CN116639151B - Unmanned vehicle control method and system based on pedestrian existence prediction in pavement blind area - Google Patents
Unmanned vehicle control method and system based on pedestrian existence prediction in pavement blind area Download PDFInfo
- Publication number
- CN116639151B CN116639151B CN202310619973.2A CN202310619973A CN116639151B CN 116639151 B CN116639151 B CN 116639151B CN 202310619973 A CN202310619973 A CN 202310619973A CN 116639151 B CN116639151 B CN 116639151B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- pedestrian
- sidewalk
- probability
- speed
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000013210 evaluation model Methods 0.000 claims abstract description 29
- 238000011217 control strategy Methods 0.000 claims abstract description 27
- 230000003449 preventive effect Effects 0.000 claims abstract description 22
- 230000000007 visual effect Effects 0.000 claims abstract description 18
- 230000009471 action Effects 0.000 claims abstract description 12
- 230000006399 behavior Effects 0.000 claims description 73
- 230000007613 environmental effect Effects 0.000 claims description 17
- 238000004891 communication Methods 0.000 claims description 10
- 230000000903 blocking effect Effects 0.000 claims description 8
- 230000001133 acceleration Effects 0.000 claims description 6
- 241000283070 Equus zebra Species 0.000 claims description 5
- 238000001556 precipitation Methods 0.000 claims description 4
- 238000003066 decision tree Methods 0.000 claims description 3
- 206010039203 Road traffic accident Diseases 0.000 abstract description 5
- 239000000523 sample Substances 0.000 abstract description 5
- 230000033001 locomotion Effects 0.000 description 8
- 238000005259 measurement Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 206010021567 Impulsive behaviour Diseases 0.000 description 2
- 238000012097 association analysis method Methods 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000008030 elimination Effects 0.000 description 2
- 238000003379 elimination reaction Methods 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 238000012098 association analyses Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000008451 emotion Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000003137 locomotive effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000004080 punching Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
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
- B60W60/00—Drive control systems specially adapted for autonomous road vehicles
- B60W60/001—Planning or execution of driving tasks
-
- 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, or advanced driver assistance systems for ensuring comfort, stability and safety or drive control systems for propelling or retarding the vehicle
- B60W30/08—Active safety systems predicting or avoiding probable or impending collision or attempting to minimise its consequences
- B60W30/095—Predicting travel path or likelihood of collision
- B60W30/0956—Predicting travel path or likelihood of collision the prediction being responsive to traffic or environmental parameters
-
- 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
- B60W40/02—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 ambient conditions
-
- 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
- B60W2554/00—Input parameters relating to objects
- B60W2554/40—Dynamic objects, e.g. animals, windblown objects
- B60W2554/402—Type
- B60W2554/4029—Pedestrians
-
- 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
- B60W2556/00—Input parameters relating to data
- B60W2556/10—Historical data
Abstract
The invention provides an unmanned vehicle control method and system based on pedestrian existence prediction in a pavement blind area, wherein the method comprises the following steps: acquiring historical environment data associated with a sidewalk, establishing a pedestrian impact walking sidewalk model, and predicting the probability of pedestrian impact walking sidewalk; historical vehicle behavior data when a pedestrian is encountered is obtained, a vehicle control behavior evaluation model is built, and the probability of the pedestrian in front of the side vehicle is predicted; calculating the existence probability of pedestrians in the blind area of the pavement visual field; judging whether a preventive control strategy needs to be adopted or not based on the existence probability of pedestrians in the blind area of the pavement view. According to the invention, the pedestrian impact walking pavement model and the vehicle control behavior evaluation model are coupled, the existence probability of pedestrians in a pavement visual field blind area is predicted, and preventive control is adopted on a scene with high probability of occurrence of ghost probe accidents, so that control actions can be adopted in advance to prevent the pedestrians before the vehicles recognize the pedestrians, and the possibility of traffic accidents is reduced.
Description
Technical Field
The invention belongs to the technical field of unmanned, and particularly relates to an unmanned vehicle control method and system based on pedestrian existence prediction of a pavement blind area.
Background
The development of unmanned automobile technology is rapid, the safety problem of the unmanned automobile is becoming an important topic, and particularly the possibility of accidents in certain fixed scenes is greatly increased, for example, when the view of vehicles is blocked at intersections with pedestrians, passers-by or non-motor vehicles crossing the road in front of the vehicles on the left side and the right side are difficult to find in time, and then measures for avoiding collision cannot be taken.
However, the recently developed technologies such as vehicle-road interconnection require too long development time and require a large amount of engineering for installing the infrastructure, which is difficult to achieve. Therefore, the traffic accident of the dead zone ghost probe can be effectively avoided by the sensor and the control strategy of the vehicle, and the traffic accident is a problem which needs to be solved by the present day.
The development of unmanned technology enables a perception system of an unmanned automobile to be identified by multi-sensor fusion, but the problems still exist for a scene with obstacle shielding. For example, the invention patent with publication number CN109949612a discloses a ghost probe accident early warning system based on information interaction, which carries out pedestrian or motor vehicle transverse passing detection through an ultrasonic radar and reminds a vehicle owner and a pedestrian through voice alarm. However, it can only warn vehicles in the blind area of pedestrian vision, but for vehicles, in the stage of passing through the crossroad by unmanned vehicles, the vision is blocked by left or right vehicles, the vehicle sensor cannot sense the information of pedestrians and vehicles in front of left or right, even if the sensor such as ultrasonic radar performs auxiliary detection, the blind area of vision caused by the blocking of surrounding vehicles can occur, and whether the situation that pedestrian impact crosses the road occurs in the blind area of the crossroad cannot be judged, so that the automatic driving vehicle cannot perform preventive control action. In addition, after the potential danger is detected, the driver actually drives the vehicle does not respond to the potential danger, and the driver can not timely take braking or the braking time is too late, so that the accident risk is increased.
Disclosure of Invention
In view of the above, the invention provides an unmanned vehicle control method and system based on pedestrian existence prediction in a pavement blind area, which are used for solving the problem that unmanned vehicles are difficult to perform preventive control when a pavement has a visual field blind area.
The invention provides a unmanned vehicle control method based on pedestrian existence prediction in a pavement blind area, which comprises the following steps:
acquiring historical environment data associated with a sidewalk, establishing a pedestrian-rushing-through sidewalk model, and predicting the probability of pedestrian rushing-through sidewalk by the pedestrian-rushing-through sidewalk model;
historical vehicle behavior data when a pedestrian is encountered is obtained, a vehicle control behavior evaluation model is established, and the probability that the pedestrian appears in front of the side vehicle is predicted through the vehicle control behavior evaluation model;
the probability of pedestrian impulse passing through the sidewalk and the probability of pedestrian occurrence in front of the side vehicles are weighted and summed, and the existence probability of the pedestrian in the vision blind area of the sidewalk is calculated;
judging whether a preventive control strategy needs to be adopted or not based on the existence probability of pedestrians in the blind area of the pavement view.
On the basis of the above technical solution, preferably, the obtaining environmental data associated with the pavement and establishing a pedestrian impact walk-through pavement model, and predicting the probability of the pedestrian impact walk-through pavement by the pedestrian impact walk-through pavement model specifically includes:
Acquiring historical environment data comprising precipitation, visibility, temperature T, crowd age distribution, zebra crossing length, whether pedestrian and vehicle traffic lights exist, red light remaining time, green light remaining time and time period, and the probability of pedestrian impulse passing through a sidewalk corresponding to the historical environment data;
taking a feature vector formed by the historical environment data as input and taking the probability of pedestrian impulse passing through the sidewalk corresponding to the historical environment data as output to train a Bayesian model, so as to obtain a pedestrian impulse passing through sidewalk model;
and acquiring real-time environment data associated with the current sidewalk to form a feature vector, inputting a pedestrian-rushing-through sidewalk model, and outputting the probability of pedestrian rushing through the sidewalk.
On the basis of the above technical solution, preferably, the obtaining the historical vehicle behavior data when the pedestrian is encountered and establishing a vehicle control behavior evaluation model, and predicting the probability that the pedestrian appears in front of the side vehicle is encountered by the vehicle control behavior evaluation model specifically includes:
acquiring historical vehicle behavior data including an original speed, a distance between the vehicle and a sidewalk, a vehicle size, a vehicle acceleration, a vehicle braking distance, a vehicle transverse deflection angle, a vehicle whistle state and frequency, and probability of an opportune pedestrian corresponding to the historical vehicle behavior data;
Taking a feature vector formed by historical vehicle behavior data as input and taking probability of a pedestrian in chance as output to train a decision tree model to obtain a vehicle control behavior evaluation model;
and acquiring real-time behavior data of a side vehicle in the same direction as the own vehicle to form a feature vector, inputting a vehicle control behavior evaluation model, and outputting the probability of the occurrence of pedestrians in front of the side vehicle.
On the basis of the above technical solution, preferably, the judging whether to adopt a preventive control strategy based on the size of the existence probability of the pedestrian in the blind area of the pavement field specifically includes:
if the pedestrian existence probability c of the blind area of the pavement is less than 60% and the vehicle speed VE of any side vehicle with shielding action on own vehicles is less than the preset vehicle speed V0, directly performing control for eliminating the blind area of the pavement without adopting a preventive control strategy;
if the speed VE of a single vehicle with shielding behavior on own vehicles is greater than or equal to the preset speed V0, a preventive control strategy is needed to be adopted to carry out follow-up running control.
On the basis of the above technical solution, preferably, the control for eliminating the blind area of the field of view specifically includes:
if only one side is blocked, determining the speed VE of the side blocking vehicle E; when 0< c <0.3, the speed VQ=VE+6 of the own vehicle, when 0.3 is less than or equal to c <0.6, the speed VQ=VE+3 of the own vehicle is controlled to drive to the head of the own vehicle to be level with the head of the side shielding vehicle, and all information on two sides of a road is identified through a laser radar arranged on the own vehicle; performing vehicle control according to the identified road information;
If both sides are blocked, determining the speeds VL and VH of the vehicles blocked at both sides, wherein VL is smaller than VH; when 0< c <0.3, the speed VQ=VL+6 of the own vehicle, and when 0.3 is less than or equal to c <0.6, the speed VQ=VL+3 of the own vehicle, and the own vehicle is controlled to run to be level with the head of the vehicle with the speed VL; identifying road information on the side where the vehicle with the speed VL is located by a laser radar arranged on the own vehicle; when a pedestrian is walking on the sidewalk, if the transverse distance between the pedestrian and the own vehicle is greater than a preset distance d, accelerating the pedestrian to pass through, and if the transverse distance is less than d, enabling vq=0; if no pedestrian is present, the own vehicle speed vq=vh+3 is controlled to drive the own vehicle until the own vehicle head is flush with the vehicle head having the speed VH; all information on two sides of a road is identified through a laser radar arranged on the own vehicle; and controlling the vehicle according to the identified road information.
On the basis of the above technical solution, preferably, the following travel control includes:
if only one side is blocked, determining the speed VE of the side blocking vehicle E and the central point YE of the length of the vehicle body, and controlling the own vehicle to run until the head of the own vehicle is level with the central point YE of the length of the vehicle body and always follow the side blocking vehicle E to run;
If both sides are blocked, determining the speeds VL and VH of the vehicles blocked at both sides, wherein VL is smaller than VH;
if VL < V0, when 0< c <0.3, the own vehicle speed vq=vl+6, and when 0.3 is less than or equal to c <0.6, the own vehicle speed vq=vl+3, and the own vehicle is controlled to run until the own vehicle head is flush with the vehicle head with the speed VL; identifying road information on the side where the vehicle with the speed VL is located by a laser radar arranged on the own vehicle; when a pedestrian is walking on the sidewalk, if the transverse distance between the pedestrian and the own vehicle is greater than a preset distance d, accelerating the pedestrian to pass through, and if the transverse distance is less than d, enabling vq=0; when no pedestrian is present, accelerating the own vehicle beyond the vehicle with the speed VL to the head of the own vehicle to be level with the central point of the vehicle with the speed VH, and braking the own vehicle if c > 0.8;
if VL>V0, adjusting own running speed, andcontrolling the own vehicle to run until the head of the own vehicle is flush with the vehicle body center point of the vehicle with the vehicle speed VL, and letting vq=vl, moving the own vehicle to follow the vehicle with the vehicle speed VL, and simultaneously, for real-time behavior data of the vehicle with the vehicle speed VH, if the condition { c is satisfied>0.8 |R EL }∪{ c>0.8|R EH -1, braking the own vehicle, otherwise not braking; r is R EL Real-time behavior data R for a vehicle with speed VL EH Is real-time behavior data of a vehicle having a vehicle speed VH.
On the basis of the above technical solution, preferably, the preset vehicle speed v0=15 km/h and the preset distance d=5 m.
In a second aspect of the present invention, there is provided a unmanned vehicle control system based on prediction of existence of pedestrians in a blind area of a pavement, the system comprising:
a first probability prediction module: the method comprises the steps of obtaining historical environment data related to a sidewalk, establishing a pedestrian-rushing-through sidewalk model, and predicting the probability of the pedestrian rushing-through sidewalk through the pedestrian-rushing-through sidewalk model;
a second probability prediction module: the method comprises the steps of acquiring historical vehicle behavior data when a pedestrian is encountered, establishing a vehicle control behavior evaluation model, and predicting the probability of the pedestrian in front of the side vehicle through the vehicle control behavior evaluation model;
pedestrian presence prediction module: the method comprises the steps of carrying out weighted summation on the probability of pedestrian impulse passing through a pavement and the probability of pedestrian occurrence in front of a lateral vehicle, and calculating the existence probability of pedestrians in a pavement vision blind area;
preventive control module: judging whether a preventive control strategy needs to be adopted or not based on the existence probability of pedestrians in the blind area of the pavement view.
In a third aspect of the present invention, an electronic device is disclosed, comprising: at least one processor, at least one memory, a communication interface, and a bus;
the processor, the memory and the communication interface complete communication with each other through the bus;
the memory stores program instructions executable by the processor which the processor invokes to implement the method according to the first aspect of the invention.
In a fourth aspect of the invention, a computer-readable storage medium is disclosed, storing computer instructions that cause a computer to implement the method according to the first aspect of the invention.
Compared with the prior art, the invention has the following beneficial effects:
1) The method simulates a driving scene model of pedestrians possibly occurring in the stage of the unmanned vehicle passing through the sidewalk, establishes a pedestrian-rushing-through sidewalk model through historical environment data related to the sidewalk, and predicts the probability of the pedestrian rushing-through sidewalk; establishing a vehicle control behavior evaluation model according to historical vehicle behavior data when a pedestrian is encountered, and predicting the probability of the pedestrian in front of the side vehicle; finally, the prediction results of the two models are coupled to obtain the existence probability of pedestrians in the visual field blind area of the sidewalk, and preventive control is adopted on the scene of the high-probability ghost probe accident, so that control actions can be adopted in advance to prevent the pedestrians before the vehicles recognize the pedestrians, and the possibility of traffic accidents is reduced;
2) According to the invention, surrounding environment information is considered, the pedestrian existence probability of a visual field blind area of a pavement is combined, two control schemes of visual field blind area elimination control and following driving control are provided for solving the problem that a visual field blind area exists in a pavement stage of vehicle passing, and if the pedestrian existence probability c of the visual field blind area of the pavement is less than 60% and the vehicle speed VE of a vehicle on any side with shielding action on a host vehicle is less than a preset vehicle speed V0, the visual field blind area elimination control is carried out; if the speed VE of a single vehicle with shielding action on own vehicles is greater than or equal to the preset speed V0, the following driving control is carried out, so that the possibility of misjudgment can be greatly reduced, and the feasibility of a control strategy is improved.
3) The invention provides two control schemes of eliminating the blind area control of the visual field and the following driving control, which considers the situation that only one side is blocked and both sides are blocked, combines the speed and the relative position measured by different sides to carry out fine control, improves the accuracy of the control strategy, can adapt to various traffic scenes of vehicles in the stage of crossing the sidewalk, and improves the adaptability of the control strategy.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of an unmanned vehicle control method based on pedestrian existence prediction in a pavement blind area;
FIG. 2 is a diagram of the existence of occlusion phenomena;
FIG. 3 is a schematic diagram of the actual front measurement range in the presence of occlusion;
fig. 4 is a schematic diagram of the measurement range in the rear and front of the control for eliminating the blind area of the visual field.
Detailed Description
The following description of the embodiments of the present invention will clearly and fully describe the technical aspects of the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, are intended to fall within the scope of the present invention.
For the stage of passing the pavement by the unmanned vehicle, as the vision is blocked by the left or right vehicle, the vehicle sensor can not sense the information of the pedestrians and vehicles in front of the left or right, and if the pedestrians or vehicles which are crossing the pavement suddenly appear in front of the side, the pedestrians and vehicles can not be braked in time, or the braking time is too late, so that traffic accidents can be generated. The traditional unmanned traffic sidewalk control strategy is easy to misjudge and has low traffic efficiency. In order to prevent the pedestrian by taking control action in advance before the vehicle identifies the pedestrian and reduce the possibility of safety accidents, the invention fully considers the psychological trend of the pedestrian in the surrounding environment and predicts the existence of the pedestrian in the blind area of the pavement with the observed vehicle motion track, thereby guiding the unmanned vehicle control strategy.
The invention is applicable to the following scenes: the own vehicle enters a pedestrian passing intersection, and any one of the left side and the right side is a straight-going lane, and the situation that vehicles exist on the same-direction lane and shielding behaviors exist is recognized, and no vehicle exists in the front 10 m. Wherein, definition of occlusion phenomenon is: due to the shielding of the left (right) vehicle, the front recognition range of the own vehicle laser radar is obviously reduced, for example, the front right radar radiation range is 90 degrees originally, and the shielding radar radiation range of the vehicle is smaller than 45 degrees. Fig. 2 is a diagram showing a shielding phenomenon, wherein the left side diagram is an overhead view, and the right side diagram is an in-vehicle view.
The required sensors of the invention are a set of solid-state laser radar sensors r1, r2, r3 and r4 which are respectively arranged at four turning angles of the vehicle, a GPS system, a network communication module and a sound sensor. Wherein, r1, r2, r3, r4 can identify circles with the radius more than 100m and the angle of 360 degrees, and the length information and the position information of the left and right vehicles can be calculated through point cloud identification. The vehicle speed and acceleration can be calculated according to the vehicle travel distance change and the time change and the speed change. And calculating the transverse steering offset angle according to the measured locomotive position information and the central point position information.
Referring to fig. 1, the invention provides a unmanned vehicle control method based on pedestrian existence prediction in a pavement blind area, which comprises the following steps:
s1, acquiring historical environment data associated with a sidewalk, establishing a pedestrian-rushing-through sidewalk model, and predicting the probability of pedestrian rushing through the sidewalk by using the pedestrian-rushing-through sidewalk model.
Definition of pedestrian impulsive walk-through behavior: if the distance between any vehicle and the zebra stripes is not less than 10m and the speed is not less than 20km/h, the pedestrian still passes through the sidewalk, and the event is regarded as the pedestrian impulse passing sidewalk behavior because the event is an impulse behavior in the psychological level.
According to the invention, the pedestrian impact walking sidewalk model is established based on the Bayesian model by acquiring the historical environment data associated with the sidewalk, so that the probability of pedestrian impact walking sidewalk is predicted.
S11, acquiring historical environment data related to the sidewalk and the probability of pedestrian impulse passing through the sidewalk corresponding to the historical environment data.
Firstly, environmental data needing to be observed in advance and the influence of the environmental data on the psychology of the pedestrians are selected, so that specific characteristic variables needed by the pedestrians to walk through the sidewalk model in a fluctuant mode are determined.
(1) Road conditions are intuitive factors affecting pedestrian impact traveling on a sidewalk, and thus as one of the environmental data to be observed, the road conditions are divided into three categories here including: the length of the zebra stripes, whether traffic lights exist or not and the remaining time of the traffic lights/green lights;
(2) weather factors also affect the emotion of pedestrians, and therefore, the weather factors should also be taken as one of environmental data affecting pedestrian mobility through sidewalks, and the influence caused by weather is divided into several categories, namely, main meteorological factors include: air temperature, precipitation, visibility, and environmental conditions under common conditions are taken as characteristic variables.
(3) The possibility of pedestrian rushing through the sidewalk needs to consider the psychology of pedestrians in traffic psychology, and on the basis of judging whether the pedestrians cross the road, the pedestrians and the drivers of the vehicles have psychological games, on the one hand, the motion states of the vehicles which are about to pass through the sidewalk comprise the speed of the vehicles, the distance between the vehicles and the sidewalk, the acceleration of the vehicles, the whistling state of the vehicles and the like, and on the other hand, the motions of the pedestrians comprise the sight line direction of the pedestrians, the advancing direction of the pedestrians, the hand motions of the pedestrians and the like. And the interaction of the two affects the psychological activities of the pedestrians. For such interactions, the different reactions that it may make can be classified by classifying pedestrians. One of the experienced adults has a more comprehensive understanding of the state of motion and traffic rules of the vehicle, so that the behavior of the vehicle must be more confident, and the probability of the vehicle making an emergency avoidance behavior is relatively low. If the vehicle is a child or an elderly person, the basic judgment on the road condition is lacking, and the proper vehicle motion state is difficult to grasp, so that the probability of making impulse behavior can be increased. Therefore, the invention divides the crowd age distribution as a category and uses the divided different crowd categories as a plurality of variables.
However, under the condition that a field of view blind area exists, the unmanned vehicle cannot recognize pedestrian information and cannot directly acquire crowd distribution, the crowd distribution is available under the condition of social analysis, the crowd distribution is closely related to the building type of the area, and the crowd of which type in one area can be estimated according to building type information.
The social crowd analysis model based on the building group deduces the age structure of the regional crowd through a clustering algorithm. The principle is that building distribution and crowd distribution of a certain number of areas are given, correlation between the building distribution and crowd distribution is learned and analyzed through a clustering algorithm, and then the most likely crowd age distribution structure of the areas can be deduced through the building distribution of the given areas.
The specific steps of deducing the crowd distribution of the sidewalk through a clustering algorithm are as follows:
1) The existing building distribution and crowd distribution data are subjected to hierarchical clustering, and the data are divided into hierarchical clusters according to the similarity of building group types and crowd types.
2) The data is divided into different subsets by selecting the appropriate hierarchy and number of clusters according to building class.
3) For each subset, a classification model is built to predict the group of people based on the group of buildings.
4) For data of a known building group category region, it is first judged which subset it belongs to, and then the corresponding classification model is used for predicting the crowd age distribution of the building group region.
The predicted regional population distribution is used as a part of the environmental data to be observed.
(4) Social factors can bring social pressure to pedestrians, impulsive behaviors are easy to make, the sources of the social pressure are different due to the fact that crowd distribution is not used, for example, the pressure of working crowds at the places comprises time limit of working, the pressure of student crowds has school time limit, and therefore besides the crowd distribution, the social factors also need to be used as environment data for influencing the impulsive walking of the pedestrians, and the time period is used as a dividing standard to be used as a main social influence factor.
Other environmental data which can influence the pedestrian impulsive behaviors can be also included in the observation range, and later factors with larger association degree with the pedestrian impulsive walk can be screened out through association analysis.
The invention collects original historical environment data associated with the sidewalk and the probability of pedestrian impulsivity passing through the sidewalk under the corresponding historical environment data, and screens out historical environment data with larger association degree with the pedestrian impulsivity passing through the sidewalk from the original historical environment data by an association analysis method based on a gray analysis theory. The association analysis method belongs to a common method in the field and is not described in detail.
The example of the historical environmental data associated with the sidewalk is shown in table 1, and includes precipitation, visibility, temperature T, crowd age distribution, zebra crossing length, whether there are historical environmental data of pedestrian and vehicle traffic lights, red light remaining time, green light remaining time, time period and the like, each type of historical environmental data corresponds to a plurality of characteristic variables, and the selected historical environmental data form a characteristic vector for training a pedestrian impulse walking sidewalk model.
Table 1 historical environmental data for pavement associations
And S12, taking a characteristic vector formed by the historical environment data as input, and taking the probability of pedestrian impulse passing through the sidewalk corresponding to the historical environment data as output to train a Bayesian model, so as to obtain the pedestrian impulse passing through sidewalk model.
And (3) forming a feature vector by the historical environment data screened in the step (S11), and calculating the posterior probability of the sample to be classified belonging to each category through a Bayesian model.
S13, acquiring real-time environment data associated with the current sidewalk to form a feature vector, inputting a pedestrian impact walking sidewalk model, and outputting the probability of pedestrian impact walking sidewalk.
S2, historical vehicle behavior data when a pedestrian is encountered is obtained, a vehicle control behavior evaluation model is built, and the probability that the pedestrian appears in front of the side vehicle is predicted through the vehicle control behavior evaluation model.
Since the probability of pedestrian impact traveling pavement model prediction is affected by the diversity of the environment and the recognition accuracy requirement of the sensor, there is a certain uncertainty, so after the environment is evaluated, whether the unmanned vehicle is to make a braking action or not, and the observation result of the vehicle (side vehicle) on a certain side running in the same direction as the own vehicle is needed to be combined for analysis. Because the track of the movement of the vehicle is regular, when the vehicle encounters an emergency situation, such as a sudden front emergence of a man, the driver reacts accordingly to avoid the collision as much as possible.
Therefore, the invention establishes a model for evaluating the control behavior of the vehicle when the pedestrian is encountered, and evaluates the possibility of the pedestrian in front of the vehicle by taking the motion trail of the side vehicle as the basis.
S21, historical vehicle behavior data of the vehicle when the pedestrian is encountered and probability of the pedestrian encountered corresponding to the historical vehicle behavior data are obtained.
The historical vehicle behavior data comprises an original speed, a distance between the vehicle and a sidewalk, a vehicle size, a vehicle acceleration, a vehicle braking distance, a vehicle transverse deflection angle, a vehicle whistle state, a frequency and the like. These data can be acquired by an IMU (inertial measurement unit) mounted on the vehicle and by stationary occasions.
S22, taking a feature vector formed by historical vehicle behavior data as input and taking probability of an opportunistic pedestrian as output to train a decision tree model, and obtaining a vehicle control behavior evaluation model.
S23, acquiring real-time behavior data of a side vehicle in the same direction as the own vehicle to form a feature vector, inputting a vehicle control behavior evaluation model, and outputting the probability that the side vehicle encounters a pedestrian in front.
The trained model is applied to an actual scene, a vehicle control behavior evaluation model is integrated on a vehicle, real-time behavior data are obtained through a laser radar, and similarly, the real-time behavior data of a side vehicle also comprise the original speed, the distance between the vehicle and a sidewalk, the vehicle size, the vehicle acceleration, the vehicle braking distance, the vehicle transverse deflection angle, the vehicle whistle state, the frequency and the like.
And S3, carrying out weighted summation on the probability of pedestrian impulse passing through the sidewalk and the probability of pedestrian occurrence in front of the side vehicles, and calculating the existence probability of the pedestrian in the blind area of the view of the sidewalk.
Assuming that the probability of pedestrian impact walking sidewalk output by the pedestrian impact walking sidewalk model is a, and the probability of pedestrian occurrence in front of a lateral vehicle output by the vehicle control behavior evaluation model is b, the pedestrian existence probability of a blind area of the sidewalk field of vision c=w 1 *a+w 2 *b，w 1 、w 2 For the weight, we can take 1 =0.3，w 2 =0.7。
S4, judging whether a preventive control strategy is needed to be adopted or not based on the existence probability of pedestrians in the blind area of the pavement view.
When the situation that the vehicle is about to drive into a road section containing a pavement is detected, the vehicle enters a pedestrian impulse pavement model according to the collected environmental information, and when the distance between the vehicle and the pavement is smaller than 10m, the vehicle control behavior evaluation model is operated, the pedestrian existence probability of a pavement vision blind area is calculated, and the vehicle pavement control is carried out according to the pedestrian existence probability of the pavement vision blind area.
In actually traveling a road with a pavement, different vehicle controls are required according to road conditions. For the time of eliminating the visual field blind area in time so as to avoid collision with pedestrians, a simpler control strategy can be adopted. And when the blind area of the visual field cannot be eliminated in time for complex road conditions, a safer control strategy is adopted to ensure the safety, and meanwhile, the traffic efficiency is ensured by combining the existence model of the blind area of the sidewalk. Therefore, the invention establishes two control schemes of eliminating the blind area control and the following driving control according to the existence probability of pedestrians in the blind area of the pavement and the environmental information.
S41, if the pedestrian existence probability c of the blind area of the pavement view is less than 60% and the speed VE of any side vehicle with shielding action on own vehicle is less than the preset speed V0, the control for eliminating the blind area of the pavement view is directly carried out without adopting a preventive control strategy.
According to multiparty tests, the braking distance of the unmanned automobile below 15km/h can be controlled within 0.5 m. Therefore, in the present embodiment, the preset vehicle speed v0=15 km/h is used as a safety boundary to ensure the driving safety.
The specific control strategy for eliminating the control of the visual field blind area is as follows:
s411, if only one side is blocked, determining the speed VE of the side blocking vehicle E; when 0< c <0.3, let own vehicle speed vq=ve+ km/h, when 0.3 is less than or equal to c <0.6, let own vehicle speed vq=ve+ km/h, control own vehicle to travel to own vehicle head and side shelter vehicle head parallel and level, at this moment can discern the whole information of road both sides through the lidar r1, r2 that are set up on own vehicle.
And carrying out vehicle control according to the identified road information: if no pedestrians are on the two sides of the road, enabling VQ=VE+V0 to accelerate to exceed other vehicles; for the situation that the pedestrian is traveling on the road, if the lateral distance between the pedestrian and the own vehicle is greater than the preset distance d, the pedestrian is accelerated to pass through, if the lateral distance is smaller than d, vq=0, and the preset distance d is a safe distance, and d=5m is preferable.
If there is a shade on both sides, determining the speeds VL and VH of the two-side-shading vehicles, wherein VL < VH, the vehicle on the lower speed side is denoted as EL, and the vehicle on the higher speed side is denoted as EH.
When 0< c <0.3, the speed vq=vl+ km/h of the own vehicle, and when 0.3 < c <0.6, the speed vq=vl+ km/h of the own vehicle, and the head of the own vehicle is controlled to be flush with the head of the vehicle EL with lower speed; the road information on the side where the vehicle EL with the lower vehicle speed is located can be identified by the laser radar r1 or r2 provided on the own vehicle; and carrying out vehicle control according to the identified road information: when a pedestrian is walking on the sidewalk, if the transverse distance between the pedestrian and the own vehicle is greater than a preset distance d, accelerating the pedestrian to pass through, and if the transverse distance is less than d, enabling vq=0;
if no pedestrian exists on the pavement, the speed VQ of the own vehicle is enabled to be=VH+3 km/h, the distance between the own vehicle and the vehicle EH with higher speed is gradually reduced, the own vehicle is controlled to run until the head of the own vehicle is level with the head of the vehicle EH with higher speed, and at the moment, all information on two sides of the pavement can be identified through the laser radars r1 and r2 arranged on the own vehicle;
and carrying out vehicle control according to the identified road information: when a pedestrian is on the left side and the right side and is on the road to pass through the sidewalk, if the transverse distance between the pedestrian and the own vehicle is greater than a preset distance d, accelerating to pass through, and if the transverse distance is less than d, enabling vq=0 km/h; if no pedestrians are on two sides of the road, vq=ve+v0 is accelerated to pass through the road beyond other vehicles.
FIG. 3 shows the actual front measurement range when the occlusion exists, and the actual front measurement range is the dark shade region with an included angle of θ1 in FIG. 3; fig. 4 shows the front measurement range after the control for eliminating the blind area of the field, and the front measurement range is the dark shade area with an included angle θ2 in fig. 4. Compared with the prior art, the control strategy for eliminating the blind area of the visual field can greatly improve the safety and the running efficiency of the unmanned vehicle on the road, and avoid accidents with other vehicles and pedestrians.
S42, if the speed VE of a single vehicle with shielding behavior on own vehicles is greater than or equal to a preset speed V0, a preventive control strategy is needed to be adopted to carry out follow-up running control.
When VE >15km/h, the strategy of eliminating the visual field blind area control is obviously dangerous, because the running speed is too high, and the situation that the braking distance is insufficient and the pedestrian collides with the situation that the own vehicle is too long when the pedestrian or the non-motor vehicle moves across the road is caused.
Based on this situation, the present invention proposes a following control strategy, i.e. during the course of the road, the own vehicle keeps a certain longitudinal distance from the side screening vehicle, vq=ve in the given case. The side shielding vehicle can clearly observe the road traffic condition of the shielded side of the own vehicle, if the condition that pedestrians or non-motor vehicles cross the road occurs, the side shielding vehicle must take braking action, and the own vehicle cannot recognize the road condition at the moment, but synchronously shields the behavior of the vehicle, and vq=0 km/h is caused when the output result c of the pedestrian existence model meets a certain condition, so that collision with pedestrians or non-motor vehicles crossing the road is prevented.
The following running control specific control strategy is:
s421, if only one side is blocked, determining the speed VE and the central point YE of the length of the vehicle E, at the moment, adjusting the running speed of the own vehicle, if the head of the own vehicle lags behind the central point YE of the length of the vehicle E, enabling VQ=VE+ km/h, otherwise enabling VQ=VE-3 km/h, controlling the own vehicle to run until the head of the own vehicle is level with the central point YE of the length of the vehicle, enabling VQ=VE, enabling the own vehicle to always follow the side blocked vehicle E to run, enabling a certain longitudinal distance to exist between the own vehicle and the side blocked vehicle E, and guaranteeing enough braking space when pedestrians appear at the current side.
If there is a shade on both sides, determining the speeds VL and VH of the two-side-shaded vehicles, where VL < VH, and similarly, the vehicle on the lower speed side is denoted as EL and the vehicle on the higher speed side is denoted as EH.
If VL < V0, when 0< c <0.3, the own vehicle speed vq=vl+ km/h, when 0.3 < c <0.6, the vehicle speed should not be too fast, the own vehicle speed vq=vl+ km/h, and the own vehicle is controlled to run until the own vehicle head is level with the vehicle head with VL; the road information on one side of the vehicle with lower running speed can be identified through the laser radar r1 or r2 arranged on the own vehicle; at this time, the next control action is performed according to the identified information: when a pedestrian is walking on the sidewalk, if the transverse distance between the pedestrian and the own vehicle is greater than a preset distance d, accelerating the pedestrian to pass through, and if the transverse distance is less than d, enabling vq=0; when no pedestrian is present, the own vehicle speed vq=vh+ km/h is set to accelerate the vehicle EL with a lower overrun speed. Then, the vehicle length center point YEH of the vehicle EH on the side with higher vehicle speed is determined, at this time, the own running speed is adjusted, if the own vehicle head lags the vehicle length center point YEH of the EH, vq=vh+ km/h is set, otherwise vq=vh-3 km/h is set, the own vehicle head is made to be level with the vehicle body center point of the vehicle EH with higher vehicle speed, vq=vh is set, that is, the own vehicle head is always kept aligned with YEH, and a certain longitudinal distance exists between the own vehicle and the EH, so that a sufficient braking space can be ensured when pedestrians appear on the side. At this time, if c >0.8, the own vehicle brakes.
If VL>V0, adjusting the own running speed, controlling the own vehicle to run to the position that the head of the own vehicle is level with the central point of the vehicle with the speed VL, enabling VQ=VL to enable the own vehicle to move along with the vehicle with the speed VL, ensuring a certain longitudinal distance between Q and EL, and providing enough braking space when pedestrians are in front of the side. Meanwhile, monitoring real-time behavior data of vehicle HEH with higher speed, if condition { c ] is met>0.8|R EL }∪{c>0.8|R EH -1, braking the own vehicle, otherwise not braking; r is R EL Real-time behavior data R for vehicle EL having low vehicle speed EH Is real-time behavior data of the vehicle EH with higher vehicle speed. { c>0.8|R EL The real-time behavior data R EL And under the current environmental data, the pedestrian existence probability c of the blind area of the visual field of the pavement is calculated through a pedestrian punching walking pavement model and a vehicle control behavior evaluation model>Logic value 0.8 (0 or 1), { c>0.8|R EH Similar to the above } { c>0.8|R EL }∪{c>0.8|R EH The expression "1" indicates that either side of the two-sided vehicle satisfies the condition c>At 0.8, the own vehicle immediately takes braking control.
Corresponding to the embodiment of the method, the invention also provides an unmanned vehicle control system based on the pedestrian existence prediction of the dead zone of the sidewalk, which comprises the following steps:
A first probability prediction module: the method comprises the steps of obtaining historical environment data related to a sidewalk, establishing a pedestrian-rushing-through sidewalk model, and predicting the probability of the pedestrian rushing-through sidewalk through the pedestrian-rushing-through sidewalk model;
a second probability prediction module: the method comprises the steps of acquiring historical vehicle behavior data when a pedestrian is encountered, establishing a vehicle control behavior evaluation model, and predicting the probability of the pedestrian in front of the side vehicle through the vehicle control behavior evaluation model;
pedestrian presence prediction module: the method comprises the steps of carrying out weighted summation on the probability of pedestrian impulse passing through a pavement and the probability of pedestrian occurrence in front of a lateral vehicle, and calculating the existence probability of pedestrians in a pavement vision blind area;
preventive control module: judging whether a preventive control strategy needs to be adopted or not based on the existence probability of pedestrians in the blind area of the pavement view.
The system embodiments and the method embodiments are in one-to-one correspondence, and the brief description of the system embodiments is just to refer to the method embodiments.
The invention also discloses an electronic device, comprising: at least one processor, at least one memory, a communication interface, and a bus; the processor, the memory and the communication interface complete communication with each other through the bus; the memory stores program instructions executable by the processor that the processor invokes to implement the aforementioned methods of the present invention.
The invention also discloses a computer readable storage medium storing computer instructions for causing a computer to implement all or part of the steps of the methods of the embodiments of the invention. The storage medium includes: a usb disk, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The system embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, i.e., may be distributed over a plurality of network elements. One of ordinary skill in the art may select some or all of the modules according to actual needs without performing any inventive effort to achieve the objectives of the present embodiment.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (9)
1. An unmanned vehicle control method based on pedestrian existence prediction in a pavement blind area is characterized by comprising the following steps:
acquiring historical environment data associated with a sidewalk, establishing a pedestrian-rushing-through sidewalk model, and predicting the probability of pedestrian rushing-through sidewalk by the pedestrian-rushing-through sidewalk model;
historical vehicle behavior data when a pedestrian is encountered is obtained, a vehicle control behavior evaluation model is established, and the probability that the pedestrian appears in front of the side vehicle is predicted through the vehicle control behavior evaluation model;
the probability of pedestrian impulse passing through the sidewalk and the probability of pedestrian occurrence in front of the side vehicles are weighted and summed, and the existence probability of the pedestrian in the vision blind area of the sidewalk is calculated;
judging whether a preventive control strategy is needed to be adopted or not based on the existence probability of pedestrians in the blind area of the pavement visual field;
the method for obtaining the environmental data associated with the sidewalk and establishing a pedestrian impact walking sidewalk model, and predicting the probability of the pedestrian impact walking sidewalk through the pedestrian impact walking sidewalk model specifically comprises the following steps:
acquiring historical environment data including precipitation, visibility, temperature, crowd age distribution, zebra crossing length, whether pedestrians and vehicles exist, red light remaining time, green light remaining time and time period, and the probability of pedestrian impulse passing through a sidewalk corresponding to the historical environment data;
Taking a feature vector formed by the historical environment data as input and taking the probability of pedestrian impulse passing through the sidewalk corresponding to the historical environment data as output to train a Bayesian model, so as to obtain a pedestrian impulse passing through sidewalk model;
and acquiring real-time environment data associated with the current sidewalk to form a feature vector, inputting a pedestrian-rushing-through sidewalk model, and outputting the probability of pedestrian rushing through the sidewalk.
2. The unmanned vehicle control method based on the pedestrian existence prediction of the blind area of the pavement according to claim 1, wherein the steps of obtaining the historical vehicle behavior data when the pedestrian is encountered, and establishing a vehicle control behavior evaluation model, and predicting the probability that the pedestrian appears in front of the side vehicle is encountered by the vehicle control behavior evaluation model specifically comprise:
acquiring historical vehicle behavior data including an original speed, a distance between the vehicle and a sidewalk, a vehicle size, a vehicle acceleration, a vehicle braking distance, a vehicle transverse deflection angle, a vehicle whistle state and frequency, and probability of an opportune pedestrian corresponding to the historical vehicle behavior data;
taking a feature vector formed by historical vehicle behavior data as input and taking probability of a pedestrian in chance as output to train a decision tree model to obtain a vehicle control behavior evaluation model;
And acquiring real-time behavior data of a side vehicle in the same direction as the own vehicle to form a feature vector, inputting a vehicle control behavior evaluation model, and outputting the probability of the occurrence of pedestrians in front of the side vehicle.
3. The unmanned vehicle control method based on the pedestrian existence prediction of the pavement blind area according to claim 2, wherein the step of judging whether the preventive control strategy is needed based on the magnitude of the pedestrian existence probability of the pavement blind area specifically comprises:
if the pedestrian existence probability c of the blind area of the pavement is less than 60% and the vehicle speed VE of any side vehicle with shielding action on own vehicles is less than the preset vehicle speed V0, directly performing control for eliminating the blind area of the pavement without adopting a preventive control strategy;
if the speed VE of a single vehicle with shielding behavior on own vehicles is greater than or equal to the preset speed V0, a preventive control strategy is needed to be adopted to carry out follow-up running control.
4. The unmanned vehicle control method based on the prediction of the existence of pedestrians in the blind area of a pavement according to claim 3, wherein the blind area eliminating control specifically comprises:
if only one side is blocked, determining the speed VE of the side blocking vehicle E; when 0< c <0.3, the speed VQ=VE+6 of the own vehicle, when 0.3 is less than or equal to c <0.6, the speed VQ=VE+3 of the own vehicle is controlled to drive to the head of the own vehicle to be level with the head of the side shielding vehicle, and all information on two sides of a road is identified through a laser radar arranged on the own vehicle; performing vehicle control according to the identified road information;
If both sides are blocked, determining the speeds VL and VH of the vehicles blocked at both sides, wherein VL is smaller than VH;
when 0< c <0.3, the speed VQ=VL+6 of the own vehicle, and when 0.3 is less than or equal to c <0.6, the speed VQ=VL+3 of the own vehicle, and the own vehicle is controlled to run to be level with the head of the vehicle with the speed VL; identifying road information on the side where the vehicle with the speed VL is located by a laser radar arranged on the own vehicle; when a pedestrian is walking on the sidewalk, if the transverse distance between the pedestrian and the own vehicle is greater than a preset distance d, accelerating the pedestrian to pass through, and if the transverse distance is less than d, enabling vq=0; if no pedestrian is present, the own vehicle speed vq=vh+3 is controlled to drive the own vehicle until the own vehicle head is flush with the vehicle head having the speed VH; all information on two sides of a road is identified through a laser radar arranged on the own vehicle; and controlling the vehicle according to the identified road information.
5. The unmanned vehicle control method based on pedestrian presence prediction of a blind area of a pedestrian as set forth in claim 3, wherein the following travel control includes:
if only one side is blocked, determining the speed VE of the side blocking vehicle E and the central point YE of the length of the vehicle body, and controlling the own vehicle to run until the head of the own vehicle is level with the central point YE of the length of the vehicle body and always follow the side blocking vehicle E to run;
If both sides are blocked, determining the speeds VL and VH of the vehicles blocked at both sides, wherein VL is smaller than VH;
if VL < V0, when 0< c <0.3, the own vehicle speed vq=vl+6, and when 0.3 is less than or equal to c <0.6, the own vehicle speed vq=vl+3, and the own vehicle is controlled to run until the own vehicle head is flush with the vehicle head with the speed VL; identifying road information on the side where the vehicle with the speed VL is located by a laser radar arranged on the own vehicle; when a pedestrian is walking on the sidewalk, if the transverse distance between the pedestrian and the own vehicle is greater than a preset distance d, accelerating the pedestrian to pass through, and if the transverse distance is less than d, enabling vq=0; when no pedestrian is present, accelerating the own vehicle beyond the vehicle with the speed VL to the head of the own vehicle to be level with the central point of the vehicle with the speed VH, and braking the own vehicle if c > 0.8;
if VL>V0, adjusting the own running speed, controlling the own vehicle to run to the same level as the vehicle body center point of the own vehicle with the vehicle speed VL, enabling the own vehicle to move along with the vehicle speed VL by VQ=VL, simultaneously, for the real-time behavior data of the vehicle with the vehicle speed VH, if the condition { c>0.8|R EL }∪{c>0.8|R EH -1, braking the own vehicle, otherwise not braking; r is R EL Real-time behavior data R for a vehicle with speed VL EH Is real-time behavior data of a vehicle having a vehicle speed VH.
6. The unmanned vehicle control method based on the pedestrian existence prediction of the pavement blind area according to claim 5, wherein the preset vehicle speed v0=15 km/h and the preset distance d=5 m.
7. An unmanned vehicle control system based on pedestrian existence prediction in a blind area of a sidewalk using the method of any one of claims 1 to 6, the system comprising:
a first probability prediction module: the method comprises the steps of obtaining historical environment data related to a sidewalk, establishing a pedestrian-rushing-through sidewalk model, and predicting the probability of the pedestrian rushing-through sidewalk through the pedestrian-rushing-through sidewalk model;
a second probability prediction module: the method comprises the steps of acquiring historical vehicle behavior data when a pedestrian is encountered, establishing a vehicle control behavior evaluation model, and predicting the probability of the pedestrian in front of the side vehicle through the vehicle control behavior evaluation model;
pedestrian presence prediction module: the method comprises the steps of carrying out weighted summation on the probability of pedestrian impulse passing through a pavement and the probability of pedestrian occurrence in front of a lateral vehicle, and calculating the existence probability of pedestrians in a pavement vision blind area;
Preventive control module: judging whether a preventive control strategy needs to be adopted or not based on the existence probability of pedestrians in the blind area of the pavement view.
8. An electronic device, comprising: at least one processor, at least one memory, a communication interface, and a bus;
the processor, the memory and the communication interface complete communication with each other through the bus;
the memory stores program instructions executable by the processor, the processor invoking the program instructions to implement the method of any of claims 1-6.
9. A computer readable storage medium storing computer instructions for causing a computer to implement the method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310619973.2A CN116639151B (en) | 2023-05-30 | 2023-05-30 | Unmanned vehicle control method and system based on pedestrian existence prediction in pavement blind area |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310619973.2A CN116639151B (en) | 2023-05-30 | 2023-05-30 | Unmanned vehicle control method and system based on pedestrian existence prediction in pavement blind area |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116639151A CN116639151A (en) | 2023-08-25 |
| CN116639151B true CN116639151B (en) | 2023-11-28 |
Family
ID=87614939
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310619973.2A Active CN116639151B (en) | 2023-05-30 | 2023-05-30 | Unmanned vehicle control method and system based on pedestrian existence prediction in pavement blind area |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116639151B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015225366A (en) * | 2014-05-26 | 2015-12-14 | 株式会社リコー | Accident prevention system, accident prevention device, and accident prevention method |
| CN109949612A (en) * | 2019-04-04 | 2019-06-28 | 哈尔滨工业大学 | Ghost probe accident pre-alarming system based on information exchange |
| CN109969172A (en) * | 2017-12-26 | 2019-07-05 | 华为技术有限公司 | Control method for vehicle, equipment and computer storage medium |
| CN111009125A (en) * | 2019-12-17 | 2020-04-14 | 武汉理工大学 | Urban intersection pedestrian and vehicle collision probability calculation method |
| CN114464012A (en) * | 2022-03-15 | 2022-05-10 | 徐州百事利电动车业有限公司 | Vehicle and pedestrian collision early warning method and system based on Internet of vehicles |
| CN114511999A (en) * | 2020-11-17 | 2022-05-17 | 郑州宇通客车股份有限公司 | Pedestrian behavior prediction method and device |
| CN115662113A (en) * | 2022-09-30 | 2023-01-31 | 合肥工业大学 | Signalized intersection people-vehicle game conflict risk assessment and early warning method |
| CN115862383A (en) * | 2023-01-20 | 2023-03-28 | 北京航空航天大学 | Vehicle blind area potential risk quantification method based on uncertainty |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10429842B2 (en) * | 2017-07-10 | 2019-10-01 | Toyota Research Institute, Inc. | Providing user assistance in a vehicle based on traffic behavior models |
| WO2020058740A1 (en) * | 2018-09-17 | 2020-03-26 | 日産自動車株式会社 | Vehicle behavior prediction method and vehicle behavior prediction device |
-
2023
- 2023-05-30 CN CN202310619973.2A patent/CN116639151B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015225366A (en) * | 2014-05-26 | 2015-12-14 | 株式会社リコー | Accident prevention system, accident prevention device, and accident prevention method |
| CN109969172A (en) * | 2017-12-26 | 2019-07-05 | 华为技术有限公司 | Control method for vehicle, equipment and computer storage medium |
| CN109949612A (en) * | 2019-04-04 | 2019-06-28 | 哈尔滨工业大学 | Ghost probe accident pre-alarming system based on information exchange |
| CN111009125A (en) * | 2019-12-17 | 2020-04-14 | 武汉理工大学 | Urban intersection pedestrian and vehicle collision probability calculation method |
| CN114511999A (en) * | 2020-11-17 | 2022-05-17 | 郑州宇通客车股份有限公司 | Pedestrian behavior prediction method and device |
| CN114464012A (en) * | 2022-03-15 | 2022-05-10 | 徐州百事利电动车业有限公司 | Vehicle and pedestrian collision early warning method and system based on Internet of vehicles |
| CN115662113A (en) * | 2022-09-30 | 2023-01-31 | 合肥工业大学 | Signalized intersection people-vehicle game conflict risk assessment and early warning method |
| CN115862383A (en) * | 2023-01-20 | 2023-03-28 | 北京航空航天大学 | Vehicle blind area potential risk quantification method based on uncertainty |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116639151A (en) | 2023-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110488802B (en) | Decision-making method for dynamic behaviors of automatic driving vehicle in internet environment | |
| EP3678911B1 (en) | Pedestrian behavior predictions for autonomous vehicles | |
| US20210339741A1 (en) | Constraining vehicle operation based on uncertainty in perception and/or prediction | |
| JP6193572B2 (en) | Vehicle or traffic control method and system | |
| US7102496B1 (en) | Multi-sensor integration for a vehicle | |
| JP4578795B2 (en) | Vehicle control device, vehicle control method, and vehicle control program | |
| US20170329332A1 (en) | Control system to adjust operation of an autonomous vehicle based on a probability of interference by a dynamic object | |
| JP5620147B2 (en) | Movable object prediction apparatus and program | |
| US20180370502A1 (en) | Method and system for autonomous emergency self-learning braking for a vehicle | |
| JP2005056372A5 (en) | ||
| CN111094095B (en) | Method and device for automatically sensing driving signal and vehicle | |
| CN110316186A (en) | Vehicle collision avoidance pre-judging method, device, equipment and readable storage medium storing program for executing | |
| US20240067215A1 (en) | Method and system for operating an autonomous agent with incomplete environmental information | |
| CN111252066A (en) | Emergency braking control method and device, vehicle and storage medium | |
| CN114430722A (en) | Security analysis framework | |
| JP4814816B2 (en) | Accident occurrence prediction simulation apparatus, method and program, safety system evaluation apparatus and accident alarm apparatus | |
| CN116206285A (en) | Traffic weakness group collision risk assessment method and system applied to automatic driving | |
| CN114446046A (en) | LSTM model-based weak traffic participant track prediction method | |
| CN116639151B (en) | Unmanned vehicle control method and system based on pedestrian existence prediction in pavement blind area | |
| CN114787894A (en) | Perceptual error model | |
| JP2021160531A (en) | Vehicle control device, vehicle control method, and program | |
| CN116842340A (en) | Pedestrian track prediction method and equipment based on ATT-GRU model | |
| CN116834655A (en) | Dangerous early warning method and device and vehicle | |
| Peng et al. | A Method for Vehicle Collision Risk Assessment through Inferring Driver's Braking Actions in Near-Crash Situations | |
| Yoshihira et al. | Developing an autonomous vehicle control system for intersections using obstacle/blind spot detection frames |
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 |