Claims:WE CLAIM:
1. A learning-based motor control system (100) for an electric vehicle comprising:
an electric vehicle motor control subsystem (110) configured to:
inject one or more random tuning parameter values corresponding to each of a plurality of real-time performance parameters to a static controller;
collect each of a plurality of real-time values corresponding to each of the plurality of real-time performance parameters sensed by a plurality of corresponding sensors based on an injection of the one or more random tuning parameter values; and
compute one or more errors based on comparison of the each of the plurality of real-time values with each of a plurality of predetermined threshold values corresponding to a plurality of predetermined performance parameters;
the static controller (120) operatively coupled to the electric vehicle motor control subsystem (110), wherein the static controller (120) is configured to :
compute one or more net error function values representative of the one or more errors; and
generate training data comprising computed values of the one or more net error functions and the one or more random tuning parameter values corresponding to each of the plurality of real-time performance parameters; and
a performance parameter optimization subsystem (130) operatively coupled to the static controller (120), wherein the performance parameter optimization subsystem (130) is configured to:
perform regression of the training data by using at least one learning technique;
minimize the one or more net error functions by using at least one optimization technique;
identify one or more optimized tuning parameter values corresponding to each of the plurality of real-time performance parameters based on regression and minimization of the one or more net error functions ; and
utilize the one or more optimized tuning parameter values for transmitting to the static controller for controlling the performance of the motor (105) to drive the electric vehicle.
2. The system (100) as claimed in claim 1, wherein the plurality of sensors (112) comprises at least one of a hall sensor, a current, a temperature sensor, an accelerometer, a gyro sensor or a combination thereof.
3. The system (100) as claimed in claim 1, wherein the plurality of real-time performance parameters comprises at least one of speed, current, torque, temperature, voltage, acceleration, load or a combination thereof.
4. The system (100) as claimed in claim 1, wherein the one or more errors comprises at least one of errors in steady state value, overshoot value, rise time, fall time or a combination thereof of the plurality of real-time performance parameters comprising at least one of speed response, current response, torque response, regenerative braking force response, temperature response or a combination thereof.
5. The system (100) as claimed in claim 1, wherein the each of the plurality of predetermined threshold values corresponding to the plurality of predetermined performance parameters comprises each of a plurality of reference values corresponding to the plurality of predetermined performance parameters stored in a performance parameter repository.
6. The system (100) as claimed in claim 1, wherein the one or more random tuning parameter values comprises at least one of a proportional gain (Kp), integral gain (Ki), derivative gain (Kd) values or a combination thereof.
7. The system (100) as claimed in claim 1, wherein the performance parameter optimization subsystem (130) is configured to derive the optimized tuning parameter values by the regression and the minimization of the one or more net error functions associated with each of the plurality of real-time performance parameters using the training data and the at least one learning technique for the regression and by using the at least one optimization technique for the minimization.
8. The system (100) as claimed in claim 1, wherein the at least one optimization technique comprises a gradient descent technique for minimization of the one or more net function values.
9. A method (300) comprising:
injecting, by an electric vehicle motor control subsystem, one or more random tuning parameter values corresponding to each of a plurality of real-time performance parameters to a static controller (310);
collecting, by the electric vehicle motor control subsystem, each of a plurality of real-time values corresponding to each of the plurality of real-time performance parameters sensed by a plurality of corresponding sensors based on an injection of the one or more random tuning parameter values (320);
computing, by the electric vehicle motor control subsystem, one or more errors based on comparison of the each of the plurality of real-time values with each of a plurality of predetermined threshold values corresponding to a plurality of predetermined performance parameters (330);
computing, by the static controller, one or more net error function values representative of the one or more errors (340);
generating, by the static controller, training data comprising computed values of the one or more net error functions and the one or more random tuning parameter values corresponding to each of the plurality of real-time performance parameters (350);
performing, by a performance parameter optimization subsystem, regression of the training data by using at least one learning technique (360);
minimizing, by the performance parameter optimization subsystem, the one or more net error functions by using at least one optimization technique (370);
identifying, by the parameter optimization subsystem, one or more optimized tuning parameter values corresponding to each of the plurality of real-time performance parameters based on regression and minimization of the one or more net error functions (380); and
utilizing, by the parameter optimization subsystem, the one or more optimized tuning parameter values for transmitting to the static controller for controlling the performance of the motor to drive the electric vehicle ( 390).
Dated this 08th day of May 2020

Signature

Vidya Bhaskar Singh Nandiyal
Patent Agent (IN/PA-2912)
Agent for the Applicant
, Description:BACKGROUND
[0001] Embodiments of the present disclosure relate to electrically operated vehicle and more particularly, to a learning-based motor control system for electric vehicles and a method to operate the same.
[0002] An electric vehicle (EV) is one that operates on an electric motor, instead of an internal-combustion engine that generates power by burning a mix of fuel and gases. A power system of the electric vehicle includes two components such as the electric motor that provides the power and a controller that controls the application of this power to the electric motor. The performance of the electric motor depends on many factors like voltage, current, speed, temperature etc. As a result, the electronic motor controller is essential which controls different aspects of the electric motor to ensure the right current and voltage is applied across the electric motor. Various motor control systems are available which controls one or more parameters of the electric motor for driving the electric vehicles.
[0003] Conventionally, the motor control systems for the electric vehicle available in market include electronic controllers which require manual intervention for controlling one or more real-time parameters by performing tuning activity. The tuning activity for various parameters includes tuning for, controlling the speed, modifying or limiting the torque, and shielding against faults and overloads and the like. However, the manual intervention involved in the tuning activity sometimes generates inaccurate results as amount of tuning required for controlling the motor of the electric vehicle based on a predefined requirement is unknown and undefined. Moreover, random tuning activity performed due to the manual intervention also compromises performance of the electric vehicle.
[0004] Hence, there is a need for an improved learning-based motor control system for an electric vehicle and a method to operate the same in order to address the aforementioned issues.
BRIEF DESCRIPTION
[0005] In accordance with an embodiment of the present disclosure a learning-based motor control system for an electric vehicle is disclosed. The system includes an electric vehicle motor control subsystem to inject one or more random tuning parameter values corresponding to each multiple real-time performance parameters to a static controller. The electric vehicle motor control subsystem also collects each multiple real-time values corresponding to each of the multiple real-time performance parameters sensed by multiple corresponding sensors based on an injection of the one or more random tuning parameter values, wherein the one or more random tuning parameter values includes, but not limited to, Kp, Ki, and Kd, associated with each performance parameter, each time the vehicle is run. The electric vehicle motor control subsystem also computes one or more errors based on comparison of the each of the multiple real-time values with each of the multiple predetermined threshold values corresponding to multiple predetermined performance parameters. The system also includes the static controller to compute one or more net error function values representative of the one or more errors. The static controller also generates training data including one or more computed net error functions and one or more random tuning parameter values such as Kp, Ki, Kd, corresponding to each of the multiple real-time performance parameters. The system also includes a performance parameter optimization subsystem to perform regression of the training data by using at least one learning technique. The performance parameter optimization subsystem also minimizes the one or more net error functions by using at least one optimisation technique. The performance parameter optimization subsystem also identifies one or more optimized tuning parameter values corresponding to each of the multiple real-time performance parameters based on regression and minimization of the one or more net error functions. The parameter optimization subsystem also utilizes the one or more optimized tuning parameter values for transmitting to the static controller for controlling the performance of the motor to drive the electric vehicle.
[0006] In accordance with another embodiment of the present disclosure, a method for operating a learning-based motor control system of an electric vehicle is disclosed. The method includes injecting one or more random tuning parameter values corresponding to each multiple real-time performance parameters to a static controller. The method also includes collecting each multiple real-time values corresponding to each of the multiple real-time performance parameters sensed by multiple corresponding sensors based on an injection of the one or more random tuning parameter values. The method also includes computing one or more errors based on comparison of each of the multiple real-time values with each of multiple predetermined threshold values corresponding to multiple predetermined performance parameters. The method also includes computing one or more net error function values representative of the one or more errors. The method also includes generating training data including computed values of the one or more net error functions and the one or more random tuning parameter values corresponding to each of the multiple real-time performance parameters. The method also includes performing regression of the training data by using at least one learning technique. The method also includes minimizing the one or more net error functions by using at least one optimization technique. The method also includes identifying one or more optimized tuning parameter values corresponding to each of the multiple real-time performance parameters based on regression and minimization of the one or more net error functions. The method also includes utilizing the one or more optimized tuning parameter values for transmitting to the static controller for controlling the performance of the motor to drive the electric vehicle.
[0007] To further clarify the advantages and features of the present disclosure, a more particular description of the disclosure will follow by reference to specific embodiments thereof, which are illustrated in the appended figures. It is to be appreciated that these figures depict only typical embodiments of the disclosure and are therefore not to be considered limiting in scope. The disclosure will be described and explained with additional specificity and detail with the appended figures.
BRIEF DESCRIPTION OF THE DRAWINGS
The disclosure will be described and explained with additional specificity and detail with the accompanying figures in which:
[0008] FIG. 1 is a block diagram of a learning-based motor control system for an electric vehicle in accordance with an embodiment of the present disclosure;
[0009] FIG. 2 illustrates a block diagram for depicting an operation of an exemplary learning-based control system in BLDC motor driver for an electric vehicle of FIG.1 in accordance with an embodiment of the present disclosure;
[0010] FIG. 3 illustrates a block diagram of a computer or a server in accordance with an embodiment of the present disclosure; and
[0011] FIG. 4 is a flow chart representing the steps involved in a method of operation of a learning-based motor control system for an electric vehicle of FIG. 1 in accordance with the embodiment of the present disclosure.
[0012] Further, those skilled in the art will appreciate that elements in the figures are illustrated for simplicity and may not have necessarily been drawn to scale. Furthermore, in terms of the construction of the device, one or more components of the device may have been represented in the figures by conventional symbols, and the figures may show only those specific details that are pertinent to understanding the embodiments of the present disclosure so as not to obscure the figures with details that will be readily apparent to those skilled in the art having the benefit of the description herein.
DETAILED DESCRIPTION
[0013] For the purpose of promoting an understanding of the principles of the disclosure, reference will now be made to the embodiment illustrated in the figures and specific language will be used to describe them. It will nevertheless be understood that no limitation of the scope of the disclosure is thereby intended. Such alterations and further modifications in the illustrated system, and such further applications of the principles of the disclosure as would normally occur to those skilled in the art are to be construed as being within the scope of the present disclosure.
[0014] The terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process or method that comprises a list of steps does not include only those steps but may include other steps not expressly listed or inherent to such a process or method. Similarly, one or more devices or sub-systems or elements or structures or components preceded by "comprises... a" does not, without more constraints, preclude the existence of other devices, sub-systems, elements, structures, components, additional devices, additional sub-systems, additional elements, additional structures or additional components. Appearances of the phrase "in an embodiment", "in another embodiment" and similar language throughout this specification may, but not necessarily do, all refer to the same embodiment.
[0015] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this disclosure belongs. The system, methods, and examples provided herein are only illustrative and not intended to be limiting.
[0016] In the following specification and the claims, reference will be made to a number of terms, which shall be defined to have the following meanings. The singular forms “a”, “an”, and “the” include plural references unless the context clearly dictates otherwise.
[0017] Embodiments of the present disclosure relate to a learning-based motor control system for an electric vehicle and a method to operate the same. The system includes an electric vehicle motor control subsystem to inject one or more random tuning parameter values corresponding to each multiple real-time performance parameters to a static controller. The electric vehicle motor control subsystem also collects each multiple real-time values corresponding to each of the multiple real-time performance parameters sensed by multiple sensors The electric vehicle motor control subsystem also computes one or more errors based on comparison of the each of the multiple real-time values with each of multiple predetermined threshold values corresponding to multiple predetermined performance parameters. The system also includes the static controller which includes several PID loops corresponding to each performance parameter. The static controller computes one or more net error function values representative of the one or more errors The static controller also generates training data including computed values of the one or more net error functions and the one or more random tuning parameter values corresponding to each of the multiple real-time performance parameters. The system also includes a performance parameter optimization subsystem to perform regression of the training data by using at least one learning technique. The performance parameter optimization subsystem also minimizes one or more net error functions by using at least one optimization technique The performance parameter optimization subsystem also identifies one or more optimized tuning parameter values corresponding to each of the multiple real-time performance parameters based on regression and minimization of the one or more net error functions. The parameter optimization subsystem also utilizes the one or more optimized tuning parameter values for transmitting to the static controller for controlling the performance of the motor to drive the electric vehicle.
[0018] FIG. 1 is a block diagram of a learning-based motor control system (100) for an electric vehicle in accordance with an embodiment of the present disclosure. The system (100) includes an electric vehicle motor control subsystem (110) to inject one or more random tuning parameter values corresponding to each of multiple real-time performance parameters to a static controller .The electric vehicle motor control subsystem also collects each of multiple real-time values corresponding to each multiple real-time performance parameters sensed by multiple corresponding sensors (112) from the electric vehicle at running condition, when the vehicle is run with randomly tuned static controller (120) several times. The one or more random tuning parameter values includes, but not limited to, Kp (proportional gain), Ki (integral gain), Kd (derivative gain) values. In one embodiment, each multiple real-time performance parameters may include, but not limited to, speed, current, torque, temperature, voltage, acceleration, load and the like. As used herein, the term ‘multiple real-time performance parameters’ is defined as multiple actual parameters which are determined by the multiple corresponding sensors (112) in real-time during operation of the vehicle. In one embodiment, the multiple sensors (112) may include, but not limited to, a hall sensor, a current sensor, a temperature sensor, an accelerometer, a gyro sensor and the like.
[0019] The electric vehicle motor control subsystem (110) also computes one or more errors for each of the performance parameters based on comparison of each multiple real-time values with each multiple predetermined threshold values corresponding to multiple predetermined performance parameters. This error generation is done each time the vehicle is run with the static controller (120) tuned with suitable, yet random Kp, Ki, Kd values and large number of errors corresponding to each performance parameters are generated every time. In one embodiment, the static controller may include a static PID controller. In one embodiment, the one or more errors may include difference in values obtained upon comparison of each multiple real-time values with each multiple predetermined threshold values. In such embodiment, the one or more errors may include, but not limited to, at least one of the errors in steady state value, overshoot value, rise time, fall time, settling time value or a combination thereof in speed response, current response, torque response, regenerative braking force response, temperature response or a combination thereof. These errors are further transmitted to the static controller by the electric vehicle motor control subsystem (110).
[0020] From the one or more errors, one or more net error function values for each of the multiple real-time performance parameters are computed by the static controller (120). The system also includes the static controller which includes of several PID loops corresponding to each performance parameter. The one or more net error functions are mathematical function of errors corresponding to each of the multiple real-time performance parameters and are generated by the static controller (120). In a specific embodiment, each multiple predetermined threshold values corresponding to the multiple predetermined performance parameters may include each multiple reference values corresponding to the multiple predetermined performance parameters stored in a performance parameter repository (not shown in FIG.1). In such embodiment, the performance parameter repository may store multiple predetermined threshold values and multiple predetermined performance parameters. In such embodiment, the performance parameter repository may be hosted/stored/installed, but not limited to, on a remote server or in a local storage device. In such embodiment, the remote server may also include a cloud server.
[0021] The static controller (120) also generates training data at a running condition of the electric vehicle, wherein the training data includes, but not limited to, one or more computed net error function values and one or more random tuning parameter values Kp, Ki, Kd, corresponding to each of the multiple real-time performance parameters. The generated training data is further sent by the static controller to the performance parameter optimization subsystem.
[0022] The system (100) also includes a performance parameter optimization subsystem (130) which uses the training data and performs regression and minimization to find the optimized tuning parameter values (Kp, Ki, Kd) for each of the multiple performance parameters. As used herein, the performance parameter optimization subsystem (130) uses at least one learning technique such as artificial intelligence technology-based technique including, but not limited to, machine learning techniques and also uses optimization techniques, to enable self-learning of optimum random tuning parameter values such as the Kp, Ki, Kd for each of the multiple performance parameters. In one embodiment, the at least one machine learning technique may include, but not limited to, an artificial neural network, a regression technique and the like for performing regression of the training data. The regression of the one or more net error functions are evaluated with respect to the one or more random tuning parameter values associated with each of the multiple performance parameters. The optimization technique may include, but not limited to, a gradient descent technique for minimization of the one or more net error functions corresponding to each of the multiple performance parameters.
[0023] The static controller (120) also utilizes the tuning parameters of the PID controller, wherein the tuning parameters include, but not limited to, proportional gain (Kp), integral gain (Ki) and differential gain (Kd) for each of the sensed performance parameter values such as the speed, current, temperature, voltage, acceleration, load and the like, monitored by the electric vehicle motor control subsystem. From such sensed values, the errors associated with the at least one performance parameter, is calculated by the electric vehicle motor control subsystem. The errors may include, but not limited to, steady state value error, an overshoot error, settling time error, rise time error, fall time error and the like. The errors associated with the at least one performance parameter computed by the electric vehicle control subsystem (110) are transmitted to the static controller (120) which is also configured to compute a mathematical function which indicates an overall error associated with the at least one performance parameter mentioned as the net error function. The net error function values may be any suitable mathematical function and is defined as, but not limited to, as sum of the squares of the one or more errors associated with at least one performance parameter. For a given Kp, Ki, Kd values, the static controller (120) generates one net error function value for each of the performance parameters. More the number of the one or more net error functions generated the better training of the parameter optimisation subsystem may be achieved. In one embodiment, the one or more random tuning parameter Kp, Ki and Kd values associated with each of the multiple performance parameter may include, but not limited to, Kps, Kis, Kds, (for speed), Kpc, Kic, Kdc (for current), KpT, KiT, KdT (for temperature) and the like.
[0024] The performance parameter optimization subsystem (130) also identifies which values of the Kp, Ki, Kd corresponding to each of the multiple real-time performance parameters produce least value of the net error function, which is the overall error. The values of Kp, Ki, Kd corresponding to the each of the multiple real-time performance parameters which produces least value of the net error function is then finalised as the final Kp, Ki, Kd values that should be used in the electric vehicle. In one embodiment, the final value of the Kp Ki and Kd are determined by using the at least one artificial intelligence technology-based technique including, but not limited to, a machine learning technique, a deep learning technique such as artificial neural networks, regression technique and the like. In such embodiment, the artificial neural network may include, but not limited to, three inputs, the inputs being the one or more tuning parameters Kp, Ki, Kd of a particular performance parameter, and single output, wherein the output being the one or more net error functions. The artificial neural network also involves optimal number of hidden layers of neurons with many suitable weights, biases and suitable activation functions. The artificial neural network is trained to predict the one or more net error functions for a given set of the Kp, Ki, Kd values of each of the multiple real-time performance parameter by providing the training data and suitable training technique which may include, but not limited to, back propagation (gradient descent) technique. Once the relation between the one or more net error functions and the one or more random tuning parameter values are established by a mathematical regression by the artificial neural network, the minimal value of the net error function of a given performance parameter is found using optimization techniques which may include, but not limited to, gradient descent technique. The corresponding tuning parameter values Kp, Ki, Kd of a performance parameter which yield the minimal value of the net error function, are identified as the final optimal Kp, Ki, Kd values for optimal control of the corresponding performance parameter.
[0025] The performance parameter optimization subsystem (130) also controls performance of the motor (105) with the derived optimized Kp, Ki, Kd values for each of the multiple performance parameter based on automated regression and minimization of the one or more net error functions. These optimum Kp, Ki, Kd values are further transferred to the Static PID controller which generates variable duty ratio/frequency PWM signals. Such pulse width modulation (PWM) signals are further given as input to one or more switches of metal–oxide–semiconductor field-effect transistor (MOSFET) driver which in turn controls the speed, current, torque, temperature and the like of the motor. In one embodiment, the motor of the electric vehicle may include, but not limited to, a brushless direct current (BLDC) motor.
[0026] FIG. 2 illustrates a block diagram for depicting an operation of a learning-based motor control system (100) for an electric vehicle of FIG.1 in accordance with an embodiment of the present disclosure. The system (100) is utilised to monitor and control performance of a motor (105) used in the electric vehicle (108). The motor (105) used in the electric vehicle includes, but not limited to, a BLDC motor (105). The BLDC motor (105) of the electric vehicle (108) starts the operation when a load voltage (107) is applied to the BLDC motor (105) via the MOSFET driver(135). Once, the load voltage (107) is applied, the BLDC motor (105) attains a corresponding speed, which is determined by a speed sensor or a hall sensor. Such speed value which is determined by the speed sensor or the hall sensor in real-time is known as actual speed value. Also, along with speed, multiple real-time values corresponding to each multiple real-time performance parameters are also sensed by corresponding multiple sensors (112). Again, each multiple real-time value is corresponding to each multiple real-time performance parameters are collected by an electric vehicle motor control subsystem (110). Here, the multiple real-time performance parameters may include, but not limited to, speed, current, torque, temperature, voltage, acceleration, load and the like. Also, the multiple sensors (112) used herein may include, but not limited to, a hall sensor, a current sensor, a temperature sensor, an accelerometer, a gyro sensor and the like.
[0027] Further, each of the multiple real-time values are compared with each of the multiple predetermined threshold values corresponding to multiple predetermined performance parameters. For example, the actual speed value of the motor, which is determined by the speed sensor or the hall sensor is compared with the predetermined threshold value of speed which is also known as reference speed value. Here, the multiple predetermined threshold values corresponding to the multiple predetermined performance parameters are stored in a performance parameter repository (115). Upon, comparison, errors are generated, wherein the errors includes difference in values which are generated between each multiple real-time values and multiple predetermined threshold values. Here, the errors may include, but not limited to, at least one of the errors in steady state value, overshoot value, rise time, settling time, fall time or a combination thereof in speed response, current response, torque response, regenerative braking force response, temperature response and the like
[0028] The large number of errors generated by the electric vehicle control subsystem (110) by running the vehicle several times with random tuning of a static controller (120) is converted into one or more net error function values, computed by the static controller (120). The training data includes, but not limited, to random tuning parameter values Kp, Ki, Kd and the computed one or more net error function values. This training data is used by a performance parameter optimization subsystem (130) to automatically find the optimum Kp, Ki, Kd values for each of the performance parameters by using at least one artificial intelligence technique including, but not limited to, a machine learning technique, which includes, but not limited, to an artificial neural network, regression technique, for regression of the training data and by using at least one optimization technique for minimization of the net error function values. For example, the optimization technique may include, but not limited to, a gradient descent technique.
[0029] Further the performance parameter optimization subsystem (130) also controls the performance of the motor with the derived optimized Kp, Ki, Kd values for each of the performance parameters based on automated regression and minimization of the one or more net error function values corresponding to each of the performance parameters. These optimum Kp, Ki, Kd values are further transferred to the static PID controller (120) controller which generates variable duty ratio/frequency PWM signals. These PWM signals are further given as an input to one or more switches of metal–oxide–semiconductor field-effect transistor (MOSFET) driver (135) which in turn controls the speed, current, torque, temperature and the like of the motor (105). In example used herein, the motor of the electric vehicle may include, but not limited to, a brushless direct current (BLDC) motor
[0030] FIG. 3 illustrates a block diagram of a computer or a server (200) in accordance with an embodiment of the present disclosure. The server (200) includes, but not limited to, processor(s) (230), and memory (210) operatively coupled to the bus (220). The processor(s) (230), as used herein, means any type of computational circuit, such as, but not limited to, a microprocessor, a microcontroller, a complex instruction set computing microprocessor, a reduced instruction set computing microprocessor, a very long instruction word microprocessor, an explicitly parallel instruction computing microprocessor, a digital signal processor, or any other type of processing circuit, or a combination thereof.
[0031] The memory (210) includes several subsystems stored in the form of executable program which instructs the processor (230) to perform the method steps illustrated in FIG. 1. The memory (210) is substantially similar to a system (100) of FIG.1. The memory (210) has following subsystems, but not limited to, an electric vehicle motor control subsystem (110), a static PID controller with PWM control (120) and a performance parameter optimization subsystem (130).
[0032] The electric vehicle motor control subsystem (110) to inject one or more random tuning parameter values corresponding to each of multiple real-time performance parameters to a static controller. The electric vehicle motor control subsystem (110) also collects each multiple real-time values corresponding to each of the multiple real-time performance parameters sensed by multiple sensors The electric vehicle motor control subsystem (110) also computes one or more errors based on comparison of the each of the multiple real-time values with each of multiple predetermined threshold values corresponding to multiple predetermined performance parameters.The static controller (120) which includes several PID loops corresponding to each multiple real-time performance parameters. The static controller (120) computes one or more net error function values representative of the one or more errors The static controller (120) also generates training data including computed values of the one or more net error functions and the one or more random tuning parameter values corresponding to each of the multiple real-time performance parameters. The performance parameter optimization subsystem (130) to perform regression of the training data by using at least one learning technique. The performance parameter optimization subsystem (130) also minimizes one or more net error functions by using at least one optimization technique The performance parameter optimization subsystem (130) also identifies one or more optimized tuning parameter values corresponding to each of the multiple real-time performance parameters based on regression and minimization of the one or more net error functions . The parameter optimization subsystem (130) also utilize the one or more optimized tuning parameter values for transmitting to the static controller for controlling the performance of the motor to drive the electric vehicle.
[0033] FIG. 4 is a flow chart representing the steps involved in a method (300) of operation of a learning-based motor control system for an electric vehicle of FIG. 1 in accordance with the embodiment of the present disclosure. The method (300) includes injecting, by an electric vehicle motor control subsystem, one or more random tuning parameter values corresponding to each of multiple real-time performance parameters to a static controller in step 310. The method (300) also includes collecting each multiple real-time values corresponding to each multiple real-time performance parameters sensed by multiple corresponding sensors from a running condition of the electric vehicle with a randomly tuned static controller in step 320. In one embodiment, collecting each multiple real-time value corresponding to each multiple real-time performance parameters may include collecting each multiple real-time values corresponding to each multiple real-time performance parameters which may include, but not limited to, speed, current, torque, temperature, voltage, acceleration, load and the like. In such embodiment, collecting each multiple real-time performance parameters may include collecting each multiple real-time performance parameters sensed by multiple sensors which may include, but not limited to, a hall sensor, a current sensor, a temperature sensor, an accelerometer, a gyro sensor and the like.
[0034] The method (300) also includes computing one or more errors based on comparison of each multiple real-time values with each multiple predetermined threshold values corresponding to multiple predetermined performance parameters from a running condition of vehicle with a randomly tuned static controller in step 330. In one embodiment, computing the one or more errors may include computing the one or more errors which may include, but not limited to, errors in steady state value, overshoot value, rise time, settling time, fall time or a combination thereof in speed response, current response, torque response, regenerative braking response, temperature response and the like based on the comparison of each multiple real-time values with each multiple predetermined threshold values. The method (300) also includes computing one or more net error function values representative of the one or more errors in step 340. In such embodiment, the one or more net error function values is a suitable mathematical function, which is a function of all the errors corresponding to a performance parameter.
[0035] The method (300) also includes generating training data which includes, but not limited to, the one or more net error function values and one or more random tuning parameter values such as Kp, Ki, Kd values by running the vehicle several times with the randomly tuned static controller in step 350. The method (300) also includes performing mathematical regression of the one or more net error functions by using at least one learningtechniques in step 360. In such embodiment, the at least one learning technique may include an artificial intelligence technique which may include, but not limited to, at least one machine learning technique, which includes, but not limited to, artificial neural network, a regression technique, and the like.The method (300) also includes minimizing the one or more net error function values by using at least one optimisation technique. In one embodiment, the at least one optimisation technique may include but not limited to a gradient descent technique in step 370.The method (300) also includes identifying one or more optimized tuning parameter values corresponding to each of the multiple real-time performance parameters based on regression and minimization of the one or more net error functions in step 380.
[0036] The method (300) also includes utilizing, the one or more optimised tuning parameter values for transmitting to the static controller for generation of variable duty ratio/frequency PWM signals for controlling performance of the motor to drive the electric vehicle in step 390. The Kp, Ki, Kd values are fed to the Static controller to run the motor in an optimum way by generating variable duty ratio/frequency PWM signals.The PWM signals obtained are provided as an input to one or more switches of metal–oxide–semiconductor field-effect transistor (MOSFET) driver which in turn controls the performance parameters of the motor .
[0037] Various embodiments of the present disclosure relate to a motor control system of the electric vehicle which automatically calibrates the multiple tuning parameters of the electric motor in order to achieve optimum performance of the electric vehicle. Moreover, the present disclosed system helps in upgrading the electric vehicle’s performance through a set of machine-generated instructions without changing the physical components of the electric vehicle.
[0038] Furthermore, the present disclosed system reduces the manual intervention involved in the tuning activity which sometimes generate inaccurate results as amount of tuning required for controlling the motor of the electric vehicle based on a predefined requirement is unknown and undefined and controls the tuning activity by self-learning to improve the performance of the electric vehicle.
[0039] It will be understood by those skilled in the art that the foregoing general description and the following detailed description are exemplary and explanatory of the disclosure and are not intended to be restrictive thereof.
[0040] While specific language has been used to describe the disclosure, any limitations arising on account of the same are not intended. As would be apparent to a person skilled in the art, various working modifications may be made to the method in order to implement the inventive concept as taught herein.
[0041] The figures and the foregoing description give examples of embodiments. Those skilled in the art will appreciate that one or more of the described elements may well be combined into a single functional element. Alternatively, certain elements may be split into multiple functional elements. Elements from one embodiment may be added to another embodiment. For example, the order of processes described herein may be changed and are not limited to the manner described herein. Moreover, the actions of any flow diagram need not be implemented in the order shown; nor do all of the acts need to be necessarily performed. Also, those acts that are not dependent on other acts may be performed in parallel with the other acts. The scope of embodiments is by no means limited by these specific examples.