Autonomous driving is generally divided into four basic modules: perception, prediction, planning, and control. Each module performs its own duties and has a clear division of labor. Today, Autobots will talk about planning and control with you. In general, planning controls serve two purposes:
One is to carry out global planning for vehicle movement (running the route from point A to point B), behavioral decision-making (judging whether to change lanes or overtake, etc.), local planning (planning local driving trajectories, avoiding obstacles, etc.);
The second is to precisely control the vehicle to drive according to the planned trajectory.
As the most downstream module in the entire autonomous driving/robot algorithm process, planning control directly determines the safety and comfort of autonomous driving. A good regulation directly affects the driving experience of the driver and passengers: whether the turn is smooth, how the car in front brakes when following the car, how to deal with the car in front, whether the vehicle can be merged into the planned lane in a timely manner when merging at a complex intersection, etc.
Behind these complex controls is the regulation and control module that plays a role. Many friends who are just getting started are at a loss. What is regulation and control? What are the subfields? What algorithms are included in each field? How are these algorithms implemented? What are the advantages and disadvantages? Which scenarios are applicable? On the whole, the vehicle starts from path planning, and makes corresponding decisions (following, changing lanes, accelerating, etc.) according to the information of perception and positioning during the driving process. Then, according to the results of the upstream decision-making, the motion planning module outputs the corresponding trajectory information, speed, acceleration and steering wheel angle information in real time, and then the control module performs corresponding control.
Moreover, with the country's strong support for the development of the new energy vehicle industry, major companies have also increased their recruitment for relevant positions. I just saw a recruitment website, and the average monthly salary of related positions has reached more than 40,000, and the annual salary is 600,000. There are many high-level positions with an annual salary of one million!
Early bird waits for no one! Scan the code to get discounts and join the study!
difficult to learn
Overall, planning control requires more theoretical knowledge than perception. At present, the mainstream regulation and control algorithms in the corporate world include control algorithms such as PID, LQR, and MPC, as well as planning algorithms such as A*, Hybrid A*, Lattice Planner, and EM Planner. During this period of time, many small partners have consulted about planning and control related issues. In fact, we are also very interested in planning and control. The quality of learning materials related to planning and control on the market is uneven. Stepped on more pits:
There are many types of planning control algorithms, and there are no systematic learning materials on the Internet. Students who are just getting started don't know where to start, and the papers are half-knowledge...
Do not understand the advantages and disadvantages of various planning control algorithms, and do not know which algorithm to choose in different scenarios
I don’t know what kind of technology stack talents are lacking in the industry, and it is easy to grasp the wrong direction in the process of self-study
After analyzing the pain points of everyone in the learning process, the Heart of Automated Driving and the regulation and control engineer of a major manufacturer in the industry jointly polished the online course "Planning Control Theory and Practical Course". If you want to get started with planning control, have an in-depth understanding of algorithm principles, or need to improve your technical capabilities in this area, do not know how to optimize, and lack practical experience in projects, then you must study this course. The course content introduces planning algorithms in detail Basic knowledge, horizontal and vertical decoupling/joint decision planning framework and commonly used control algorithms (PID, LQR, MPC, etc.) .
The course starts from the outline and definition of the most basic planning control module, and then goes to the explanation of the basic knowledge of planning algorithms, involving related planning algorithms based on search/sampling/vehicle kinematics/numerical optimization, and then to the explanation of decision-making planning framework (horizontal and vertical solutions) coupling, horizontal and vertical combination), and finally explained several commonly used control algorithms (PID, LQR, MPC) and discussed the challenges faced by PnC. The actual combat involves Dijkstra, A*, RRT*, State Lattice Planner, QP path optimization and based on Trajectory tracking algorithm for MPC .
Let’s take a look at the outline of this course first, it’s full of dry goods, and it really helps students with zero foundation to learn efficiently and quickly master every knowledge point
Combination of actual combat and theory of the project, and the actual combat code after class of the actual combat course, which can be mastered quickly after learning and practicing.
A total of 5 major combat projects
The course includes a complete [teacher teaching] + [teaching assistant answering questions] service to ensure that every little partner can learn knowledge happily.
Combat 1: Achieving A*, Dijkstra, a must for school recruitment interviews;
Combat 2: Realize the RRT* algorithm;
Combat 3: Implement State Lattice Planner;
Combat 4: Realize the path optimization algorithm based on QP;
Combat 5: Realize the trajectory tracking algorithm based on MPC!
Courseware codes are readily available
Detailed explanation, not only the theory, but also the code and practice must be explained thoroughly!
Through a full set of video explanations, it will help you build the basic framework of the model in your mind, so that you can thoroughly understand every knowledge point, thereby improving your writing speed.
Instructors
Ning Yuan, a member of the cutting-edge technology research team of the heart of automatic driving, has been deeply involved in the field of automatic driving algorithms for many years. He is currently a senior algorithm engineer in the team of a leading automatic driving company in the industry. He has rich experience in the research of automatic driving planning control algorithms and engineering implementation.
Course harvest
Have a deep understanding of the theoretical basics of planning algorithms, and have greatly improved in code implementation;
Have an in-depth understanding of the decision-making planning framework, and master the common horizontal-vertical decoupling and horizontal-vertical joint planning frameworks;
Master commonly used control algorithms (PID, LQR, MPC);
After completing this course, you can reach the level of an automatic driving regulation engineer for about one year;
Get to meet many industry practitioners and study partners!
suitable for the crowd
Bachelor/Master/PhD degree in vehicle engineering, automation, automotive electronics, computer science, software engineering, motion control, etc.;
Algorithm engineers related to autonomous driving planning and control;
Friends who want to transfer to the automatic driving regulation algorithm;
The basics required for this course
Have a certain programming foundation: C/C++ or Python;
Certain basics of advanced mathematics, linear algebra and matrix theory;
Class time and learning style
On July 18, 2023, the road to learning officially began. After two months, offline video lectures were given. The lecturer answers questions in the WeChat learning group, and solves problems such as algorithms, codes, and environment configuration in the course one by one!
course consultation
Early bird waits for no one! Scan the code to get discounts and join the study!
Scan the code to add assistant consulting courses!
(WeChat: AIDriver004)
