【Jiaqin Comment】The truck automatic driving scheme invented by Zhijia Technology makes the truck driving control algorithm have strong anti-interference and robustness through the linear control method, and at the same time compensates according to the estimated disturbance size, which can strengthen the consistency of the horizontal control, make the response of the truck closer to the design indicators, and also improve the comfort of the user.
Jiwei network news, Zhijia Technology recently announced a cooperation with Aeva, Aeva will help Zhijia Technology autonomous driving heavy trucks to clearly perceive the environment at a long distance, reduce response delays, and use Aeva's unique real-time speed information to deal with safety hazards and extreme road scenarios.
Why is self-driving technology for trucks more complex than other models? This is determined by the structure of the truck, as shown in the figure above, the structure of the truck is divided into two parts: the tractor and the trailer, of which the center of the rear axle is the particle of the tractor, and the center of the rear axle of the trailer is the particle of the trailer. Usually, the full load of the truck is 20-25 times that of the ordinary car, the length of the truck is 4-5 times that of the ordinary car, and because the structure of the truck is not a uniform whole (including tractors and trailers), so in the road driving process, the truck's control of the vehicle itself is more complex and more demanding than that of the car.
Trucks are driving in the lane, and in addition to keeping the tractor in and out of the lane, you also need to keep the trailer in the lane. Due to the different weights of the truck, the half-load, and the full load, the driving speed is different, if the driving operation is improper during the driving process, such as improper control of the steering wheel angle and steering angle speed, there will be a hanging phenomenon of the truck in the straight lane driving process, and it may also scratch during the cornering process.
In the existing scheme, the advanced automatic driving technology of the truck is often difficult to accurately model the truck, and it is difficult to accurately estimate the changes in the different mass center of mass and so on, and at the same time, in the process of control, there are external disturbances affecting the control of the vehicle, such as wind speed, road conditions, etc., which are external disturbances to the vehicle control system, affecting the performance of the control, and even making the system unstable, resulting in safety accidents.
In order to improve the automatic driving control effect of trucks and reduce the safety risks of truck driving, Zhijia Technology applied for a technical solution entitled "A Lateral Control Method, System and Storage Medium for Autonomous Trucks" (application number: 201911328780.1) on December 20, 2019, and the applicant was Suzhou Zhijia Technology Co., Ltd.
According to the technical solution currently disclosed in the patent, let's take a look at this technical solution.
As shown in the figure above, it is a schematic diagram of the principle of the lateral control method of the self-driving truck invented in the patent, and the algorithm of the method is mainly composed of two parts: the expansion state observer ESO and the LQR controller. By controlling the signal and the measurement signal of the vehicle, the extended state observer can directly measure the state quantity, and estimate the state amount and disturbance that are difficult to measure, and then compensate for the external disturbance to achieve the purpose of eliminating external disturbance and vehicle model uncertainty disturbance.
After receiving the reference trajectory of the upstream input, the LQR controller finds the trajectory point closest to the truck and calculates the tracking error e, and then uses the tracking error and the information fed by the expansion state observer ESO as the state variable X of the linear quadratic regulator. By designing a properly designed LQR closed-loop feedback controller, it is possible to ensure that the design system has sufficient phase margin and amplitude margin.
As shown in the figure above, a schematic diagram of the flow of this lateral control method for autonomous trucks, the system is initialized with a linearized vehicle dynamics model, the time-varying parameters of the linear quadratic regulator, and the parameters of the expansion state observer. Secondly, you need to enter the corresponding motion trajectory, and find the closest point to the truck, and calculate the error state variable, that is, the tracking error.
Then, based on the truck's state and the control input at the previous moment, an estimate of the truck's state and an equivalent estimate of the disturbance are obtained by the expansion state observer. On the basis of these data, the linearization model of the truck can be used to calculate the control law of the linear quadratic regulator.
Finally, when the overall disturbance compensation is obtained by using the expansion observer observation, the control output of the compensation disturbance is filtered and limited, and then sent to the truck wire control at a fixed frequency as a control signal, and the control signal here is the steering wheel command. By cycling through the above steps, the output of the truck is consistent with the desired output, resulting in better driving control.
The above is the truck automatic driving scheme invented by Zhijia Technology, which makes the truck driving control algorithm have strong anti-interference and robustness through linear control methods, and compensates according to the estimated disturbance size, which can strengthen the consistency of horizontal control, make the response of the truck closer to the design indicators, and also improve the comfort of the user.)
(Proofreading/Holly)