CN112026757A - Self-adaptive active anti-collision brake system with autonomous training and learning function - Google Patents

Abstract

The invention provides a self-adaptive active anti-collision brake system with autonomous training and learning, which belongs to the technical field of vehicle auxiliary driving and mainly comprises 6 parts, namely a millimeter wave radar sensor, a monocular optical sensor, a brake motor controller, a precise stepping brake motor, a Beidou high-precision combined navigation communication terminal and a display terminal; the invention can solve the defects that the existing AEBS control parameter is single, and the load change of a commercial truck, the adhesion coefficient change of the commercial truck in rainy and snowy weather, the long-term wear and aging of a vehicle braking system and the single and unchanged AEB control parameter cannot be met. The driver is provided with more comfortable, safe and reliable early warning and emergency braking experience.

Description

An adaptive active anti-collision braking system with autonomous training and learning

technical field

The invention belongs to the technical field of assisted driving of vehicles, in particular to an adaptive active anti-collision braking system with autonomous training and learning.

Background technique

Vehicle Active Collision Avoidance System (AEBs) is a kind of vehicle assisted driving system. It is mainly reflected in the fact that when an emergency occurs, the driver does not ignore the road surface due to factors such as fatigue and distraction, and there are hidden dangers of accidents. AEBS will first use various means to timely. It prompts the danger signal that appears on the road, and at the same time, when the driver fails to give the proper braking feedback in time, the emergency braking is started, so as to avoid the accident or reduce the loss of the accident. The existing AEBS system, regardless of whether it is installed in the front or in the rear, ignores the influence of vehicle weight changes, road conditions, weather changes, etc. on the vehicle's braking ability, while a fixed set of AEBs parameters has many incompatibility in different scenarios. .

SUMMARY OF THE INVENTION

In order to realize the adaptive adjustment of AEBS control parameters with the environment, the present invention provides an adaptive active anti-collision braking system with autonomous training and learning.

The technical scheme that realizes the present invention is as follows:

An adaptive active anti-collision braking system with autonomous training and learning, mainly composed of millimeter-wave radar sensor, monocular optical sensor, brake motor controller, precision stepping brake motor, Beidou high-precision integrated navigation communication terminal, and 6 display terminals parts;

Millimeter wave radar sensor to detect forward obstacle information and transmit it to Beidou high-precision integrated navigation communication terminal;

Monocular optical sensor, used to detect forward obstacle information and road lane line information, and transmit it to Beidou high-precision integrated navigation communication terminal;

The Beidou high-precision integrated navigation communication terminal receives the obstacle and road lane line information, and fuses the measurement information of its own two sensors. On the one hand, it outputs the distance and relative speed between the vehicle and the most dangerous obstacle, and the road Lane line position; on the other hand, the position, speed, and acceleration information of the vehicle are extracted, and output to the brake controller for evaluating the strength of braking;

The brake motor controller, on the one hand, receives the distance and relative speed information between the vehicle and the most dangerous obstacle, sends out an early warning signal to the display terminal according to the distance collision time TTC, and at the same time, according to the TTC information and vehicle speed, according to the pre-designed intensity control Accurate stepper brake motor for emergency braking; on the other hand, according to the current vehicle speed, combined with the vehicle braking capability, it generates the appropriate braking signal corresponding to spot braking/full braking, and uses the speed provided by the Beidou high-precision integrated navigation communication terminal and acceleration information to estimate the current braking signal, estimate the deviation range between the actual braking acceleration and the configured acceleration parameters, and record, and use the accumulated deviation range to analyze the steady-state change of the vehicle's braking. The braking parameters are inverted, and the influence caused by the load, weather changes and the running time of the braking mechanism is corrected according to the actual situation, and the corrected control parameters are sent to the cloud through the Beidou high-precision integrated navigation communication terminal; The confirmed braking parameters are used for braking control at this time;

Under the control of the brake controller, the precise stepping brake motor realizes the braking force output of the brake pedal;

The HMI display terminal realizes sound and light warning under the control of the brake controller.

Further, the HMI display terminal of the present invention also displays the distance and type of obstacles within the lane line.

Beneficial effects:

First, the present invention can solve the shortcomings of the existing AEBS control parameters, which cannot meet the change of commercial truck load, the change of the adhesion coefficient of commercial vehicles in rainy and snowy weather, the long-term wear and aging of the vehicle brake system, and the single and unchanged AEB control parameters. Provide drivers with a more comfortable, safe and reliable warning and emergency braking experience.

Second, the present invention avoids the complex model analysis of the braking force model of commercial vehicles, and the complex model design of multiple factors such as vehicle load, vehicle mass center distribution, road conditions, brake drum/disc wear conditions, etc. The actual braking force produced by the sub-controllable braking control is accurate and timely.

Third, in the design of the present invention, the AEBS control parameters of the cloud to be sent and uploaded to be updated are considered, and the parameters can be checked and verified repeatedly.

Description of drawings

FIG. 1 is an overall block diagram of an active anti-collision braking system with autonomous training and learning according to the present invention.

FIG. 2 is a frame diagram of the Beidou high-precision integrated navigation communication terminal unit of the present invention.

Fig. 3 is a flow chart of the signal control flow of the AEB configuration parameter self-learning of the present invention.

Detailed ways

To make the purposes, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention.

As shown in FIG. 1 , this embodiment provides an adaptive active anti-collision braking system with autonomous training and learning, which specifically includes a 77GHz millimeter-wave radar sensor, a monocular optical sensor, a brake motor controller, a precision stepping brake motor, The Beidou high-precision integrated navigation communication terminal and display terminal consist of 6 parts.

The 77GHz millimeter-wave radar sensor is used to detect forward obstacle information, including the distance and relative speed of the obstacle, and transmit it to the Beidou high-precision integrated navigation communication terminal; the 77GHz millimeter-wave radar sensor is a high-frequency electronic scanning radar. The frequency band is 76-77GHz, and it has both medium and long-distance scanning capabilities. Long distance 175m, forward angle ±10°, close range 60m, forward angle ±45° detection range.

Monocular optical sensor, used to detect forward obstacle information and road lane line information, etc., and transmit it to Beidou high-precision integrated navigation communication terminal; monocular optical sensor, based on megapixel high-definition color camera and high-performance SOC embedded platform , through the rational use of resources such as multi-core heterogeneous and accelerators, to meet the needs of deep learning computing, and to achieve high-performance vision products based on deep learning. Using image recognition, tracking, risk behavior assessment and other high-tech, to achieve based on lane line detection, vehicle detection, pedestrians and cyclists, traffic sign recognition and other applications.

Millimeter-wave radar and monocular optical composite detection of obstacles can effectively make up for the shortcomings of false alarms and poor ranging accuracy caused by a single sensor, and better improve the complex conditions of the system during vehicle driving. Millimeter-wave radar cannot distinguish the height of obstacles ahead, and non-obstruction targets such as manhole covers and height-limiting poles will be mistakenly regarded as vehicles or pedestrians. In this case, the monocular optical sensor can effectively identify such non-obstacle targets and reject such targets. The monocular optical sensor cannot effectively identify normal targets in the face of insufficient light, rain, snow and fog, and its ranging error is too large to be used as a single source of braking control.

The Beidou high-precision integrated navigation communication terminal integrates 4G/5G communication module, Beidou high-precision positioning module, and inertial measurement unit. 4G/5G is used to receive and check the standard brake control parameters issued by the cloud. Beidou high-precision positioning + inertial measurement unit, completes high-precision integrated navigation, realizes high-precision position, speed, acceleration measurement, and provides data support for evaluating each braking. Specifically: Receive the obstacle and road lane line information of the 77GHz millimeter-wave radar sensor and the monocular optical sensor, integrate the ranging and speed measurement features and advantages of the two sensors, and output the vehicle and the most dangerous obstacle for the brake controller on the one hand. The distance and relative speed of the object, and the position of the road lane line, etc.; on the other hand, the position, speed, and acceleration information of the vehicle are extracted, and output to the brake controller for evaluating the strength of braking.

The brake motor controller, on the one hand, receives the distance and relative speed information between the vehicle and the most dangerous obstacle provided by the Beidou high-precision integrated navigation communication terminal, and represents the vehicle at any time according to the distance collision time TTC=relative vehicle distance/relative speed The time it takes to collide with the target obstacle will send an early warning signal to the HMI display terminal. According to the TTC information and vehicle speed, the precise stepping brake motor will be controlled for emergency braking according to the pre-designed strength; on the other hand, according to the current vehicle speed, combined with The braking ability of the vehicle is to generate a suitable braking signal corresponding to the point braking/full braking. The speed v and acceleration a information provided by the Beidou high-precision integrated navigation communication terminal are used to estimate the braking signal this time, and the actual braking signal is estimated. The deviation range between the dynamic acceleration and the configured acceleration parameter is recorded, and the steady-state change of the vehicle's braking can be analyzed according to the long-term data accumulation, and the local AEBS braking parameter acceleration and response time can be inverted and corrected according to the actual situation. Due to the influence of load, weather changes and the running time of the brake mechanism, the corrected control parameters are sent to the cloud through the Beidou high-precision integrated navigation communication terminal for backup, and the cloud is analyzed and confirmed.

The cloud is used to evaluate the rationality of AEBS braking parameters based on a large amount of accumulated braking intensity information, mainly for the decrease in braking force caused by vehicle aging and wear. The braking operation is carried out using the issued new braking parameters.

In the vehicle braking model of this embodiment, the vehicle braking model uses the basic law of physics v ² =2as. Under the condition that the acceleration a of the vehicle is constant, the greater the speed v, the greater the braking distance s required, and the square times In order to determine the vehicle braking time and the early warning time, of course, the brake reaction time T1 and the time T2 required for the brake to achieve the maximum braking are also considered.

Under the control signal of the brake controller, the precise stepping brake motor realizes the accurate motor lever displacement and realizes the accurate and controllable braking force output of the brake pedal.

The HMI display terminal can be an Android smart interface or an HMI display, which can realize sound and light warning under the control of the brake controller. At the same time, the distance and type of obstacles in the lane line are displayed.

The present invention avoids analyzing the internal and external factors affecting the braking capability of the vehicle, such as changes in vehicle load, changes in road conditions, weather changes, and aging of vehicle conditions. With the help of AEBS's precise brake stepper motor and high-precision Beidou satellite navigation integrated navigation terminal, AEBS can choose the right time in each transportation task, conduct different levels of brake tests, and collect the generated braking capacity. Learn and compare the braking capacity parameter with the braking force parameter pre-stored by AEBS or issued by the cloud, and calculate the most mandatory power possessed by the vehicle in this environment. Then AEBS automatically calculates the appropriate safe braking distance and early warning distance, and informs the driver of early warning and emergency braking.

As shown in Figure 2, in this embodiment, the main component block diagram of the Beidou high-precision integrated navigation communication terminal of the vehicle is the core measurement component for AEBS to realize self-learning. Through the Beidou + MEMS combination algorithm running on the processor, Realize the capture and measurement of instantaneous speed and acceleration information. The positioning and speed measurement accuracy of Beidou satellite navigation remains stable without divergence. However, MEMS are small in size and low in cost, but there is a random drift phenomenon, which will diverge over time, so that MEMS cannot be used as a measuring device alone. However, the integrated navigation algorithm combines the advantages of MEMS and Beidou navigation systems, and uses MEMS to measure vehicle position, speed, and attitude values with poor accuracy, and combines the position and speed calculated by Beidou receiver. The measurement error of MEMS is estimated and compensated by data fusion, and then high-precision vehicle position, speed and acceleration information are obtained.

As shown in Figure 3, it is a closed-loop process of self-learning. When the optical and radar sensors are suitable for sensing, the brake controller performs precise braking control on the precise stepping brake motor according to the pre-designed strength, considering that the brake pedal is stepped on. After the braking force takes effect, the vehicle speed and acceleration are measured through the Beidou high-precision integrated navigation communication terminal, and fed back to the brake controller, and compared with the preset braking parameters, through the preset parameter table, select The safest and most comfortable braking parameters are controlled by AEB.

To sum up, the above are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims (2)

1. A self-adaptive active anti-collision brake system with autonomous training and learning is characterized by mainly comprising 6 parts, namely a millimeter wave radar sensor, a monocular optical sensor, a brake motor controller, a precise stepping brake motor, a Beidou high-precision combined navigation communication terminal and a display terminal;

the millimeter wave radar sensor is used for detecting forward obstacle information and transmitting the forward obstacle information to the Beidou high-precision integrated navigation communication terminal;

the monocular optical sensor is used for detecting forward obstacle information and road and lane line information and transmitting the forward obstacle information and the road and lane line information to the Beidou high-precision integrated navigation communication terminal;

the Beidou high-precision combined navigation communication terminal receives the information of the obstacles and the lane lines of the road and fuses the measurement information of two sensors of the Beidou high-precision combined navigation communication terminal, so that on one hand, the distance and the relative speed between the vehicle and the most dangerous obstacle and the lane line position of the road are output for the brake controller; on the other hand, the position, the speed and the acceleration information of the vehicle are extracted and output to the brake controller for evaluating the braking strength;

the brake motor controller receives the distance between the vehicle and the most dangerous obstacle and the relative speed information, sends out an early warning signal to a display terminal according to the distance collision time TTC, and controls the accurate stepping brake motor to brake emergently according to the strength designed in advance according to the TTC information and the vehicle speed; on the other hand, according to the current vehicle speed condition, combining the vehicle braking capacity, generating a corresponding braking signal suitable for inching braking/full braking, estimating the current braking signal by using the speed and acceleration information provided by the Beidou high-precision integrated navigation communication terminal, estimating the deviation range of the actually generated braking acceleration and the configured acceleration parameter, recording, analyzing the steady state change of the vehicle brake by using the accumulated deviation range, carrying out inversion on the local AEBS braking parameter, correcting the influence caused by load, weather change and the running time of a braking mechanism according to the actual condition, and sending the corrected control parameter to the cloud end through the Beidou high-precision integrated navigation communication terminal; receiving the confirmed braking parameters sent by the cloud, and performing braking control by using the braking parameters;

the accurate stepping brake motor realizes the brake force output of the brake pedal under the control of the brake controller;

and the HMI display terminal realizes acousto-optic early warning under the control of the brake controller.

2. The adaptive active anti-collision braking system with autonomous training and learning of claim 1, wherein the HMI display terminal further displays the distance and type of obstacles in the lane line of the vehicle.

Images