible implementation example of the above system control 10, for an angle control application 20 steering 𝛿 in a “steer-by-wire” system. O In this case, the digital controller 14 is configured in order to receive input from a generation stage of target 40, a yaw rate target 𝜔 , generated on the 𝑧 _ 𝑔𝑡 𝑡 base of a steering angle imparted by the driver 𝛿 . 𝑟 𝑖𝑣 𝑑 25 The digital controller 14 is then configured so ̃ to generate an output command 𝛿 , intended to be The One supplied to an actuator device coupled to the steering, to determine the steering actuation angle inside the motor vehicle 1. The compensator 16 is in this case configured so 5 to be generated at the output, as compensation signal S , comp a compensating steering angle 𝛿 , for the 𝑝 𝑐𝑜 𝑚 control action compensation, before sending the actual command to the real motor vehicle 1, indicated here with 𝛿 . CMD_real 10 In this case, the digital twin model 12 receives also in input, as quantities associated with the motor vehicle behavior 1, operating signals, from driver's side of motor vehicle 1, accelerator (“Throttle”) and brakes (“Brake”). 15 From what has been discussed, the following clearly emerge: advantages that this solution allows you to obtain. In any case, it is highlighted again that this control solution, allows you to obtain performance improved compared to known solutions, thanks to the fact that 20 the main control ring is closed on a model ideal free from measurement noise and uncertainties. In this regard, with particular reference to the implementation example in Figure 6, Figures 7A-7C and 8A-8C show performance graphs of 25 control, respectively for a known type solution (i.e., with closed control loop on the real vehicle) and for the present solution (using the model of digital twin for closing the control ring). In particular, the trends illustrated are obtained at 5 PID controller calibration parity (e.g. calibrated with a 7 Hz bandwidth). Figures 7A and 8A compare the velocity trend angular 𝜔 , showing both the reference ('ref') and 𝑧 the measured trend ('meas') following the action of 10 control; Figures 7B and 8B compare the progress of the steering column movement, again showing both the command ('Cmd') and the measured trend ('meas') in following the control action; and Figures 7C and 8C compare the trend of the steering angle (SWA – 15 Steering Wheel Angle), showing both the steering wheel angle given by the driver ('drv'), both the actual command ('Cmd') provided by the control action. In general, the above graphs highlight how the control action implemented by using the 20 digital twin model 12 allows to obtain a noise-free and generally replicating performance faithfully the corresponding reference quantity (or control target). In particular, in the 'known solution' one could 25 reduce the oscillations of the control action by derogating controller performance, reducing bandwidth (e.g., below 7 Hz); conversely, the 'present solution' guarantees better performance, since with a bandwidth which can be higher (e.g., 7 Hz) 5 the control action does not exhibit oscillations. As previously indicated, this solution also allows you to calibrate and validate multiple easily the control system (since the calibration in simulation it can be the same as the one used in 10 car). Finally, it is clear that what has been described can be changes and variations can be made without having to leave within the scope of this invention, as defined by claims attached. 15 It should be noted in particular that the solution described can find advantageous application within a vehicle, regardless of its type vehicle or by the controlled actuator device.

Claims

CLAIMS 1. Control system (10), configured to implement real-time control of at least one actuator device (7) on board a vehicle (1), comprising a first control loop (10a) including: a digital twin model (12), configured to implement a virtual representation replicating the behavior of the vehicle (1); and a digital controller (14), configured to generate an output command signal (SCMD) to be supplied to the actuator device (7) of the vehicle (1) based on a closed feedback control loop on said digital twin model (12).

2. System according to claim 1, wherein said digital twin model (12) is configured to output a state vector (r ), including a plurality of quantities associated with the dynamic behavior of the vehicle (1); and wherein said first control loop (10a) comprises a subtractor stage (15), configured to implement a difference between a control target signal (St) and a state signal (Sf ) output by the digital twin model (12) and included in said state vector (r) and to generate an error signal (Se) input to the digital controller (14); said digital controller (14) being configured to generate said command signal (Scmd) as a function of said error signal (Se).

3. System according to claim 2, wherein said command signal (Scmd) is supplied as input to said digital twin model (12), so as to replicate the command action implemented on the vehicle (1) and so that said digital twin model (12) modifies the state vector (r) supplied as output accordingly.

4. System according to any of the preceding claims, wherein said digital twin model (12) is configured to implement a “double track” type model of the dynamic behaviour of the vehicle (1), capable of being executed in real time in an electronic control unit (6) of the vehicle (1).

5. System according to claim 4, wherein said dynamic model is described by the following equations: vx= αχ + ωζ· Vy -τ'-- . . Vy CLy ωζ ' Vx where: ax = ~' [(Fx, fi + Fx, fr) cOs^f) — (Fy, fi + Fy, fr) sin(^f) + Fx,ri + Fx,rr]; ay = m' [(Fx, fi + Fx, fr) sin(^f) + (Fy, fi + Fy, fr) c°s(Sf) + Fy,ri + Fy,rr] , where: Fx and Fy e represent respectively a longitudinal force and a lateral force at the wheels of the vehicle (1) , with the suffixes 'fl', 'fr', 'rl' and 'rr' indicating respectively a left front wheel, a right front wheel, a left rear wheel and a right rear wheel; m is the mass of the vehicle (1); CG is the centre of gravity of the vehicle (1); ωζ is the angular velocity around a vertical axis passing through the centre of gravity CG; δβ is the steering angle; vx and vy are respectively a longitudinal velocity and a lateral velocity of the vehicle (1); and ax and a.y are respectively a longitudinal acceleration and a lateral acceleration of the vehicle (1).

6. System according to any of the preceding claims, comprising a second control loop (10b), configured to generate a compensation of the control action implemented by the first control loop (10a) before providing an actual command to the actuator device (7) of the vehicle (1), based on a difference between an actual behavior of the vehicle (1) and a behavior replicated by said digital twin model (12).

7. System according to claim 6, wherein said second control loop (10b) comprises a compensator (16), configured to generate as output a compensation signal (Scomp), based on a difference between - 22 measurements (r) provided by a plurality of sensors arranged on board the vehicle (1) and states (r) provided as output by said digital twin model (12); and a first combination stage (17), configured to receive as input said compensation signal (Scomp) and also said command signal (SCMD) and to generate an actual command signal (SCMD_real), suitable to be supplied to the actuator device (7) of the vehicle (1) to implement an actual control action.

8. System according to claim 7, wherein said first combining stage (17) is a summing stage and is configured to generate said real command signal (SCMD_real) as a sum between said compensation signal (Scomp) and command signal (SCMD).

9. System according to claim 7 or 8, wherein said second control loop (10b) comprises a second combination stage (18), configured so as to receive as input said states (r ) provided as output by the digital twin model (12) and said measurements (r) provided by the sensors located on board the vehicle (1), and so as to provide as output a difference vector (Δγ), based on a difference between said states (r) and said measurements (r); said difference vector (Δγ) being provided as input to the compensator (16) and indicative of the difference between the actual behaviour of the vehicle (1) and the behaviour - 23 replicated by the digital twin model (12).

10. System according to any of claims 6-9, further comprising an estimation stage (10c), including a state estimator (20), configured to implement estimation algorithms, to generate as output estimates (y) of a plurality of quantities associated with the behavior of the vehicle (1); wherein said second control loop (10b) further comprises a concatenator stage (19), adapted to receive as input both the measurements (r) provided by the sensors located on board the vehicle (1), and the estimates (y) provided by the state estimator (20); said second control loop (10b) being configured to compare both the measurements (r) and the estimates (y) with the states (r ) output from the digital twin model (12) to generate said control action compensation.

11. System according to claim 10, wherein said state estimator (20) is based on a Kalman filter, which implements a nonlinear two-degree-of-freedom model with correction.

12. System according to claim 10 or 11, wherein said state estimator (20) is configured to receive as input a correction factor, which is a function of a deviation between at least some of the estimates (jz) and corresponding measurements (y) provided by respective sensors on board the vehicle (1).

13. System according to claim 12, wherein said estimation stage (10c) comprises: a subtractor stage (21) capable of receiving said measurements (y) at a sum input and the corresponding estimates (j) at a difference input and providing a relative difference (Δ) as output; and a correction stage (22), configured so as to apply a correction factor (K) to said difference (Δ) and provide a relative correction factor (K-Δ) as input to the state estimator (20), on the basis of which said state estimator (20) is able to correct the estimation algorithms implemented for generating the estimates (j).

14. Vehicle (1), comprising at least one electronic control unit (6), configured to implement the control system (10) according to any of the preceding claims.

15. Automotive control software storable in, and executable by, at least one electronic control unit (6) of a vehicle (1), such that, when executed, the control unit (6) becomes configured to implement the control system (10) according to any of the preceding claims.