Technical characteristics and wide application of embedded systems

Nuli-IT

Technical characteristics and wide application of embedded systems [Copy link]

Based on the concept of embedded system, this paper expounds the key technologies, embedded development and wide applications of embedded system. This paper analyzes the technical characteristics of embedded system, introduces it from two aspects: embedded processor and embedded operating system, and emphasizes some processing methods and processes that are different from other operating systems; on this basis, this paper expounds the development process of embedded software, and combines the author's embedded software development practice to emphasize some development skills of embedded software. This paper introduces some popular applications of embedded system at present, and finally gives the author's own experience and looks forward to the prospects of embedded system. In the post-PC era with the rapid development of digital information technology and network technology, embedded system has been widely penetrated into scientific research, engineering design, military technology, various industries and commercial culture and art, as well as people's daily life. With the further development and promotion of various embedded products at home and abroad, embedded technology is more and more closely integrated with people's lives. The concept of embedded system appeared around 1970. At that time, many embedded systems did not use operating system. They only used a simple loop control to process external control requests in order to realize a certain control function. When the application system becomes more and more complex and the scope of application becomes wider and wider, each new function may need to be designed from scratch. The lack of operating system has become the biggest disadvantage. The emergence of C language makes operating system development simple. Since the 1980s, a variety of commercial embedded operating systems have emerged, with the most famous ones being VxWorks, pSOS, and Windows CE. Most of these operating systems are developed for proprietary systems. In addition, the open-source embedded Linux has also been increasingly used recently due to its powerful network functions and low cost. 01—Technical characteristics of embedded systems Embedded systems usually include two parts: hardware and operating systems that constitute the basic operating environment of the software. The operating environment and application scenarios of embedded systems determine that embedded systems have some characteristics that are different from other operating systems. 1.1 Embedded processors Embedded processors can be divided into three categories: embedded microprocessors, embedded microcontrollers, and embedded DSPs (Digital Signal Processors). Embedded microprocessors are CPUs that correspond to the microprocessors of general-purpose computers. In applications, microprocessors are generally assembled on specially designed circuit boards, and only embedded-related functions are retained on the motherboard, which can meet the requirements of small size and low power consumption of embedded systems. Current embedded processors mainly include: PowerPC, Motorola 68000, ARM series, etc. Embedded microcontrollers, also known as single-chip microcomputers, encapsulate CPU, memory (a small amount of RAM, ROM or both) and other peripherals in the same integrated circuit. The most common one is 8051. Embedded DSP is specially used to process discrete time signals very quickly, improving compilation efficiency and execution speed. In the fields of digital filtering, FFT, spectrum analysis, image processing analysis, etc., DSP is entering the embedded market in large numbers. 1.2 Microkernel Structure Most operating systems are divided into at least two levels: kernel layer and application layer. The kernel only provides basic functions, such as establishing and managing processes, providing file systems, managing devices, etc. These functions are provided to users in the form of system calls. Some desktop operating systems, such as Windows and Linux, have introduced many functions into the kernel, and the kernel of the operating system has become larger and larger. The larger the kernel, the more resources it occupies, and it is very troublesome to trim it. Most embedded operating systems use a microkernel structure, and the kernel only provides basic functions, such as: task scheduling, communication and synchronization between tasks, memory management, clock management, etc. Other application components, such as network functions, file systems, GUI systems, etc., all work in user mode and work in the form of system processes or function calls. Therefore, the system is customizable, and users can choose the corresponding components according to their needs. 1.3 Task Scheduling In embedded systems, tasks are threads. Most embedded operating systems support multitasking. The implementation of multitasking is actually to switch and schedule between multiple tasks by the CPU. Each task has its priority, and different tasks may have the same or different priorities. There are three ways to schedule tasks: preemptive scheduling, non-preemptive scheduling, and time slice round-robin scheduling. Non-preemptive scheduling means that once a task obtains the CPU, it will run exclusively on the CPU unless it decides to give up the right to use the CPU for some reason; preemptive scheduling is based on task priority, and the currently running task can give way to other tasks with higher priority in the ready state at any time; when two or more tasks have the same priority, different tasks use the CPU in turn until the CPU time slice allocated by the system is used up, which is time slice round-robin scheduling. At present, most embedded operating systems use priority-based preemptive scheduling for tasks of different priorities, and time slice round-robin scheduling for tasks of the same priority. 1.4 Hard Real-Time and Soft Real-Time Some embedded systems have high requirements for time, which are called real-time systems. There are two types of real-time systems: hard real-time systems and soft real-time systems. Soft real-time systems do not require that a task must be completed within a certain time, but only require that each task run as fast as possible; hard real-time systems have strict requirements on system response time. Once the system response time cannot be met, it may cause system crashes or fatal errors. Generally, they are widely used in industrial control. 1.5 Memory Management Some desktop operating systems designed for processors with memory management units (MMUs), such as Windows and Linux, use the concept of virtual memory. The virtual memory address is sent to the MMU. Here, the virtual address is mapped to the physical address, and the actual memory is divided into pages of the same size, and the process is loaded in a paging manner. Before a program is run, it is not necessary to load all of it into the memory, but only those pages that are currently to be run are loaded into the memory for operation. Most embedded systems are designed for processors without MMUs, and cannot use the processor's virtual memory management technology. Instead, they use real memory management strategies. Therefore, the access to memory is direct. It does not need to go through the MMU to access the address, but directly sends it to the address line for output. The addresses accessed in all programs are actual physical addresses. Moreover, most embedded operating systems do not protect the memory space, and each process actually shares a running space. Before a process is executed, the system must allocate enough continuous address space for it, and then load it all into the continuous space of the main memory. It can be seen that the developers of embedded systems have to participate in the memory management of the system. Starting from compiling the kernel, the developer must tell the system how much memory this development board has; when developing applications, the memory allocation must be considered and the size of the running space required by the application must be paid attention to. In addition, due to the use of real memory management strategy, user programs are in the same address space as the kernel and other user programs. When developing programs, it is necessary to ensure that the address space of other programs is not violated, so that the program will not destroy the normal operation of the system or cause abnormal operation of other programs; therefore, the developers of embedded systems must be particularly careful about some memory operations in the software. 1.6 Kernel loading mode The kernel of the embedded operating system can be run directly on Flash or loaded into memory for operation. The way Flash works is to burn the kernel's executable image to Flash, and start execution from a certain address in Flash when the system starts. This method is actually adopted by many embedded systems. The kernel loading method is to store the kernel's compressed file on Flash, read the compressed file when the system starts, decompress it in the memory, and then start execution. This method is relatively complicated, but the running speed may be faster because the access rate of RAM is higher than that of Flash. Due to the memory management mechanism of the embedded system, the embedded operating system uses static linking for user programs. In an embedded system, the application and the operating system kernel code are compiled and linked to generate a binary image file to run. 02—Technology related to embedded system development Compared with developing applications in a Windows environment, embedded system development is very different. Different hardware platforms and operating systems bring many additional development complexities. 2.1 Embedded Development Process In the embedded development process, there are two roles: the host machine and the target machine: the host machine is the computer that performs the compilation, linking, and addressing processes; the target machine refers to the hardware platform that runs the embedded software. First, the application must be converted into binary code that can be run on the target machine. This process includes three steps: compilation, linking, and addressing. The compilation process is implemented by a cross compiler. The so-called cross compiler is a compiler that runs on one computer platform and generates code for another platform. Common cross compilers include GNU C++/C++ (gcc). All target files generated by the compilation process are linked into one target file, which is called the linking process. The addressing process assigns the physical memory address to each relative offset of the target file. The file generated by this process is a binary file that can be executed on the embedded platform. Another important step in the embedded development process is to debug the application on the target machine. Embedded debugging uses a cross debugger, generally using the host-target debugging method, which is connected by a serial port line, Ethernet, or BDM line. Cross debugging includes task-level, source-level and assembly-level debugging. During debugging, the application and operating system kernel on the host machine need to be downloaded to the RAM of the target machine or burned directly into the ROM of the target machine. The target monitor is the agent (Debugger Agent) that the debugger uses to control the application running on the target machine. It is pre-fixed in the Flash and ROM of the target machine. It automatically starts after the target machine is powered on and waits for commands from the debugger on the host machine. It cooperates with the debugger to complete the download, operation and basic debugging functions of the application and returns the debugging information to the host machine. 2.2 Migrate software to embedded platforms The software development model chosen by most embedded developers is to write software on a PC first and then migrate the software. When writing software on a PC, pay attention to the portability of the software, choose a programming language with high portability (such as C language), call operating system functions as little as possible, and pay attention to shielding byte order, byte alignment and other issues caused by different hardware platforms. The following are some of our experiences in migrating the protocol stack. 2.2.1 Byte Order Byte order refers to the order in which data that occupies more than one byte of memory is stored in memory. There are usually two byte orders: little endian and big endian. Little endian means that low byte data is stored at low memory addresses and high byte data is stored at high memory addresses; big endian means that high byte data is stored at low addresses and low byte data is stored at high addresses. PCs based on the X86 platform are in little endian byte order, while some embedded platforms are in big endian byte order. Therefore, for data types with more than 1 byte, such as int, uint16, and uint32, their storage order should be changed on these embedded platforms. Generally, we believe that the order of bytes transmitted in the air, namely the network byte order, is the standard order. Considering the consistency with the protocol and the interoperability with other similar platform products, when sending data packets in the program, the host byte order is converted to the network byte order, and when receiving data packets, the network byte order is converted to the host byte order. 2.2.2 Byte alignment The addressing mode of some embedded processors determines that data types such as int16 and uint16 that occupy 2 bytes in memory can only be stored at even memory addresses, and data types such as int32 and uint32 that occupy 4 bytes can only be stored at memory addresses that are integer multiples of 4; data types that occupy 8 bytes can only be stored at memory addresses that are integer multiples of 8; and data types that occupy only 1 byte in memory can be stored at any address. Due to these restrictions, there are great differences when programming on these platforms. First, there will be holes between structure members, such as this structure: typedef struct test{ char a; uint16 b; }TEST The structure TEST occupies three bytes of memory on a single-byte aligned platform, but may occupy three or four bytes on the embedded platform described above, depending on the storage address of member a. When the storage address of a is an even number, the structure occupies four bytes, and there is a one-byte hole between a and b. For both parties in communication, both operate on the structure members, this situation will not go wrong, but if one party reads the content byte by byte (most communication protocols do this), the content of other bytes will be read incorrectly. Secondly, if the data in the memory is read by forced type conversion, the difference in byte alignment will cause data reading errors. Because if the pointer points to the base memory address, we want to obtain the data occupying two bytes of memory and store it in a uint16 variable. The result of forced type conversion is to obtain the data at the address pointed by the pointer and the previous address, and not to obtain the data at the address pointed by the pointer and the next address as we wish, which leads to data reading errors. There are many ways to solve byte alignment, such as adding the compilation option --pack-struct in the GCC project management file MakeFile; but this method can only remove the holes in the structure, and cannot solve the errors caused by forced type conversion. In order to enhance the portability of the software and interoperability with other similar platform products, we have added a function for unpacking at the receiving data packet and a function for assembling at the sending data packet. These two functions solve the problem of byte order and byte alignment. That is, when assembling a packet, a certain segment of data of different data types in the memory is placed at the specified address according to the format string in the parameter, and the packet is sent to the lower layer; when unpacking the packet, the received data in the memory is stored in different types of variables or structure members according to the format string in the parameter. Different data types are processed differently in the function. 2.2.3 Bit segment Since the spatial allocation direction of the bit segment varies with different hardware platforms, for the X86 platform, the bit segment is allocated from right to left; while for some embedded platforms, the bit segment is allocated from left to right. The different allocation order leads to data access errors. One way to solve this problem is to use conditional compilation to define bit segments of different orders for different platforms; or to add bit segment processing to the two functions described above. 2.2.4 Code optimization Embedded systems have higher requirements for the quality of application software. Therefore, in embedded development, special attention should be paid to code optimization to improve code efficiency as much as possible and reduce code size. Although modern C and C++ compilers provide a certain degree of code optimization, most of the optimization techniques performed by the compiler only involve the balance between execution speed and code size. It is impossible to make the program both fast and small, so necessary measures must be taken when writing embedded software. 1) Improve code efficiency ① switch-case statements. Switch-case statements are often used in programs. Each test and jump implemented by machine language is just to decide what to do next, which wastes processor time. In order to improve speed, specific situations can be sorted according to their relative frequency of occurrence. That is, put the most likely situation first and the least likely situation last, which will reduce the average code execution time. ② Global variables. Using global variables is more efficient than passing parameters to functions. This eliminates the need to push parameters into the stack before the function call and pop parameters out of the stack after the function is completed. Of course, using global variables will have some negative effects on the program. 2) Reduce the size of the code Embedded system programming should avoid using standard library routines, because many large library routines try to handle all possible situations, so they take up a lot of memory space, so the use of standard library routines should be reduced as much as possible. 2.3 Avoid memory leaks User memory space (heap) is the remaining space after global data and task stack space are allocated in RAM. In order for the program to have enough memory to run, the memory applied for must be released in time after it is not used, so as to ensure that there is space when it is applied again. If there is a memory leak in the program (that is, the memory is not released in time after application), the program will eventually fail to run due to insufficient memory space. 03—Wide application of embedded systems The application prospects of embedded systems are very broad. People will be exposed to embedded products at all times and everywhere, from washing machines and refrigerators at home, to bicycles and cars as means of transportation, to remote conference systems in offices, etc. In particular, the emergence of small-scale wireless access protocols represented by Bluetooth has quietly raised the concept of embedded radio. When the price of embedded radio chips is acceptable, its application may be ubiquitous. At home, in the office, and in public places, people may use dozens or even more such embedded radio chips to form a wireless network for some electronic information equipment and even electrical equipment; in the car and on the road, people can use such embedded radio chips to realize remote office and remote control, and truly carry the network with them. The following are some specific applications. 3.1 Embedded Mobile Database The so-called mobile database is a database that supports mobile computing, which has two meanings: ① Users can access database resources online while moving. ② Users can carry the database with them. Typical applications include checking the nearest hospital in a moving ambulance. The system consists of a front-end mobile terminal and a back-end synchronization server. The mobile terminal has an embedded real-time operating system and an embedded database. Renmin University of China is developing this system. They have cooperated with Hopen to develop the front-end mobile terminal "Elf". 3.2 Application of embedded systems in smart home networks Smart home network (E-Home) refers to the establishment of a communication network in a home to provide the necessary access for family information; under the control of the home network operating system, through the corresponding hardware and actuators, all home appliances and devices on the home network can be controlled and monitored. The network structure must include a home gateway. The home gateway mainly realizes the signal integration of the control network and the information network and interfaces with the outside world for remote control and information exchange. Whether it is a gateway or a control module on each home appliance, an embedded operating system is required. These operating systems must have the characteristics of embedded, good real-time performance, and multi-user. The intelligent multimedia home gateway developed by Nanjing Dongda Mobile Internet Technology Co., Ltd. uses embedded Linux as the operating system of the embedded device. The mutual communication between devices complies with the Bluetooth communication protocol, which can support multiple devices to access fixed telephone networks, the Internet and other external networks at the same time. 3.3 Embedded Voice Chip The embedded voice chip is based on an embedded operating system and adopts technologies such as voice recognition and speech synthesis, phonetics hierarchy system and text processing model. It can be applied to multiple fields such as handheld devices and smart home appliances, giving these devices humanized interaction methods and convenient usage methods. It can also be applied to toys to realize voice-controlled toys, simulated pets, and toys that talk to people. It can also be applied to vehicle-mounted communication equipment to realize human-computer communication. This chip is used in mobile communication devices. For example, when a short message comes on the mobile phone, we don’t have to look at it laboriously, but can hear the sound. 3.4 Embedded products based on short-range wireless communication protocols The combination of short-range wireless access protocols represented by Bluetooth and embedded systems will surely promote the widespread application of embedded systems. Recently, embedded products based on these protocols have emerged in an endless stream, including various telephone systems, wireless briefcases, various digital electronic devices and applications in e-commerce. These products, with their miniaturization and low cost, have opened up broad prospects for their application in home and office automation, e-commerce, industrial control, intelligent buildings and various special occasions. Conclusion - The above focuses on the characteristics of embedded systems, the development and application of embedded products. During the development process, we realized that there is essentially no difference between embedded system development and previous development work. The only change is that each hardware platform is unique. This difference leads to many additional development complexities. Therefore, in the embedded development process, special attention should be paid to the software creation process; moreover, before developing embedded products, we should have a better understanding of the selected embedded hardware platform, have the corresponding hardware knowledge, and work closely with hardware engineers; when selecting embedded operating systems and hardware platforms, we should choose appropriate systems and platforms based on the needs of the applications to be developed and cost considerations. With the rapid development of science and technology today, embedded products will be more and more widely used. We believe that as long as we follow the development rules of embedded products and adapt to market needs, we will be able to develop more and more embedded products.

Jacktang

There are many ways to improve the efficiency of code, which is also related to personal skills