CN112141101A - Method and system for pre-aiming safety path based on CNN and LSTM - Google Patents
Method and system for pre-aiming safety path based on CNN and LSTM Download PDFInfo
- Publication number
- CN112141101A CN112141101A CN202011057237.5A CN202011057237A CN112141101A CN 112141101 A CN112141101 A CN 112141101A CN 202011057237 A CN202011057237 A CN 202011057237A CN 112141101 A CN112141101 A CN 112141101A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- steering wheel
- path
- wheel angle
- driver
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 230000033001 locomotion Effects 0.000 claims abstract description 16
- 230000001133 acceleration Effects 0.000 claims description 36
- 238000012937 correction Methods 0.000 claims description 17
- 238000006073 displacement reaction Methods 0.000 claims description 17
- 238000004364 calculation method Methods 0.000 claims description 13
- 238000004458 analytical method Methods 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 8
- 230000004044 response Effects 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 7
- 239000000126 substance Substances 0.000 claims description 5
- 238000012546 transfer Methods 0.000 claims description 4
- 230000002457 bidirectional effect Effects 0.000 claims description 2
- 230000008685 targeting Effects 0.000 claims 7
- 238000013277 forecasting method Methods 0.000 claims 1
- 238000013527 convolutional neural network Methods 0.000 abstract description 8
- 238000013528 artificial neural network Methods 0.000 abstract description 7
- 239000013598 vector Substances 0.000 description 10
- 238000010586 diagram Methods 0.000 description 8
- 238000006243 chemical reaction Methods 0.000 description 7
- 238000005457 optimization Methods 0.000 description 7
- 230000009466 transformation Effects 0.000 description 6
- 230000008859 change Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 230000009471 action Effects 0.000 description 4
- 210000002569 neuron Anatomy 0.000 description 4
- 230000008904 neural response Effects 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000013213 extrapolation Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000015654 memory Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000005562 fading Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000012916 structural analysis Methods 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units
- B60W30/10—Path keeping
- B60W30/12—Lane keeping
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W10/00—Conjoint control of vehicle sub-units of different type or different function
- B60W10/20—Conjoint control of vehicle sub-units of different type or different function including control of steering systems
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
- B60W40/10—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to vehicle motion
- B60W40/105—Speed
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W2050/0001—Details of the control system
- B60W2050/0043—Signal treatments, identification of variables or parameters, parameter estimation or state estimation
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W2520/00—Input parameters relating to overall vehicle dynamics
- B60W2520/10—Longitudinal speed
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W2710/00—Output or target parameters relating to a particular sub-units
- B60W2710/20—Steering systems
Landscapes
- Engineering & Computer Science (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Human Computer Interaction (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Steering Control In Accordance With Driving Conditions (AREA)
- Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
Abstract
The invention relates to a method and a system for pre-aiming a safe path based on CNN and LSTM, wherein the method at least comprises the following steps: predicting an ideal steering wheel angle of the initial driver following model based on the expected vehicle speed information and the expected path information; optimizing the ideal steering wheel corner into an actual steering wheel corner based on vehicle motion information fed back by a vehicle model; and the vehicle model optimizes vehicle attitude information based on the actual steering wheel angle information and fits to obtain expected path information. The invention decides the steering wheel angle applied to the automobile by the driver according to the preset road track and the expected speed information, thereby deciding the driving path of the driver. And continuously adjusting the internal weight of the neural network through the LSTM, thereby reducing the deviation of the expected path and the actual path and obtaining a good driver preview following model followed by the preview path.
Description
Technical Field
The invention relates to the technical field of intelligent driving, in particular to a method and a system for previewing a safe path based on CNN and LSTM. The invention also relates to the technical field of automobiles, and relates to an automobile for previewing a safe path.
Background
The path tracking is one of the cores of the automatic driving bottom layer control, and the safety and the stability of the vehicle are determined by the performance of the path tracking. In the field of path tracking research, the more common methods are: PID, driver Predictive Model, Model Predictive Control (MPC), and the like. The model predictive control can be used for different working conditions, but has higher requirements on a vehicle model and larger calculated amount.
The driver preview model was developed by the hoodie courier, giga, and was operated based on steering angle inputs. The preview path refers to a forward path according to the current head direction and the heading angle direction of the vehicle. The driver preview model is simple in structure, small in calculation amount and capable of coping with more working conditions, and therefore the driver preview model is widely adopted in the field of path tracking. The existing driver preview model adopts an ideal working condition as a calculation basis during preview path tracking calculation, does not consider the error of preview path tracking calculation caused by the objective system delay, and brings potential safety hazard for automatically driving vehicles.
Patent document CN108944930B discloses an automatic car following method and system based on LSTM for simulating the characteristics of drivers, wherein the automatic car following method includes step 1, which establishes a training sample library according to driving training data collected at different places and under different weather conditions: the driving training data includes input feature data and corresponding driving maneuver data: the input characteristic data comprises the speed of the vehicle, the speed of the vehicle ahead, the acceleration of the vehicle and the distance between the vehicles, and the driving operation data comprises sensor data of an accelerator pedal and a brake pedal; step 2, establishing an LSTM network model: the LSTM network model comprises l input layers, 2 hidden layers and 1 output layer; wherein: an input layer: defining an input vector XL as a 4-dimensional vector, wherein 4 elements respectively represent a vehicle distance, a vehicle speed of a vehicle ahead and an acceleration of the vehicle; all input vectors are spread out in time series: wherein the input vector XL refers to an input at a time t; a first hidden layer: set as a fully connected layer, comprising 32 neurons: the first hidden layer is a dropout layer and is used for preventing overfitting of the model and automatically discarding a certain proportion of neurons in the fully-connected layer; a second hidden layer: set as a fully connected layer, comprising 41 neurons: the second hidden layer is a dropout layer and is used for preventing the model from being over-fitted and automatically discarding a certain proportion of neurons in the fully-connected layer; an output layer: set as softmax classifier: the output vector OL is obtained by jointly calculating the current time t and the previous partial memory and corresponds to the discrete longitudinal vehicle operation vector: the vehicle operation information in the output vector OL is a 41-dimensional vector, which respectively corresponds to the operation states of an accelerator pedal and a brake pedal of the target vehicle, wherein the 1 st to 20 th elements represent accelerator operations and respectively correspond to different accelerator pedal opening degrees: the 21 st element represents no operation; elements 22-41 represent braking operations, corresponding to different brake pedal opening degrees respectively; and 3, performing off-line training on the LSTM network model according to the training samples. The invention utilizes the characteristic that an LSTM recurrent neural network is good at processing time sequence characteristic data to simulate the behavior of a driver following a vehicle, but the invention can not decide the steering wheel angle applied to the vehicle by the driver according to the preset road track and the expected vehicle speed information and can not decide the driving path of the driver.
Patent document CN111284478A discloses a preview path tracking calculation method and a tracking calculation module, the method mainly includes the following steps, S1, obtaining the delay Δ T of the system of the automatic driving vehicle; s2, acquiring a global path, and extrapolating and calculating the position of the vehicle after the system delay delta T in the global path based on a uniform acceleration model; s3, calculating a preview error; and S4, calculating according to the relationship between the preview error and the tire rotation angle to obtain a tire rotation angle, calculating according to the tire rotation angle to obtain a steering wheel rotation angle, and sending the steering wheel rotation angle to a chassis CAN to execute steering wheel rotation angle control. According to the method, the system delay is introduced into the calculation of the preview path by accurately acquiring the system delay which objectively exists, the calculation error of the preview path can be reduced, the vehicle position extrapolation prediction is carried out by using the uniform acceleration model, and the relatively accurate vehicle extrapolation prediction position can be obtained through a small calculation amount. The invention does not solve the technical problem of reducing the deviation of the intended path from the actual path.
Patent document CN107264534B discloses an intelligent driving control system based on a driver experience model, comprising: the system comprises a driver model module, a pre-aiming control model module, a path tracking control module and a whole vehicle system module, wherein the driver model module is used for collecting driver operation information and expected path information and predicting a driving state parameter of a vehicle at a future moment by combining vehicle attitude information; the pre-aiming control model module is used for calculating a transverse correction angle and judging a correct driving mode of the vehicle at the next moment; the path tracking control module is used for converting parameters respectively output by the driver model module and the pre-aiming control model module into control quantities executable by the whole vehicle system module; and the whole vehicle system module is used for responding to the control quantity obtained from the tracking control module to control the vehicle to move and feeding back vehicle movement parameters to the driver model module, the pre-aiming control model module and the path tracking control module.
Furthermore, on the one hand, due to the differences in understanding to the person skilled in the art; on the other hand, since the inventor has studied a lot of documents and patents when making the present invention, but the space is not limited to the details and contents listed in the above, however, the present invention is by no means free of the features of the prior art, but the present invention has been provided with all the features of the prior art, and the applicant reserves the right to increase the related prior art in the background.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a method for pre-aiming a safe path based on CNN and LSTM, which is characterized by at least comprising the following steps:
predicting an ideal steering wheel angle of the initial driver following model based on the expected vehicle speed information and the expected path information;
optimizing the ideal steering wheel corner into an actual steering wheel corner based on vehicle attitude information fed back by a vehicle model;
and the vehicle model optimizes vehicle attitude information based on the actual steering wheel angle information and fits to obtain expected path information.
Based on the defect that the error of the driver model preview path in the prior art is larger than the actual path, the method corrects the actual steering wheel corner by calculating the error of the steering wheel corner, and is favorable for obtaining a more accurate driving path of the driver.
Preferably, the method further comprises: adjusting weight information inside the LSTM in a manner that actual path information, the expected path information is input, thereby optimizing the initial driver following model. According to the method, the internal weight of the neural network is continuously adjusted through the LSTM, so that the deviation between the expected path and the actual path is reduced, and the pre-aiming path following driver pre-aiming following model with smaller deviation is obtained.
Preferably, the method for optimizing the ideal steering wheel angle to the actual steering wheel angle includes:
and analyzing the correction quantity of the steering wheel angle based on the lateral acceleration error in the vehicle posture information fed back by the vehicle model, and calculating the actual steering wheel angle according to the correction quantity of the steering wheel angle and the ideal steering wheel angle. The correction of the steering wheel angle is beneficial to compensating G caused by the reflection lag of the driver and various reasonsayThe error caused by the discrepancy of the gain of the lateral acceleration of the actual vehicle enables the driver to have strong robustness in the preview following model.
Preferably, the method of optimizing vehicle attitude information further comprises:
predicting attitude information of the vehicle based on a two-degree-of-freedom vehicle model, wherein,
carrying out automobile dynamics analysis and dynamics analysis on the two-freedom-degree automobile model to obtain a motion differential equation of the two-freedom-degree automobile;
by centroid slip angle beta and yaw angular velocity wrAs state variables, steering wheel angle as input quantityCalculating a corresponding state differential equation according to the motion differential equation of the two-degree-of-freedom automobile;
thereby connecting the centroid slip angle beta and the yaw rate wrAnd lateral acceleration ayAs an output of the state space, an output equation is obtained as:
wherein k is1Represents the cornering stiffness, k, of the front wheel2The cornering stiffness of the rear wheels is represented, m represents the mass of the whole vehicle, u represents the component of the absolute speed of the vehicle in the x-axis direction of a vehicle coordinate system, a represents the distance from the center of mass to the front axis, b represents the distance from the center of mass to the front axis, and beta represents the steering wheel angle.
The centroid slip angle beta and the yaw angular velocity w output by the inventionrAnd lateral acceleration ayAnd the operation stability of the driver following model can be analyzed quickly and conveniently.
Preferably, the method of optimizing the initial driver-following model comprises at least:
correcting the weight in the LSTM according to the preset step number by taking the actual path information as a training set and the expected path information as a test set,
where step number representation LSTM predicts the output of a subsequent step number based on the previous step number input. By adding the LSTM, the driving path decided by the method has good following performance no matter where the curvature change is large or small.
Preferably, the calculation method for predicting the ideal steering wheel angle of the initial driver following model is as follows:
wherein the content of the first and second substances,
representing ideal lateral directionAcceleration, GayShowing the steady-state gain of lateral acceleration to steering wheel angle, feRepresenting the lateral displacement of the vehicle, y representing the lateral displacement of the vehicle at time t, vyRepresents the lateral speed of the vehicle at time T, and T represents the preview time.
Preferably, the calculation method for analyzing the correction amount of the steering wheel angle includes:
wherein, ayRepresents lateral acceleration, 1/(1+ t)hs) represents the transfer function, thRepresents a motion response lag time constant, H represents an acceleration feedback coefficient, and s represents the number of steps.
The invention relates to a CNN and LSTM-based system for previewing a safe path, which is characterized by at least comprising a vehicle speed judgment module, a path judgment module, a driver model module and a vehicle model module, wherein the vehicle speed prediction module and the path judgment module are respectively in data connection with the driver model module, and the driver model module and the vehicle model module are in unidirectional or bidirectional data connection; wherein the content of the first and second substances,
the driver model module calculates an ideal steering wheel corner according to the expected vehicle speed sent by the vehicle speed judging module and the expected path information sent by the path judging module; and is
The driver model module optimizes the ideal steering wheel corner into an actual steering wheel corner according to the vehicle attitude information fed back by the vehicle model module;
and the vehicle model module optimizes vehicle attitude information based on the actual steering wheel corner information sent by the driver model module and fits to obtain expected path information.
The system for pre-aiming the safe path continuously adjusts the internal weight of the neural network through the LSTM, so that the deviation between the expected path and the actual path is reduced, and a good pre-aiming path following driver model is obtained. The invention can effectively serve the intelligent driving system, and improve the driving safety of the driver according to the corresponding decided path.
Preferably, the method for the driver model module to optimize the ideal steering wheel angle to the actual steering wheel angle comprises:
analyzing the correction quantity of the steering wheel angle based on the lateral acceleration error in the vehicle posture information fed back by the vehicle model module,
and calculating the actual steering wheel angle according to the correction quantity of the steering wheel angle and the ideal steering wheel angle.
The invention provides an intelligent vehicle which is characterized in that a preview safety system is installed on the vehicle, and the preview safety system decides a driving path of a driver according to the preview safety path method.
Drawings
FIG. 1 is a schematic diagram of the structure of the logic module of the present invention;
FIG. 2 is a schematic view of the driver preview following model framework of the present invention;
FIG. 3 is a direction control driver preview following model of the present invention under a constant vehicle speed low curvature road;
FIG. 4 is a schematic diagram of a driver preview follow model in one embodiment;
FIG. 5 is a schematic view of angles and coordinates of a two-degree-of-freedom automobile model;
FIG. 6 is a schematic diagram of a path in the case of a double-shift road;
FIG. 7 is a schematic diagram of path fitting under a double-shift road condition;
FIG. 8 is a block diagram of the LSTM neural network of the present invention
FIG. 9 is a LSTM-based driver preview following model framework diagram;
FIG. 10 is a schematic of a driver's predicted path following model decision without using LSTM structural optimization;
FIG. 11 shows a schematic of a path followed by a model decision for driver preview using LSTM structural optimization.
List of reference numerals
10: a vehicle speed prediction module; 20: a path prediction module; 30: a driver model module; 40: a vehicle model module; 31: steering wheel angle information; 41: vehicle attitude information; a: a preview point.
Detailed Description
The following detailed description is made with reference to the accompanying drawings.
In the prior art, the deviation between the expected path and the actual path is large, so that the driving safety of a driver needs to be improved. The invention decides the steering wheel angle applied to the automobile by the driver according to the preset road track and the expected speed information, thereby deciding the driving path of the driver. And continuously adjusting the internal weight of the neural network through the LSTM, thereby reducing the deviation of the expected path and the actual path and obtaining a good pre-aiming path following driver model.
The invention relates to a method and a system for previewing a safe path based on CNN and LSTM, which can also be an optimization method of a driver model previewing path of a Long Short-Term Memory network (LSTM), and can also be an establishment method and an establishment system of a driver following model. The invention also can be an intelligent automobile, a CNN and LSTM-based system for previewing the safe path is arranged on the intelligent automobile, or the driver model is optimized according to the optimization method of the driver model previewing path.
The steering wheel angle information 31 in the present invention includes an ideal steering wheel angle, a steering wheel angle correction amount, and an actual steering wheel angle.
The vehicle posture information 41 of the present invention includes at least vehicle posture information such as position information and speed information of the vehicle, and may include other information related to the vehicle posture.
As shown in fig. 1, the system for predicting a safe path of the present invention includes at least a vehicle speed determination module 10, a path determination module 20, a driver model module 30, and a vehicle model module 40. The vehicle speed prediction module 10 and the path judgment module 20 are respectively in data connection with the driver model module 30. The driver model module 30 and the vehicle model module 40 establish a one-way or two-way data connection.
The vehicle speed determination module 10 is used to demand and describe the expected vehicle speed from the vehicle speed on the road. The path determination module 20 is used to describe the expected road trajectory when the driver follows a clear road centerline.
The driver model module 30 is used to build a driver model based on the expected vehicle speed and the expected road estimates, and to optimize the driver model. Specifically, the driver model module 30 calculates an ideal steering wheel angle according to the expected vehicle speed sent by the vehicle speed judgment module and the expected path information sent by the path judgment module; and the driver model module optimizes the ideal steering wheel angle to be the actual steering wheel angle according to the vehicle attitude information fed back by the vehicle model module.
And the vehicle model module optimizes vehicle attitude information based on the actual steering wheel corner information sent by the driver model module and fits to obtain expected path information. The vehicle model module 40 transmits the vehicle position information and the speed information to the driver model module 30, and controls the vehicle according to the steering wheel angle transmitted by the driver model module 30.
The method for pre-aiming the safe path based on the CNN and the LSTM comprises the following steps:
an ideal steering wheel angle is calculated based on the expected vehicle speed and the expected path.
And the initial driver follows the lateral acceleration error information fed back by the model vehicle model to optimize an ideal steering wheel corner and form an actual steering wheel corner.
And inputting actual steering wheel angle information into the two-degree-of-freedom automobile model to predict an actual driving path.
And analyzing the stability of the vehicle on the basis of the track of the double-lane road on the basis of the transformation of the coordinate system, wherein the double-lane road is obtained by performing curve fitting on the sudden change track.
Actual travel path information is input based on the LSTM structure to optimize the driver's preview following model.
The method of calculating the ideal steering wheel angle based on the expected vehicle speed and the expected path is as follows.
Fig. 2 is a frame diagram of a driver following model. Drawing (A)In 2, P(s) is a preview link, F(s) is a forward correction link, B(s) is a feedback estimation link, G(s) is the dynamic characteristic of the automobile, f is an expected track characteristic quantity, y is the position of the current automobile motion track, f(s) is a forward correction link, B(s) is a feedback estimation link, G(s) is the dynamic characteristic of the automobile, f is an expected track characteristic quantity, y is a position of thepFor the prediction of the characteristic variables of the vehicle position at a future time, which are estimated by the preview stage on the basis of the current vehicle motion state, ypThe estimation link estimates the estimated value of the characteristic quantity at the future moment according to the current state information, and the deviation of the estimated values of the two characteristic quantities is the control information applied to the automobile.
Assuming that the vehicle is traveling at a constant speed, the desired path is followed and the lateral displacement of the vehicle is small relative to the longitudinal displacement. Under this assumption, the driver's tracking of the road trajectory is the tracking of the desired lateral displacement of the vehicle. Let f (t) be the expected lateral displacement of the vehicle and y (t) be the current actual lateral displacement of the vehicle. According to the theory of 'preview-follow' driver modeling, a driver can acquire information in front of a road through 'preview'. Assuming that the preview time of the driver is T, the operation of the current steering wheel by the driver is expected to be that the actual lateral displacement y (T + T) of the vehicle is close to the expected lateral displacement f of the automobile as much as possible after the preview time TeWherein f ise＝f(t+T)。
As shown in FIG. 3, the lateral displacement of the vehicle at time t is y and the lateral velocity is vyThe lateral deviation between the preview point A with the preview time T and the vehicle is delta fp. From the assumed conditions,. DELTA.fp≈fe-y. Assuming that the vehicle is at the time of the day with an ideal lateral acceleration
Making a uniform acceleration movement in the lateral direction, the vehicle can be brought to the desired trajectory after time T, then
Since y (T + T) ═ f (T + T), then
Lateral acceleration
At the speed, the steady-state gain of the lateral acceleration of the automobile to the steering wheel angle is GayTo achieve ideal lateral acceleration
Ideal steering wheel angle that should be applied:
the method of optimizing the ideal steering wheel angle based on the feedback information of the lateral acceleration is as follows.
In the actual driving situation of the driver, the driver has physiological limitation, the using state of the vehicle is very complex, and the control of the vehicle by only adopting the steering wheel angle which is ideally decided is unrealistic.
The physiological limitation of the driver mainly comes from the reaction lag of the driver, and can be divided into two types, namely the nerve reaction lag and the action reaction lag of the driver. The neural response lag of the driver describes the perception of various information by the driver, and the lag is usually a pure lag and is expressed by a transfer function exp (-t)ds) represents tdIs the neural response lag time. The action reaction lag of the driver describes the operation process of the driver on the automobile, the lag can be generally described by a first-order inertia link, and the transfer function is 1/(1+ t)hs) where t ishIs the motion response lag time constant. As shown in FIG. 4, the steering wheel angle after setting the physiological limit of the driver issw0. The lateral dynamic characteristics of the vehicle are nonlinear due to the reaction lag of the driver, and the complex driving conditions (such as uneven road surface) of the automobile are usedsw0The lateral acceleration a of the vehicle actually generated during the directional controlyAnd ideal acceleration
The difference causes the tracking precision of the driver preview following model to be reduced. Therefore, the invention adopts a lateral acceleration error feedback mode to correct the steering wheel rotation angle decided by the driving and driver preview following model module, and the correction quantity is
H represents an acceleration feedback coefficient.
The steering wheel angle is corrected to compensate for the steady gain G caused by the delay in the driver's response and various factorsayThe error caused by the discrepancy of the gain of the lateral acceleration of the actual vehicle enables the driver to have strong robustness in the preview following model. For example, when a vehicle is driven on a low-adhesion road surface and the lateral acceleration gain is reduced due to the restriction of the road surface friction factor, the steady-state gain G obtained on a high-adhesion road surface is usedayAnd the determined steering wheel angle is smaller, and the actual lateral acceleration of the automobile is smaller. An additional steering wheel angle is added after lateral acceleration feedback is used to compensate for the steady state gain GayErrors due to mismatch. Steering wheel corner finally decided by driver preview following modelswIs composed ofsw0And correction amount deltaswAnd (4) summing. And inputting the actual lateral displacement and the expected lateral displacement after coordinate transformation into a display for observation and comparison.
Preferably, the invention adopts a two-degree-of-freedom automobile model as a vehicle model in a driver-automobile closed loop system. By turning of steering wheelswThe vehicle model is used as an input to obtain the corresponding automobile motion situation, including the vehicle attitude information. According to the assumption of a two-degree-of-freedom automobile model, the automobile is simplified into a two-wheel automobile model with only two degrees of freedom of lateral direction and transverse swing, and as shown in FIG. 5, the cornering stiffness of the left wheel and the right wheel of the same axle is the same; and the front (or rear) wheel has a constant lateral deviation degree which is the sum of the lateral deviation rigidity of the tires on both sides.
In fig. 5, the origin of the vehicle coordinate system coincides with the center of mass of the automobile, and the physical meaning of each parameter is as follows:
indicating a front wheel turning angle; beta represents the centroid sideDeflection angle; xi represents a heading angle; u. of1,u2Respectively representing the speed of the middle points of the front axle and the rear axle; alpha is alpha1,α2Respectively representing front and rear shaft side deflection angles; fY1,FY2Representing the lateral reaction forces of the ground to the front and rear wheels, respectively, i.e. the cornering forces; v. of1Representing the absolute speed of the vehicle; u represents the component of the absolute speed of the automobile in the x-axis direction of the vehicle coordinate system; v represents the component of the absolute speed of the automobile in the y-axis direction of the vehicle coordinate system; w is arRepresenting the yaw rate of the automobile; a represents the distance of the centroid to the front axis; b represents the distance of the center of mass to the rear axis; l represents the wheel base.
The method for analyzing the corresponding vehicle motion situation by the vehicle model of the vehicle model module 40 is as follows.
And (3) carrying out automobile kinematic analysis on the two-degree-of-freedom automobile model to obtain the components of the automobile mass center absolute acceleration in the directions of the x axis and the y axis of the automobile coordinate system:
performing dynamic analysis on the vehicle motion model to obtain an external force sigma FYExternal moment sigma MzThe relationship with the vehicle parameters is:
wherein k is1，k2The cornering stiffness of the front and rear wheels.
The differential equation of motion of the two-degree-of-freedom automobile obtained by the formulas (1) and (2) is as follows:
selecting a centroid slip angle beta and a yaw angular velocity wrIs a state variable, i.e. x ═ β wr]T. The steering wheel angle is used as an input value, i.e., u. The corresponding state differential equation can be listed according to the above formula
The corresponding matrix form is:
wherein
Connecting the centroid slip angle beta and the yaw angular velocity wrAnd lateral acceleration ayAs the output of the state space, the attitude information of the vehicle is obtained, which is convenient for analyzing the operation stability of the driver preview following model in the follow-up process, namely, y ═ beta ωr ay]T. The output equation y ═ Cx + Du is given by equation (5):
wherein the content of the first and second substances,
and m is the mass of the whole vehicle.
Preferably, the present invention obtains an actual traveling trajectory of the automobile by performing coordinate transformation on a vehicle coordinate system.
Because the model adopted by the driver-automobile closed-loop system is a two-degree-of-freedom automobile model, the lateral acceleration output by the two-degree-of-freedom automobile model is based on the vehicle coordinate system. This also corresponds to the driver's preview following model's imitation of the driver's driving behavior, since the driver is sitting in the car and he is looking ahead. The direction referred to when information on the road ahead coincides with the direction in which the vehicle is traveling. In order to obtain the actual driving trajectory of the vehicle in the subsequent simulation analysis, it is necessary to perform coordinate transformation on the vehicle coordinate system. Transformation of the vehicle coordinate system to the geodetic coordinate system may employ the following transformation equations.
Wherein (x (t), y (t)) are coordinates of the automobile in a geodetic coordinate system at a certain moment, (x (t), y (t)) are coordinates in a vehicle coordinate system, and ψ (t) is a heading angle of the automobile.
Preferably, the invention evaluates the emergency obstacle avoidance capability of the vehicle based on the double-moving-line path.
The double-moving-line path is a typical test working condition for testing the operation stability of the vehicle, and the emergency obstacle avoidance capability of the vehicle is evaluated by operating the vehicle to enable the vehicle to generate higher lateral acceleration alternately.
The dimensions of the double lane shift test path are shown in FIG. 6. The actual road dimensions are performed according to the ISO 3888-1:1999 standard. In the figure s1＝15m，s2＝30m，s3＝25m，s4＝25m，s5＝15m，s6＝15m，B1Vehicle width of 1.1 times plus 0.25m, B2Vehicle width of 1.2 times plus 0.25m, B31.3 vehicle width times plus 0.25m, where B width equals 3.5 m.
The central line of the double-lane-shifting road is taken as an ideal road track to be tracked, and fig. 6 shows that the double-lane-shifting road has sudden changes at a plurality of corner points, and the sudden changes are not beneficial to the stability of the tracked vehicle. And (5) carrying out cubic spline curve fitting on the sudden change position, so that the first derivative of the fitted road at the sudden change position is continuous. The fitted curve is shown in fig. 7, and its expression is:
in fig. 7, the abscissa x is the actual longitudinal displacement of the vehicle, and the ordinate y is the actual lateral displacement of the vehicle. In FIG. 7, a0＝s1；a1＝s1+s2；a2＝s1+s2+s3；a3＝s1+s2+s3+s4。
d＝a1-a0；d′＝a2-a3。
e′1＝-6a3a2B/d′3；e′2＝3(a3+a2)B/d′3；e′3＝-2B/d′3。
In a vehicle test or simulation, the abscissa represents time in a normal case, and the abscissa can be converted into time by the relationship x ═ ut between the vehicle longitudinal displacement and the vehicle speed, so that f (x) of equation (6) can be converted into a function f (t) of time t:
in the formula (7), g0＝e0；gi＝eiuj(j＝1,2,3)；h0＝e′0；hi＝e′iuj(j＝1,2,3)。
The specific implementation parameters of the present invention are as follows.
One of the preferred vehicle parameters is:
the mass m of the whole vehicle is 1765 kg; the axle distance L is 2.6 m; the front wheelbase a is 1.2 m;
the rear wheelbase b is 1.4 m; yaw moment of inertia I of whole vehiclez＝2700kg·m2(ii) a Front and rear axle yaw stiffness k1,k2160000N/rad; the vehicle speed was 22 m/s.
Preferably, one of the parameters of the driver preview following model is as follows:
the hysteresis of the driver's action response and the hysteresis of the neuro response vary from driver to driver and are influenced by factors such as sex, age, personality and health. Generally, the action reaction lags behind th0.05-0.20 s, and the neural response lags behind tdTaking the mixture for 0.2-0.6 s. Preferably t ish＝0.10s,td0.4 s. According to the work of a plurality of researchers, the preview time T of the driver is generally 0.8-1.5 s, and preferably T is 0.8 s.
Preferably, the relationship between the vehicle speed and the lateral acceleration of the two-degree-of-freedom vehicle model in the linear interval to the gain Gay of the steering wheel angle is as follows:
wherein, IsThe steering system angle transmission ratio of the steering wheel angle to the front wheel angle; k is a stability factor and is an important parameter for representing the steady-state response of the automobile, the general vehicle has moderate understeer, and the value range of K is generally 0.002-0.004 s2/m2Herein K-0.0032637 s2/m2。
Preferably, the method optimizes the decision path of the driver aiming following the model based on the LSTM principle.
LSTM has found widespread use because it reduces the problem of recurrent neural network fading and explosive gradients, and can better address the long term dependence problem.
The LSTM structure shown in FIG. 8 acquires the input information C at the previous timen-1And hn-1Then, the input signal and the input gate inForgetting door fnAnd an output gate onSuch asThe output information at the next time can be obtained by the following relational expression.
in＝σ(Wixxn+Wihhn-1+Wiccn-1+bi)，
fn＝σ(Wfxxn+Wfhhn-1+Wfccn-1+bf)，
on＝σ(Woxxn+Wohhn-1+Woccn-1+bo)，
hn＝on⊙tanh(cn)，
Wherein, W*(ix, ih, ic, fx, fh, fc, ox, oh, oc, cx, ch) represents a weight matrix, b*(═ i, f, o, c) denotes a bias vector, which denotes the scalar product of two vectors, tanh (·) denotes a hyperbolic tangent function:
the use of these gates allows the LSTM to decide whether to retain or weigh new information over existing memory, which makes the LSTM effective for use in decision path optimization for driver preview following models.
The driver preview following model of the invention considers that when the automobile runs on the double-moving-line road, the automobile does not stop suddenly and does not move instantly. This means that the trajectory of the vehicle is continuous, as is the steering wheel angle of the driver. Thus, the driver model can be built by a sequential data modeling approach. Meanwhile, on an expressway, two main factors affecting the driver's steering operation are vehicle speed and road curvature.
Fig. 9 shows a frame diagram of the LSTM-based driver preview follow model. As shown in FIG. 9, input data P of the input layer of LSTM1The method comprises the following steps: historical velocity vn-s+1 vn-s+2 … vn]Historic road curvature [ rho ]n-s+1 ρn-s+2 … ρn]Future road curvature [ rho ]n+1 ρn+2 … ρn+s]And historical turning angle of steering wheeln-s+1 n-s+2 … n]Output data P of the output layer2Is the turning angle of the future steering wheeln+1 n+2 … n+s]. The step number S indicates that LSTM predicts the number of steps to be output later from the previous step number input.
Preferably, the invention adopts the step number of 5, namely, the actual driving path data is used as a training set in Simulink, the expected track data is used as a test set, and the weight in the LSTM is corrected by every 5 steps of data, so that the driver can follow the model more accurately.
FIG. 10 shows the driver's preview following the model decision path without using LSTM structural optimization. FIG. 11 shows the path followed by the model decision for driver preview using LSTM structural analysis.
Comparing fig. 10 and 11, without using LSTM structural optimization, the driver preview following model of the present invention has better following performance when expecting small changes in path curvature. But with greater fluctuations at faster changes in the turn (the turn). After the LSTM structure analysis is utilized, the method has good following performance no matter where the curvature change is large or small. It can therefore be concluded that the LSTM can effectively optimize the decision path that the driver aims at following the model.
It should be noted that the above-mentioned embodiments are exemplary, and that those skilled in the art, having benefit of the present disclosure, may devise various arrangements that are within the scope of the present disclosure and that fall within the scope of the invention. It should be understood by those skilled in the art that the present specification and figures are illustrative only and are not limiting upon the claims. The scope of the invention is defined by the claims and their equivalents.
Claims (10)
1. A method for pre-targeting a secure path based on CNN and LSTM, the method comprising at least:
predicting an ideal steering wheel angle of the initial driver following model based on the expected vehicle speed information and the expected path information;
optimizing the ideal steering wheel corner into an actual steering wheel corner based on vehicle attitude information fed back by a vehicle model;
and the vehicle model optimizes vehicle attitude information based on the actual steering wheel angle information and fits to obtain expected path information.
2. The method of pre-targeting a secure path based on CNN and LSTM as claimed in claim 1, wherein the method further comprises: adjusting weight information inside the LSTM in a manner that actual path information, the expected path information is input, thereby optimizing the initial driver following model.
3. The CNN and LSTM based pre-targeting safety path method of claim 2, wherein the method of optimizing the ideal steering wheel angle to an actual steering wheel angle comprises:
analyzing the correction amount of the steering wheel angle based on the lateral acceleration error in the vehicle posture information fed back by the vehicle model,
and calculating the actual steering wheel angle according to the correction quantity of the steering wheel angle and the ideal steering wheel angle.
4. The CNN and LSTM based method of pre-targeting a safe path of claim 3, wherein said method of optimizing vehicle pose information further comprises:
predicting attitude information of the vehicle based on a two-degree-of-freedom vehicle model, wherein,
carrying out automobile dynamics analysis and dynamics analysis on the two-freedom-degree automobile model to obtain a motion differential equation of the two-freedom-degree automobile;
by centroid slip angle beta and yaw angular velocity wrThe state variable is a state variable, the steering wheel angle is used as an input quantity, and a corresponding state differential equation is calculated according to the motion differential equation of the two-degree-of-freedom automobile;
thereby connecting the centroid slip angle beta and the yaw rate wrAnd lateral acceleration ayAs an output of the state space, an output equation is obtained as:
wherein k is1Represents the cornering stiffness, k, of the front wheel2The cornering stiffness of the rear wheels is represented, m represents the mass of the whole vehicle, u represents the component of the absolute speed of the vehicle in the x-axis direction of a vehicle coordinate system, a represents the distance from the center of mass to the front axis, and b represents the distance from the center of mass to the front axis, and represents the steering wheel angle.
5. The method of pre-targeting a safe path based on CNN and LSTM according to claim 4, wherein the method of optimizing the initial driver following model comprises at least:
correcting the weight in the LSTM according to the preset step number by taking the actual path information as a training set and the expected path information as a test set,
where step number representation LSTM predicts the output of a subsequent step number based on the previous step number input.
6. The CNN and LSTM based safe path forecasting method according to any of claims 1-5, characterized in that the calculation method for predicting the ideal steering wheel angle of the initial driver following model is:
wherein the content of the first and second substances,
representing the ideal lateral acceleration, GayShowing the steady-state gain of lateral acceleration to steering wheel angle, feRepresenting the lateral displacement of the vehicle, y representing the lateral displacement of the vehicle at time t, vyRepresents the lateral speed of the vehicle at time T, and T represents the preview time.
7. The method of pre-targeting a safe path based on CNN and LSTM according to any one of claims 1 to 5, wherein the calculation method of analyzing the correction amount of the steering wheel angle is:
wherein, ayRepresents lateral acceleration, 1/(1+ t)hs) represents the transfer function, thRepresents a motion response lag time constant, H represents an acceleration feedback coefficient, and s represents the number of steps.
8. The system for previewing the safe path based on the CNN and the LSTM is characterized by at least comprising a vehicle speed judgment module (10), a path judgment module (20), a driver model module (30) and a vehicle model module (40), wherein the vehicle speed prediction module (10) and the path judgment module (20) are respectively in data connection with the driver model module (30), and the driver model module (30) and the vehicle model module (40) are in unidirectional or bidirectional data connection; wherein the content of the first and second substances,
the driver model module (30) calculates an ideal steering wheel angle according to the expected vehicle speed sent by the vehicle speed judging module (10) and the expected path information sent by the path judging module (20); and is
The driver model module (30) optimizes the ideal steering wheel angle to an actual steering wheel angle according to the vehicle posture information fed back by the vehicle model module (40);
the vehicle model module (40) optimizes vehicle attitude information and fits to obtain expected path information based on the actual steering wheel angle information sent by the driver model module (30).
9. The CNN and LSTM based pre-targeting safety path system of claim 8, wherein the method of the driver model module (30) optimizing the ideal steering wheel angle to an actual steering wheel angle comprises:
analyzing the correction amount of the steering wheel angle based on the lateral acceleration error in the vehicle posture information fed back by the vehicle model module (40),
and calculating the actual steering wheel angle according to the correction quantity of the steering wheel angle and the ideal steering wheel angle.
10. An intelligent vehicle, characterized in that a preview safety system is installed on the vehicle, and the preview safety system decides the driving path of a driver according to the preview safety path method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011057237.5A CN112141101B (en) | 2020-09-29 | 2020-09-29 | Method and system for pre-aiming safety path based on CNN and LSTM |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011057237.5A CN112141101B (en) | 2020-09-29 | 2020-09-29 | Method and system for pre-aiming safety path based on CNN and LSTM |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112141101A true CN112141101A (en) | 2020-12-29 |
| CN112141101B CN112141101B (en) | 2022-10-04 |
Family
ID=73894387
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011057237.5A Active CN112141101B (en) | 2020-09-29 | 2020-09-29 | Method and system for pre-aiming safety path based on CNN and LSTM |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112141101B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113211438A (en) * | 2021-05-08 | 2021-08-06 | 东方红卫星移动通信有限公司 | Wheel type robot control method and system based on pre-aiming distance self-adaption |
| CN113359764A (en) * | 2021-07-02 | 2021-09-07 | 合肥工业大学 | GRU-based driver model optimization method |
| CN113734198A (en) * | 2021-09-03 | 2021-12-03 | 智己汽车科技有限公司 | Target relative course obtaining method and device |
| CN114114929A (en) * | 2022-01-21 | 2022-03-01 | 北京航空航天大学 | Unmanned vehicle path tracking method based on LSSVM |
| CN114148320A (en) * | 2021-12-08 | 2022-03-08 | 华人运通(上海)自动驾驶科技有限公司 | Path tracking control method and device |
| CN114228715A (en) * | 2021-11-30 | 2022-03-25 | 武汉理工大学 | Vehicle queue combined control method, device, equipment and storage medium |
| CN114896694A (en) * | 2022-05-17 | 2022-08-12 | 哈尔滨工业大学（威海） | Path tracking control method based on two-point preview |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107380161A (en) * | 2017-07-18 | 2017-11-24 | 吉林大学 | A kind of active steering control device for aiding in driver to realize desired ride track |
| CN110626340A (en) * | 2019-09-30 | 2019-12-31 | 南京航空航天大学 | Intelligent automobile path tracking control system and method based on wolf algorithm |
| CN110641465A (en) * | 2019-10-25 | 2020-01-03 | 长安大学 | Lane keeping system and method based on vehicle speed |
| US20200234066A1 (en) * | 2019-01-18 | 2020-07-23 | Toyota Research Institute, Inc. | Attention-based recurrent convolutional network for vehicle taillight recognition |
| CN111552322A (en) * | 2020-04-29 | 2020-08-18 | 东南大学 | Unmanned aerial vehicle tracking method based on LSTM-particle filter coupling model |
| CN111610780A (en) * | 2019-02-25 | 2020-09-01 | 广州汽车集团股份有限公司 | Automatic driving vehicle path tracking control method and device |
-
2020
- 2020-09-29 CN CN202011057237.5A patent/CN112141101B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107380161A (en) * | 2017-07-18 | 2017-11-24 | 吉林大学 | A kind of active steering control device for aiding in driver to realize desired ride track |
| US20200234066A1 (en) * | 2019-01-18 | 2020-07-23 | Toyota Research Institute, Inc. | Attention-based recurrent convolutional network for vehicle taillight recognition |
| CN111610780A (en) * | 2019-02-25 | 2020-09-01 | 广州汽车集团股份有限公司 | Automatic driving vehicle path tracking control method and device |
| CN110626340A (en) * | 2019-09-30 | 2019-12-31 | 南京航空航天大学 | Intelligent automobile path tracking control system and method based on wolf algorithm |
| CN110641465A (en) * | 2019-10-25 | 2020-01-03 | 长安大学 | Lane keeping system and method based on vehicle speed |
| CN111552322A (en) * | 2020-04-29 | 2020-08-18 | 东南大学 | Unmanned aerial vehicle tracking method based on LSTM-particle filter coupling model |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113211438A (en) * | 2021-05-08 | 2021-08-06 | 东方红卫星移动通信有限公司 | Wheel type robot control method and system based on pre-aiming distance self-adaption |
| CN113211438B (en) * | 2021-05-08 | 2023-06-16 | 东方红卫星移动通信有限公司 | Wheel type robot control method and system based on pretightening distance self-adaption |
| CN113359764A (en) * | 2021-07-02 | 2021-09-07 | 合肥工业大学 | GRU-based driver model optimization method |
| CN113734198A (en) * | 2021-09-03 | 2021-12-03 | 智己汽车科技有限公司 | Target relative course obtaining method and device |
| CN114228715A (en) * | 2021-11-30 | 2022-03-25 | 武汉理工大学 | Vehicle queue combined control method, device, equipment and storage medium |
| CN114228715B (en) * | 2021-11-30 | 2023-06-13 | 武汉理工大学 | Vehicle queue joint control method, device, equipment and storage medium |
| CN114148320A (en) * | 2021-12-08 | 2022-03-08 | 华人运通(上海)自动驾驶科技有限公司 | Path tracking control method and device |
| CN114148320B (en) * | 2021-12-08 | 2023-10-20 | 华人运通(上海)自动驾驶科技有限公司 | Path tracking control method and equipment |
| CN114114929A (en) * | 2022-01-21 | 2022-03-01 | 北京航空航天大学 | Unmanned vehicle path tracking method based on LSSVM |
| CN114896694A (en) * | 2022-05-17 | 2022-08-12 | 哈尔滨工业大学（威海） | Path tracking control method based on two-point preview |
| CN114896694B (en) * | 2022-05-17 | 2024-06-04 | 哈尔滨工业大学（威海） | Path tracking control method based on two-point pre-aiming |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112141101B (en) | 2022-10-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112141101B (en) | Method and system for pre-aiming safety path based on CNN and LSTM | |
| CN111890951B (en) | Intelligent electric automobile trajectory tracking and motion control method | |
| CN112622903B (en) | Longitudinal and transverse control method for autonomous vehicle in vehicle following driving environment | |
| JP4143104B2 (en) | Vehicle control device | |
| Kosecka et al. | A comparative study of vision-based lateral control strategies for autonomous highway driving | |
| CN110568760A (en) | Parameterized learning decision control system and method suitable for lane changing and lane keeping | |
| CN107085424A (en) | Pre- for automatic Pilot takes aim at crosswise joint | |
| CN114379583B (en) | Automatic driving vehicle track tracking system and method based on neural network dynamics model | |
| CN112829766B (en) | Adaptive path tracking method based on distributed driving electric vehicle | |
| CN113009829B (en) | Longitudinal and transverse coupling control method for intelligent internet motorcade | |
| CN110217227A (en) | A kind of braking in a turn joint collision avoidance control method suitable for ice-snow road operating condition | |
| Németh et al. | Nonlinear analysis and control of a variable-geometry suspension system | |
| CN111559389A (en) | Control method of intelligent automobile under variable adhesion coefficient repeatability track | |
| Dekkata et al. | Improved steering and adaptive cruise control for autonomous vehicles using model predictive control | |
| CN112644488A (en) | Adaptive cruise system | |
| CN113359764B (en) | GRU-based driver model optimization method | |
| CN114987537A (en) | Neural network dynamics-based road adaptive drift control system and method for automatic driving vehicle | |
| CN115432009A (en) | Automatic driving vehicle trajectory tracking control system | |
| CN113341994B (en) | Intelligent automobile path tracking control method based on optimal control of piecewise affine system | |
| CN113184040B (en) | Unmanned vehicle line-controlled steering control method and system based on steering intention of driver | |
| Kanchwala et al. | Model predictive tracking controller for a high fidelity vehicle dynamics model | |
| Thommyppillai et al. | Advances in the development of a virtual car driver | |
| CN116834754A (en) | Transverse and longitudinal cooperative control method for self-adaptive speed regulation of automatic driving vehicle | |
| CN114323698B (en) | Real vehicle experiment platform testing method for man-machine co-driving intelligent vehicle | |
| CN115817509A (en) | Multi-axis distributed driving vehicle steering auxiliary track tracking method based on AMPC |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |