Embedded general learning route arrangement

Hello everyone , I'm wheat .

Engaged in embedded related industries , Almost soon 10 Years time. , Go a lot of detours , Stepped on many pits .

A lot of people will ask , Does embedded really have no future ？ In fact, I can't answer this . From the perspective of development , Everything has a cycle .

This has nothing to do with embedded , Industry related , The growth of an industry , Related to capital , It is related to Guo Jia's development . Therefore, only the right time, place and people , And this man and , It's your choice .

This is my view , You are also welcome to express your views .

If you plan to get started with embedded learning , Let's talk about the embedded learning route .

What is embedded ？

Embedded is embedded system ,IEEE（ Institute of electrical and Electronics Engineers ） It is defined to control 、 A device for monitoring or assisting in the operation of machines and equipment , It's a kind of A dedicated computer system .

The commonly accepted definition of embedded system in China is Application centric , With Based on computer technology , Software and hardware can be tailored , Adapt application system to function 、 reliability 、 cost 、 Volume 、 Special computer system with strict requirements on power consumption .

Embedded is a comprehensive direction of many disciplines , There are many subdivisions . In fact, it can directly cut into a specific field , Then introduce the learning route , For example, MCU development ,Linux application development ,Linux Drive development .

But it's not hard to find , Although there are many subdivisions , But the knowledge that needs to be learned overlaps , Some are independent .

Embedded corresponding recruitment positions , The basic skill tree requirements are not much different , However, there are many jobs that are deeply bound to the industry , It will require industry experience and relevant skills precipitated in this industry ;

• For example, automotive electronics requirements can,autosar;

• Internet of things requires network protocols , Single chip microcomputer ,RTOS,WiFi,Bluetooth wait ;

• BSP The engineer requires to master Linux Kernel and various subsystems ;

• Robot related requirements ROS;

There are many more positions , Here are just a few more , Now I'd like to share what I think is the best way to learn .

Here is my mind map . To put it simply, the traditional learning route .

Programming based

C Language

This must be mastered .C The language is very flexible , It can be used to develop single chip computers , Drive development , Firmware development , Kernel development , application development , Almost can be used .

For example, the famous operating system Linux, Mainly C The realization of language , Of course, there are many others based on C Language project , I won't introduce them here .

Study C Language , I prefer the courses of Weng Kai from Zhejiang University .

C++

C++ In embedded, it is more about application development , Because of C Good compatibility , So it is still used in hardware C Programming , And the more complex business logic part , Object oriented C++ Encapsulate and implement .

C++ and C Is also very practical .

Data structure and algorithm

Use data structures rationally , It will get twice the result with half the effort , Commonly used data structures , Like queues , Linked lists are actually very common in embedded development , Of course, hashmap, Binary trees and so on .

If programming languages are weapons , So data structure is a move . What kind of scene , What kind of moves to use , It is roughly written in the data structure .

Other programming languages

There are many other programming languages , Here are a few .

shell

This is a scripting language , stay Linux It is quite common in , Generally speaking , master shell, Can greatly improve work efficiency , This is highly recommended to master , You can read more if you don't understand man page;

python

Very convenient , Greatly improved efficiency . Used to verify some algorithm functions , Or office automation .

Hardware knowledge

The hardware knowledge here is not for the route of pure hardware development , But do embedded software development , What hardware knowledge is needed , Let's talk about it in detail ;

Hardware foundation

Have the most basic hardware knowledge , If it is Do driver development , such as BSP The engineer , There must be two key points , Can understand the chip Datasheet, Can understand the schematic diagram , Because we need to develop the board machine .

Common communication protocols

such as UART,I2C,SPI Protocol, etc , These three are relatively common , It's relatively simple . There are many other communication protocols , I won't go into it here .

Common on-chip peripherals

such as gpio,adc,dac,watchdog,timer,pwm these , The principle is the same , But in general SOC Basically, they are all similar peripherals .

Use of common tools

Oscilloscope , A multimeter , The use of power , At least don't often connect the positive and negative poles wrong . The oscilloscope is convenient for checking the timing , It is convenient to troubleshoot problems during development .

Embedded system

Embedded operating system , It is usually divided according to specific requirements , Common real-time systems are ucos,freertos,rtthread,vxworks wait , and Linux The system is also very commonly used ;

Real time systems

Generally speaking , Real time systems are mostly used in control situations , such as ucos,freertos,rtthread,vxworks wait , Are real-time systems .

Linux System

Linux The system is widely used , Smart hardware , Monitoring equipment , Router , mobile phone , Cars, etc. , All have its figure . The general direction is mainly divided into Linux Drive and Linux application , The boundary between the two has been blurred . The technical architecture is divided into user space and kernel space , As shown in the figure below ;

If it's application development , In fact, it is similar to pure software development .

Basic content , In fact, there is also a part mentioned above ,Linux System programming , Multithreading , Basic knowledge at the operating system level , Mutually exclusive , Semaphore , Network programming and so on .

Specific application , For example, the development of audio and video , Motion control , Robot development and so on ;

Recommend 《Unix Environment advanced programming 》, This is the foundation. .

drive

The requirements for mastering the underlying knowledge are relatively high , At least trigger Linux Compilation of , Tailoring , Transplant and so on .

According to specific hardware requirements , Add the corresponding driver to the system , Because the system has specified the corresponding drive framework , Therefore, we need to understand various kernel subsystems and various types of drivers . For example, character device driver , Block device driver , The use of the device tree .

Deep learning subsystem is also a difficulty , Can help better understand Linux, For example, process scheduling （SCHED）、 memory management （MM）、 Virtual file system （VFS）、 Network interface （NET） And interprocess communication （IPC）.

Compare the recommended learning resources , Teacher songbaohua Linux kernel 4.0 Drive books , Teacher weidongshan's driving video , Internet age , Just search it and you'll have .

All in all , Technology lies in accumulation .

other

So much for that , If you want to keep the technology refined , Need to ensure continuous learning , See more excellent source code , Find excellent open source projects .

Unconventional learning route

After mastering some basic knowledge , You can also try this learning route , Study in the form of projects .

Set a target project , for example ;

• stay Linux Implement a program on the system , If you have to deal with hardware , For example, make a face recognition access control system ;

• For example, it should be transplanted on a board Linux System , And add drivers ;

• Using SCM to develop a small project , Complete the design of software and hardware ;

These are three different directions , Because everyone's specific situation is different , You can do it according to your own technology stack and hobbies , I can't think of doing anything , Find some open source projects , Engrave it yourself , It's also very good .

With a goal , Disassemble the task , There will be many knowledge blind spots in this process , I will use a lot of knowledge I mentioned above , At this time, you have to study , To absorb , Study and practice with purpose , The effect will be better .

summary

This article generally introduces the embedded learning route , Relatively broad and brief , Because the embedded knowledge system is really huge , Write purely from a personal perspective , There are many unilateral and limited places .

Hope to help some students who want to join this direction , If there are errors or improvements in this article , Your suggestions are welcome .