Development of motor drive control system based on AUTOSAR

1. Introduction

Pure electric vehicles and hybrid electric vehicles are important directions for the development of the new energy vehicle industry. At the same time, Pan Asia's "electrification, intelligence, networking, and digitalization" strategy has made the development of future automotive electronic and electrical architecture systems more and more complex. The hierarchical, modular, and platform-based technologies represented by the automotive open system architecture AUTOSAR are important trends in automotive electronic software development. Among the three major electronic control systems of electric vehicles (motor control, battery management, and vehicle control), motor control is one of the cores, and the research and design of its software architecture is of great significance to the development of automotive electronic control systems. This report takes the drive motor for electric vehicles as the research object, and based on the AUTOSAR development architecture, explores the design and development of the software architecture of the motor drive control system, and on this basis, conducts a detailed study of the motor overmodulation control algorithm and the resolver soft decoding technology.

2. The significance of developing electric vehicle motor control software based on AUTOSAR

Among the three major electronic control units of electric vehicles, motor drive control is the core, and its performance has a direct impact on the vehicle's power and maneuverability. Compared with traditional motor speed control systems and servo motor systems, the development of automotive drive motor systems has higher requirements for safety and reliability in addition to performance requirements such as high power density and wide speed range. Improving the reusability of automotive motor control software, enhancing the configurability of system software, and improving the reliability and stability of system software are of great significance to the development of automotive motor control systems.

3. Research on resolver decoding

For motor vector control, it is often necessary to obtain the rotor position angle of the motor. Common methods for measuring angles include magnetic encoders, photoelectric encoders, eddy current sensors, and rotary transformers. Among them, the magnetic encoder is a shaft angle sensor based on the magnetoresistance effect or the Hall effect. The output signal is a sine and cosine function based on the rotor position. It has a simple structure and strong robustness. It is not affected by humid environments, but is limited by high temperatures and air gaps. The photoelectric encoder is small in size, high in resolution, and has strong anti-electromagnetic interference capabilities, but the speed is limited, and the maximum measurable speed is around 3000rpm. The eddy current sensor has high sensitivity, fast response speed, and is less affected by the environment, but its accuracy is limited. The rotary transformer has high reliability, strong adaptability to different environments, and is not affected by factors such as temperature and vibration. Therefore, it is widely used in servo systems and industrial automation fields such as elevators, radars, and airborne instruments.

The resolver outputs a set of sine and cosine signals containing rotor position information, which need to be properly processed to obtain the corresponding rotor position. The processing of the resolver output analog signal can be converted into a digital signal by a dedicated integrated circuit, namely the RDC circuit, also known as the resolver decoding chip. Currently, the common ones are the AD2S (AD2S1210) series chips of AD Company in the United States and the AU6802 series chips of Tamagawa Company in Japan. They can generate an excitation signal and send it to the excitation winding of the resolver, and then process the analog signal of the return winding to obtain the rotor position signal, which is output in the form of an encoder or SPI. In addition, there is another method that uses DSP and peripheral conditioning circuits to generate excitation signals and modulate, filter and process the signals output by the resolver as well as the angle identification algorithm to obtain the rotor position signal. The use of DSP for angle signal processing eliminates the need for a dedicated decoding chip, greatly reducing development costs.

Common angle identification algorithms include the inverse trigonometric function method, the calibration table lookup method, and the phase-locked loop-based angle tracking observer method. Among them, the inverse trigonometric function method is simple to implement, but it introduces a division and an inverse tangent operation, which takes up more resources and cannot use the waveform of the entire output signal; the calibration table lookup method requires the use of sensors to calibrate the resolver signal and store the angle value corresponding to the resolver output signal for table lookup; the phase-locked loop-based angle tracking observer method includes a second-order angle observer and a third-order angle observer. The second-order angle observer method has a certain filtering effect, improves anti-interference ability, and can simultaneously estimate the rotor position and speed value of the motor. When the speed changes (speed up or speed down), the rotor position angle calculated by the second-order observer will have a steady-state error. The third-order angle observer has better transient detection performance and is often used for rotor position observation when the motor starts and accelerates.

4. Development of motor drive control system based on AUTOSAR

This section mainly studies the software layering of motor control system software according to the AUTOSAR standard, including application layer, basic software layer and real-time operating environment. The application layer builds the motor control model and generates software code through MATLAB/Simulink, the basic software layer realizes the peripheral function module drive by calling the Infineon underlying driver software package programming, and the real-time operating environment realizes data exchange and service calls between different software layers through custom interface functions.

In traditional motor control, especially in the field of automation, the software development layering is not obvious, and the nested relationship between software and hardware is seriously coupled. This makes it difficult to determine whether the cause of the problem lies in the hardware driver or the software algorithm design once the software encounters problems and defects during the development process, which affects the efficiency of software development, increases development costs, and also makes the software have potential risks. AUTOSAR is an automotive open architecture standard that realizes the separation of software and hardware by software layering and module division, and divides the system software architecture into application software layer (APP), basic software layer (BSW) and real-time operating environment layer (RTE). As shown in the figure below:

In the motor control system, the motor control algorithm, motor state detection, safety monitoring and other functions related to the motor control function can be used as applications corresponding to the application software layer in the AUTOSAR architecture. In the application software layer, different functional modules are encapsulated in the form of software components SWC, which is convenient for designers to further develop. The basic software layer corresponds to the controller peripheral drivers, communication services and interrupt services related to motor control, and the controller hardware functions are encapsulated in modular form, which is conducive to its transplantation between different hardware platforms. The data exchange and service calls between the application software layer and the basic software layer are realized through the real-time operating environment layer. RTE is equivalent to a virtual function bus, so that developers do not need to consider the underlying hardware and communication network when designing application software, and focus on the design of functional software. The figure below shows the layered design of the electric vehicle drive motor ECU software architecture in this article. The following will carry out a detailed design of each layer of software based on this architecture.

5. Application layer software design (ASW)

The application layer software is the specific implementation of the motor control algorithm and its safety monitoring functions. It is necessary to first determine the input and output data of the system, the software components SWC contained in the system and its system constraints. The ECU of this system is the motor controller, and its software components can be divided into: motor control algorithm SWC, data solution SWC, safety monitoring function SWC and other software components according to the corresponding functions, as shown in Figure 2.3. At the same time, it is necessary to determine the data input and output of each software component and the data exchange and service call between software components. In the application software layer APP, the purpose of software component division is to reduce coupling through functional module division, which is conducive to the update and upgrade of system software.

After completing the design of each software component, it is necessary to design the Runnable Entity RE (Runnable Entity). The Runnable Entity is the smallest code fragment of the software component and the specific implementation of the software component function. Among them, in the data solution software component SWC, the input data includes current sampling and rotor position sensor signals obtained from the basic software layer. The corresponding motor phase current, speed and rotor position signals need to be obtained through the current calculation, speed and rotor position solution of the Runnable Entity RE.