CN113763695A

CN113763695A - Dispatching method and system for automatic driving vehicle

Info

Publication number: CN113763695A
Application number: CN202111041922.3A
Authority: CN
Inventors: 李昌远; 阮春彬; 张皓
Original assignee: Beijing Voyager Technology Co Ltd
Current assignee: Beijing Voyager Technology Co Ltd
Priority date: 2021-09-07
Filing date: 2021-09-07
Publication date: 2021-12-07

Abstract

The embodiment of the application discloses a scheduling method of an automatic driving vehicle. The dispatching method of the automatic driving vehicle comprises the following steps: acquiring the position information of at least one automatic driving vehicle in a target area through a network; determining a hot spot area and a non-hot spot area of the target area according to the position information of the at least one automatic driving vehicle; obtaining vehicle information of at least one candidate autonomous vehicle via a network, the at least one candidate autonomous vehicle being located in the non-hotspot area; determining at least one target autonomous vehicle among the at least one candidate autonomous vehicle according to vehicle information of the at least one candidate autonomous vehicle; sending a scheduling instruction to the at least one target autonomous vehicle, the scheduling instruction instructing the at least one target autonomous vehicle to travel from the non-hot spot area to the hot spot area.

Description

Dispatching method and system for automatic driving vehicle

Technical Field

The present disclosure relates to the field of automatic driving technologies, and in particular, to a method and a system for scheduling an automatic driving vehicle.

Background

Similar to a traditional taxi or a network appointment, the self-driving vehicle service has the phenomenon of uneven transportation capacity, for example, fewer vehicles are used in an area with higher service demand; the number of vehicles is relatively high in the area with low service demand. The phenomenon of uneven transport capacity not only affects passenger experience, but also affects the effective utilization of the vehicle. Accordingly, it is desirable to provide a method and system for scheduling autonomous vehicles.

Disclosure of Invention

One of the embodiments of the present specification provides a scheduling method for an autonomous vehicle, the method including: acquiring the position information of at least one automatic driving vehicle in a target area through a network; determining a hot spot area and a non-hot spot area of the target area according to the position information of the at least one automatic driving vehicle; obtaining vehicle information of at least one candidate autonomous vehicle via a network, the at least one candidate autonomous vehicle being located in the non-hotspot area; determining at least one target autonomous vehicle among the at least one candidate autonomous vehicle according to vehicle information of the at least one candidate autonomous vehicle; sending a scheduling instruction to the at least one target autonomous vehicle, the scheduling instruction instructing the at least one target autonomous vehicle to travel from the non-hot spot area to the hot spot area.

One of the embodiments of the present specification provides a system for scheduling of autonomous vehicles, the system comprising: the system comprises a determining module, a processing module and a processing module, wherein the determining module is used for acquiring the position information of at least one automatic driving vehicle in a target area through a network, and determining a hot spot area and a non-hot spot area of the target area according to the position information of the at least one automatic driving vehicle; a selection module, configured to obtain vehicle information of at least one candidate autonomous vehicle via a network, where the at least one candidate autonomous vehicle is located in the non-hotspot area, and determine at least one target autonomous vehicle in the at least one candidate autonomous vehicle according to the vehicle information of the at least one candidate autonomous vehicle; a scheduling module configured to send a scheduling instruction to the at least one target autonomous vehicle, where the scheduling instruction instructs the at least one target autonomous vehicle to travel from the non-hot spot area to the hot spot area.

One of the embodiments of the present application provides an apparatus for dispatch of an autonomous vehicle, the apparatus comprising at least one processor and at least one memory; the at least one memory is for storing computer instructions; the at least one processor is configured to execute at least some of the computer instructions to implement the method of scheduling of autonomous vehicles as described above.

One embodiment of the present application provides a computer-readable storage medium, where the storage medium stores computer instructions, and when the computer reads the computer instructions in the storage medium, the computer executes the method for scheduling an autonomous vehicle as described above.

Drawings

The present description will be further explained by way of exemplary embodiments, which will be described in detail by way of the accompanying drawings. These embodiments are not intended to be limiting, and in these embodiments like numerals are used to indicate like structures, wherein:

FIG. 1 is a schematic diagram of an application scenario of a dispatch system for an autonomous vehicle according to some embodiments of the present application;

FIG. 2 is a block diagram of an exemplary processing device according to some embodiments of the present application;

FIG. 3 is an exemplary flow chart of a method of scheduling autonomous vehicles according to some embodiments of the present application;

FIG. 4 is an exemplary flow chart of a method of obtaining a target determination model according to some embodiments of the present application;

FIG. 5 is an exemplary flow chart of a training process for a goal determination model according to some embodiments of the present application; and

FIG. 6 is a schematic diagram of exemplary hardware components and/or software components of an exemplary computing device according to some embodiments of the present application.

Detailed Description

In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings used in the description of the embodiments will be briefly described below. It is obvious that the drawings in the following description are only examples or embodiments of the present description, and that for a person skilled in the art, the present description can also be applied to other similar scenarios on the basis of these drawings without inventive effort. Unless otherwise apparent from the context, or otherwise indicated, like reference numbers in the figures refer to the same structure or operation.

It should be understood that "system", "apparatus", "unit" and/or "module" as used herein is a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, other words may be substituted by other expressions if they accomplish the same purpose.

As used in this specification and the appended claims, the terms "a," "an," "the," and/or "the" are not intended to be inclusive in the singular, but rather are intended to be inclusive in the plural, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that steps and elements are included which are explicitly identified, that the steps and elements do not form an exclusive list, and that a method or apparatus may include other steps or elements.

Flow charts are used in this description to illustrate operations performed by a system according to embodiments of the present description. It should be understood that the preceding or following operations are not necessarily performed in the exact order in which they are performed. Rather, the various steps may be processed in reverse order or simultaneously. Meanwhile, other operations may be added to the processes, or a certain step or several steps of operations may be removed from the processes.

FIG. 1 is a schematic diagram of an application scenario of a dispatch system for an autonomous vehicle according to some embodiments of the present application. The dispatch system 100 may be an online transportation service platform for a transportation service. Such as a network appointment service, a taxi call service, a cargo delivery service, a docking service, etc. As shown in FIG. 1, a dispatch system 100 for an autonomous vehicle can include a server 110, a network 120, a requester terminal 130, a vehicle 140, a memory 150, and a location system 160.

The server 110 may process data and/or information obtained from at least one component of the system 100 (e.g., the requester terminal 130, the vehicle 140, the memory 150, and the location system 160) or an external data source (e.g., a cloud data center). For example, the server 110 may accept a service order from the requester terminal 130. As another example, server 110 may also retrieve historical data from storage 150. As another example, the server 110 may obtain the real-time location of the vehicle from the vehicle 140.

In some embodiments, the server 110 may include a processing device 112. The processing device 112 may process information and/or data related to the scheduled system of the autonomous vehicle to perform one or more of the functions described herein. For example, processing device 112 may determine a dispatch instruction based on the capacity condition and send the dispatch instruction to the target autonomous vehicle instructing the target dispatch vehicle to travel from a non-hot spot area to a hot spot area.

The network 120 may provide a conduit for the exchange of information. In some embodiments, network 120 may include one or more network access points. One or more components of system 100 may connect to network 120 through an access point to exchange data and/or information. In some embodiments, at least one component in system 100 may access data or instructions stored in memory 150 via network 120.

Requester terminal 130 may refer to one or more terminal devices or software used by a user. The requester terminal 130 may process information and/or data associated with the service order to perform one or more of the functions disclosed in this specification. For example, a service requester may send a service request order to the server 110 through the requester terminal 130. In some embodiments, the requester terminal 130 may comprise one or more combinations of a mobile device, a tablet computer, a laptop computer, a device built into a motor vehicle, and the like.

Vehicle 140 refers to an autonomous vehicle that may be used to dispatch transportation services of system 100. In this application, an autonomous vehicle may refer to a smart car that automatically controls the vehicle to perform a particular operation (e.g., driving, parking, light control, etc.) via an in-vehicle computer system. In the present application, the vehicle 140 may include a fully autonomous vehicle and/or a dual-purpose vehicle for unmanned and manned. In some embodiments, the vehicle 140 may receive the command from the server 110 and perform a corresponding task, such as refueling, charging, service, order taking, scheduling, etc., in accordance with the command. The vehicle 140 may include any type of automobile, such as, for example, a sedan, an off-road vehicle, a commercial vehicle, a van, a bus, etc., or any combination thereof.

In some embodiments, the memory 150 may store data and/or instructions that the processing device 112 may perform or use to perform the exemplary methods described in this specification. For example, the memory 150 may store location information of the vehicle 140, historical thermodynamic diagrams, scheduling instructions, vehicle information of the vehicle 140, and the like. As another example, the memory 150 may store programs and/or software for generating thermodynamic diagrams and/or scheduling instructions. In some embodiments, the memory 150 may be directly connected to the server 110 as a back-end memory. In some embodiments, memory 150 may be part of server 110, requester terminal 130, and/or vehicle 140.

The positioning system 160 may determine the current position of the vehicle 140. In some embodiments, the positioning system 160 may include a combination of one or more of the Global Positioning System (GPS), the global navigation satellite system (GLONASS), the beidou navigation system, the Galileo positioning system (Galileo), the quasi-zenith satellite system (QAZZ), and the like. In some embodiments, the positioning system may include one or more satellite vehicles, such as satellite 160-1, satellite 160-2, and satellite 160-3. The positioning system 160 may transmit the current location of the vehicle 140 to the network 120 or the vehicle 140 via, for example, a wireless connection.

FIG. 2 is a block diagram of an exemplary processing device according to some embodiments of the present application. As shown in fig. 2, the system 200 may include: the system comprises a position information acquisition module 210, a determination module 220, a vehicle information acquisition module 230, a selection module 240 and a scheduling module 250.

The location information acquisition module 210 may be configured to acquire location information of at least one autonomous vehicle (e.g., vehicle 140) within a target area via a network (e.g., network 120).

The determining module 220 may be configured to determine a hot spot area and a non-hot spot area of the target area according to the location information of the at least one autonomous vehicle. In some embodiments, the determination module 220 may perform region division on the target region, dividing the target region into a plurality of partitions. For each zone, the determination module 220 may determine heat information for the zone based on the number of available autonomous vehicles (simply "available vehicles") associated with the zone and the number of pending orders associated with the zone. From the heat information, the determining module 220 may determine whether the partition is a hot spot region or a non-hot spot region. In some embodiments, the determination module 210 may be further configured to: hot spot areas and non-hot spot areas are shown in a cellular thermodynamic diagram.

Vehicle information acquisition module 230 may be configured to acquire vehicle information of at least one candidate autonomous vehicle (simply "candidate vehicle") located in the non-hotspot area via a network (e.g., network 120).

The selection module 240 may be configured to determine at least one target autonomous vehicle (simply "target vehicle") among the at least one candidate autonomous vehicle based on vehicle information of the at least one candidate autonomous vehicle. In some embodiments, the selection module 240 may also be configured to: selecting the at least one target autonomous vehicle from the at least one candidate autonomous vehicle according to a scheduling preset condition. In some embodiments, for each of the candidate vehicles, the selection module 240 may determine the fitness of the candidate vehicle through a goal determination model based on the vehicle information. The selection module 240 may determine the at least one target autonomous vehicle based on a fitness of the at least one candidate autonomous vehicle. In some embodiments, for each of the at least one target autonomous vehicle, the selection module 240 may determine the dispatchable attribute of the target vehicle based on the location information and the current time.

The scheduling module 250 may be configured to send a scheduling directive to the at least one target autonomous vehicle, which may instruct the at least one target autonomous vehicle to travel from the non-hot spot area to the hot spot area.

FIG. 3 is an exemplary flow chart of a method of scheduling autonomous vehicles according to some embodiments of the present application. In some embodiments, flow 300 may be performed by server 110 (e.g., processing device 112, processor 620, and/or one or more of the modules described in fig. 2).

At step 310, location information of at least one autonomous vehicle (e.g., vehicle 140) within a target area is obtained via a network (e.g., network 120).

The target area may refer to a geographic area serviced by the vehicle 140. For example, the target area may include one or more administrative areas. As an example, the target area may include beijing. As yet another example, the target area may include the haih lake district of beijing. As yet another example, the target area may include beijing and zhangjiakou.

The location information may include the real-time location of the vehicle 140 located by the location system 160. The vehicle 140 may transmit the location information to the processing device 112 via the network 120. The location information may include one or a combination of location coordinates (e.g., which may be expressed in terms of latitude and longitude), altitude, speed, acceleration, direction, corresponding time, etc.

Step 320, determining a hot spot area and a non-hot spot area of the target area according to the position information of the at least one automatic driving vehicle.

In some embodiments, the target area may be divided into a plurality of zones. The area division may be a uniform division or a non-uniform division. For example, an area is divided equally by a grid having a length and a width of 1 km. For example, a large-area zone where vehicles cannot pass, such as a lake, is divided into one area, and the land portion is uniformly divided. For another example, a region may have one or more centers that are unevenly spaced outwardly from a center, with portions closer to the center being spaced less apart and portions farther from the center being spaced more apart. In some embodiments, the locations in the regions may be continuous or discontinuous. For example, a region separated by a river is divided into the same division, and the river is divided into another division. In some embodiments, the region division may be performed in one time, or may be a combination of multiple division results. For example, regions are merged or subdivided according to some conditions on the basis of the result of the previous division. For another example, in the process of using the division result, the division result can be continuously adjusted according to actual needs. The representation of the area may include, but is not limited to, a description of coordinates, a description of latitude and longitude, and/or other ways in which location information may be determined.

In some embodiments, the area division may be performed in a grid. For example, the target region may be divided into squares or regular hexagons (honeycomb shapes) of equal size, or the like. In some embodiments, the zone partitions may be performed in clusters. For example, the real-time locations of the vehicles 140 may be clustered to form different zones based on the location information. As another example, pick-up points of the target area to which orders are to be assigned may be clustered to form different zones. In some embodiments, the zone division may be performed according to certain rules (e.g., administrative area, geographic information). Wherein the administrative region includes, but is not limited to, one or more of province, city, county, town, street, etc. For example, the target area is Beijing, and the sea area may be divided into one partition. The geographic information may include, but is not limited to, one or more of a landscape, a weather, precipitation, geology, hydrological information, and the like. For example, locations with an average altitude within a certain threshold range may be divided into one partition.

The above description of the region division method is merely a specific example and should not be considered as the only feasible embodiment. It is clear that, after understanding the basic principles of the various partitioning methods, it is possible for a person skilled in the art to carry out various modifications and changes in form and detail of the specific embodiments and steps of the partitioning method without departing from this principle, but these modifications and changes are still within the scope of the above description. For example, a random division may be used, and some positions may be randomly selected to be classified into a partition. For example, the area division method already existing in another system may be directly adopted.

In some embodiments, for each bay, the heat information for the bay may be determined based on the number of available autonomous vehicles (simply "available vehicles") associated with the bay and the number of pending orders associated with the bay. From the heat information, it can be determined whether the partition is a hot spot region or a non-hot spot region.

The available vehicles may include one or more of a pool vehicle having a remaining capacity (e.g., remaining number of passengers, remaining cargo weight, and/or volume), a vehicle that has not yet received an order (an idle vehicle), a vehicle that is about to complete an order within a first predetermined time period (e.g., 1 minute, 2 minutes, 5 minutes, 10 minutes, etc.), and the like. The available vehicles associated with the bays may include one or a combination of more of the available vehicles currently located within the bays, the available vehicles that will be entering the bays for a second predetermined period of time (e.g., 1 minute, 2 minutes, 5 minutes, 10 minutes, 20 minutes, 30 minutes, etc.), the available vehicles that will be exiting the bays for a third predetermined period of time (e.g., 1 minute, 2 minutes, 5 minutes, 10 minutes, 20 minutes, 30 minutes, etc.), and the like. The number of available vehicles associated with the zone may be determined based on the location information.

A pending order may refer to an order that has not been picked. The pending orders associated with a bay may include one or a combination of a number of pending orders currently within the bay (e.g., pick-up points within the bay), a number of orders to be picked up predicted within a fourth predetermined time period (e.g., 1 minute, 2 minutes, 5 minutes, 10 minutes, 20 minutes, 30 minutes, etc.), a number of orders to be picked up predicted within a fifth predetermined time period (e.g., 1 minute, 2 minutes, 5 minutes, 10 minutes, 20 minutes, 30 minutes, etc.), a number of pending orders to be added predicted within a fifth predetermined time period, etc. In some embodiments, the first preset time period, the second preset time period, the third preset time period, the fourth preset time period, and the fifth preset time period are future time periods.

In some embodiments, the heat information is the number of available vehicles associated with a zone-the number of pending orders associated with a zone. If the difference is a positive number, the capacity in the partition can be considered to be sufficient, and the partition can be determined as a non-hot spot area; if the difference is negative and the absolute value is less than or equal to a first difference threshold (e.g., 0, 10, 20, 50, 100, etc.), then the partition is deemed to be sufficiently powerful and may be determined to be a non-hot spot region; if the difference is negative and the absolute value is greater than the first difference threshold, the partition capacity may be considered to be a gap, and the partition may be determined to be a hot spot area. The smaller the negative difference (the larger the absolute value of the negative difference), the larger the capacity gap of the partition.

In some embodiments, the heat information is the number of available vehicles/pending orders associated with the zone. If the ratio is greater than or equal to a first ratio threshold (e.g., 1, 0.9, 0.8, 0.7, 0.6, 0.5, etc.), then the capacity within the zone may be deemed sufficient and the zone may be determined to be a non-hot spot region; if the ratio is less than the first ratio threshold, the partition capacity may be considered to be a gap, and the partition may be determined to be a hot spot area. The smaller the ratio is, the larger the capacity gap of the partition is.

In some embodiments, the popularity information may be determined based on the estimated number of orders to be picked up in the fourth preset time period, the number of orders to be processed currently in the partition, and the estimated number of orders to be processed to be increased in the fifth preset time period. For example, the heat information is the estimated number of orders to be taken in the fourth preset time period- (the current number of orders to be processed in the partition + the estimated number of orders to be processed in the fifth preset time period). If the absolute value of the difference is less than or equal to a second difference threshold (e.g., 0, 10, 20, 50, 100, etc.), then the partition is deemed to be sufficiently powerful and may be determined to be a non-hot spot region; if the absolute value of the difference is greater than the second difference threshold, the partition capacity may be considered to be a gap, and the partition may be determined to be a hot spot area. The larger the absolute value of the difference is, the larger the capacity gap of the partition is.

For another example, the hotness information is the estimated number of orders to be picked up in the fourth preset time period/(the current number of orders to be processed in the partition + the estimated number of orders to be processed in the fifth preset time period). If the ratio is greater than or equal to a second ratio threshold (e.g., 1, 0.9, 0.8, 0.7, 0.6, 0.5, etc.), then the partition is deemed to be sufficiently powerful and may be determined to be a non-hot spot region; if the ratio is less than the second ratio threshold, the partition capacity may be considered to be a gap, and the partition may be determined to be a hot spot region. The smaller the ratio is, the larger the capacity gap of the partition is.

In some embodiments, the first difference threshold, the second difference threshold, the first ratio threshold, and the second ratio threshold may be default values of the system 100, or may be adjusted according to actual requirements. For example, when the current time is at a peak of service (e.g., a commute peak 7: 00-9: 00, 17: 00-19: 00), the pending orders are relatively more, a larger first difference threshold (e.g., 50, 100, etc.) or second difference threshold (e.g., 50, 100, etc.) may be set, or a smaller first ratio threshold (e.g., 0.7, 0.6, 0.5, etc.) or second ratio threshold (e.g., 0.7, 0.6, 0.5, etc.) may be set. For another example, when the current time is in the service idle period (e.g., 10: 00-16: 00), the pending orders are relatively few, and a smaller first difference threshold (e.g., 0, 10, 20, etc.) or second difference threshold (e.g., 0, 10, 20, etc.) may be set, or a larger first ratio threshold (e.g., 1, 0.8, 0.9, etc.) or second ratio threshold (e.g., 1, 0.8, 0.9, etc.) may be set.

In some embodiments, the processing device 112 may display hot spot areas and/or non-hot spot areas. The display form may include, but is not limited to, a combination of one or more of voice, text, graphics, video, and the like. For example, hot spot regions and non-hot spot regions may be displayed in different colors, with darker colors giving greater zonal capacity gaps. For example, hot spot regions are displayed in red and non-hot spot regions are displayed in green.

In some embodiments, the processing device 112 (e.g., input/output 660) may exhibit hot spot areas and non-hot spot areas in a cellular thermodynamic diagram. Specifically, the shape of each partition of the target area is a honeycomb shape (regular hexagon), hot spot areas are displayed in red, and non-hot spot areas are displayed in green. In some embodiments, the honeycomb thermodynamic diagram may further include a histogram corresponding to the hot spot region, and the higher the columns of the histogram, the larger the capacity gap of the hot spot region represented by the honeycomb. The cellular thermodynamic diagram can visually display hot spots and non-hot spots so that a dispatcher can distinguish the hot spots from the non-hot spots in real time.

At step 330, vehicle information of at least one candidate autonomous vehicle (referred to as "candidate vehicle") is obtained via a network (e.g., network 120), the at least one candidate autonomous vehicle being located in the non-hotspot region. In some embodiments, at least one available vehicle may be selected as a candidate vehicle in a non-hotspot region based on the location information. In some embodiments, the vehicle information may include one or more of usage (e.g., taxi, express, special car, windmill, carpool, commercial car, rental car, shared car, etc.), remaining energy (e.g., fuel, electricity), remaining capacity (e.g., remaining number of passengers, remaining cargo weight and/or volume), pickup status (e.g., remaining capacity, not yet picked up, about to complete an order within a first predetermined time period), goodwill, hot spot distance, etc.

In some embodiments, the hotspot distance is determined based on a calculated distance of the candidate autonomous vehicle from at least one reference point of a hotspot region. The reference point includes at least one of a center point or at least one boundary point of the hotspot region. In some embodiments, the calculated distance between the candidate vehicle and the at least one reference point may be determined prior to determining the hotspot distance of the candidate vehicle based on the at least one calculated distance.

In some embodiments, the calculated distance of the candidate vehicle from a reference point may be related to one or more of: a straight-line distance, at least one route distance, at least one travel time. The travel time is related to the road condition. For example, more than one travel route is included between the candidate vehicle and the reference point. Firstly, determining the route distance and the driving time of the more than one driving routes; a weighted average of the one or more route distances is then determined to determine a calculated distance of the candidate vehicle from the reference point. The weight for each route distance may be determined based on the route distance and the corresponding travel time. For example, the calculated distance may reflect the distance traveled by the candidate vehicle to the reference point and the time, so the greater the distance of the route, the greater the corresponding weight; the longer the travel time, the greater the corresponding weight.

For example, if the at least one reference point comprises more than one reference point, more than one calculated distance needs to be determined. The hotspot distance may be determined by an arithmetic or weighted average of the one or more calculated distances. The weight corresponding to each calculated distance may be determined based on the corresponding calculated distance value. For example, the hotspot distance may reflect the distance between the candidate vehicle and the hotspot region, and thus, the greater the calculated distance, the greater the corresponding weight.

Step 340, determining at least one target autonomous vehicle (simply referred to as "target vehicle") among the at least one candidate autonomous vehicle according to the vehicle information of the at least one candidate autonomous vehicle.

In some embodiments, the at least one target autonomous vehicle may be selected from the at least one candidate autonomous vehicle according to a scheduled preset condition, the scheduled preset condition being associated with vehicle information and a scheduling requirement. The scheduling requirements may reflect requirements and limitations for the vehicle information of the target vehicle. For example, the desirability of requiring the target vehicle in the dispatch requirement needs to be greater than a desirability threshold (e.g., 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, etc.). As another example, the dispatch requirements may require that the target vehicle be a business vehicle. As another example, the dispatch request may require the target vehicle to be a free vehicle.

In some embodiments, for each of the candidate vehicles, a fitness of the candidate vehicle may be determined by a target determination model based on the vehicle information. The fitness may refer to a probability that the candidate vehicle receives the dispatching instruction or a probability that the candidate vehicle meets the order within a sixth preset time after driving to the hot spot region according to the dispatching instruction. The target determination model is a machine learning model with preset parameters. In some embodiments, the machine learning model may include a neural network model. In some embodiments, the Neural Network model may include a Convolutional Recurrent Neural Network (CRNN), a Convolutional Neural Network (CNN), a Deep Convolutional Neural Network (DCNN), a Recurrent Neural Network (RNN), or a Long Short Term Memory (LSTM) model, among others. In some embodiments, candidate vehicles having a suitability greater than a suitability threshold (e.g., 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, etc.) may be determined as target vehicles. The acquisition process of the target determination model can be referred to the related description of fig. 4 and fig. 5.

In some embodiments, the target determination model may predict the fitness based on the input vehicle information. Specifically, for each candidate vehicle, the vehicle information of the candidate vehicle may be encoded as a vector as an input object determination model, and the final layer neural network outputs the fitness of the candidate vehicle through each layer of neural network in the object determination model. As an example, the input vehicle information may include: express bus, 70% of residual electric quantity, 5 seats, 80% of favorable degree and 5km of hot spot distance, and the order has not been received. The output result is: the fitness is 90%.

In some embodiments, the at least one target vehicle may be determined based on a dispatch requirement and the suitability of the at least one candidate vehicle. For example, candidate vehicles may be screened based on the dispatch requirements, the fitness of the screened vehicles may be determined using the target determination model, and the target vehicle may be determined based on the fitness. For another example, the target determination model may be used to determine the fitness of the candidate vehicles, and then the candidate vehicles may be screened based on the scheduling requirements and the fitness to determine the target vehicle.

In some embodiments, for each of the at least one target autonomous vehicle, a dispatchable attribute of the target vehicle may be determined based on the location information and the current time, and the dispatchable attribute may be used to determine whether an order may be dispatched to the target vehicle when the target vehicle is within a first preset interval. The first preset interval may be related to time and/or location. For example, the first preset section needs to satisfy a time condition (e.g., at commute peak 7: 00-9: 00, 17: 00-19: 00) and/or a location condition (e.g., a straight-line distance and/or a traveling distance of the target vehicle from the hot spot area is greater than a distance threshold (e.g., 2km, 1.5km, 1km, 500m, 400m, 300m, etc.) during traveling of the target vehicle to the hot spot area). As another example, a dispatchable attribute may indicate that the target vehicle is or is not permitted to be dispatched all the way to the hot spot area while the target vehicle is traveling to the hot spot area. The first preset interval comprises a distance interval and a time interval of the target vehicle driving to the hot spot area.

Step 350 of sending a scheduling instruction to the at least one target autonomous vehicle, the scheduling instruction instructing the at least one target autonomous vehicle to travel from the non-hot spot area to the hot spot area.

In some embodiments, the scheduling instructions may include at least one of a scheduling location, a scheduling departure time, a scheduling arrival time, a scheduling route. For example, a target autonomous vehicle may be sent a dispatch instruction to dispatch to a central customs subway station. In some embodiments, the scheduling instructions may further include whether the target autonomous vehicle is allowed to pick up orders in a first preset interval.

In some embodiments, after the target autonomous vehicle receives the dispatch instruction, the in-vehicle computing system may choose to accept or reject the dispatch instruction. In some embodiments, the target autonomous vehicle is not allowed to reject the maneuver instruction after receiving the maneuver instruction, which may automatically control an in-vehicle computing system to execute the maneuver instruction to travel to the maneuver location.

The dispatch for traditional service vehicles (e.g., net appointments, taxis, etc.) is typically a dispatch for the driver. In the present application, the scheduling of autonomous vehicles is aimed at. After the automatic driving vehicle receives the dispatching instruction, the in-vehicle computer system can automatically analyze the dispatching instruction and select acceptance or rejection; or after the automatic driving vehicle receives the dispatching instruction, the automatic driving vehicle is not allowed to reject the dispatching instruction, and the dispatching instruction can automatically control the in-vehicle computer system to execute the dispatching instruction. Through the method and the device, the scheduling problem of the automatic driving vehicle is solved, the automatic driving vehicle can be effectively scheduled in time, and the utilization rate of the automatic driving vehicle is improved.

FIG. 4 is an exemplary flow diagram illustrating determination of a targeting model according to some embodiments of the present description. In some embodiments, flow 400 may be performed offline by server 110 (e.g., processing device 112, processor 620, and/or one or more of the modules described in fig. 2) or an external device.

At step 410, at least one historical scheduling instruction is obtained.

The historical scheduling instructions may refer to scheduling instructions that have been sent to the autonomous vehicle. In some embodiments, the at least one historical scheduling instruction may be retrieved from the memory 150.

And step 420, marking the at least one historical scheduling instruction to obtain a training label.

In some embodiments, if the historic dispatch instructions allow the autonomous vehicle to reject the historic dispatch instructions, the autonomous vehicle acceptance or rejection historic dispatch instructions are labeled as a label for the training sample, e.g., the autonomous vehicle acceptance historic dispatch instruction is labeled as "1" and the autonomous vehicle rejection historic dispatch instruction is labeled as "0". In some embodiments, if the historical scheduling command does not allow the autonomous vehicle to reject the historical scheduling command, and the historical scheduling command may automatically control the autonomous vehicle to travel to the hot spot area after being sent to the autonomous vehicle, whether the order is marked as the training sample within a sixth preset time (e.g., 1 minute, 2 minutes, 5 minutes, 10 minutes, 20 minutes, 30 minutes, etc.) after the autonomous vehicle is scheduled to travel to the hot spot area is received, for example, whether the order is marked as "1" within the sixth preset time after the autonomous vehicle is scheduled to travel to the hot spot area, and the missed order is marked as "0".

Step 430, extracting sample characteristics of the at least one historical scheduling instruction.

In some embodiments, the sample features may include: the vehicle information of the autonomous vehicle that received the historic dispatch instruction (corresponding to the "vehicle information" described in step 340 of fig. 3) and the dispatch location of the historic dispatch instruction.

Step 440, obtain an initial model. The initial model may be an untrained model.

And 450, training the initial model according to the training labels and the sample characteristics to obtain a target determination model.

In some embodiments, the training labels and the sample features may be used as training samples to train the initial model to obtain the target determination model. The training process for the object determination model can be seen in fig. 5 and its associated description.

FIG. 5 is an exemplary flow diagram of a training process for a goal determination model, shown in some embodiments herein.

In some embodiments, the initial model may be trained by performing one or more iterations. The parameters of the initial model may be initialized prior to one or more iterations. For example, the connection weights of the nodes of the initial model and/or the deviation vectors of the nodes may be initialized by assigning random values in the range of-1 to 1. As another example, the weights of all connections of the initial model may be assigned a same value, ranging from-1 to 1, such as 0. Still by way of example, the bias vectors for the nodes in the initial model may be initialized by assigning random values ranging from 0 to 1. In some embodiments, the parameters of the initial model may be initialized based on a gaussian random algorithm, an Xavier algorithm, etc., and then one or more iterations may be performed to update the parameters of the initial model until a preset condition is satisfied.

As shown in fig. 5, the process 500 includes the following steps:

step 510, inputting the sample characteristics into the intermediate model to obtain the sample fitness. The intermediate model includes the initial model if the current iteration is a first iteration of the training process. If the current iteration is not the first iteration of the training process, the intermediate model comprises an updated model generated in the last iteration.

Step 520, determining a loss function according to the fitness of the training labels and the samples.

In some embodiments, the difference in fitness of the training labels to the sample may be used to determine a loss function.

At step 530, it is determined whether the current iteration satisfies a predetermined condition. The preset condition may indicate whether the initial model is sufficiently trained. For example, in the current iteration, the preset condition may be satisfied if the value of the loss function is minimal or less than a first threshold. For another example, if the value of the loss function converges, the preset condition may be satisfied. Convergence may be considered if the change in the value of the loss function in two or more consecutive iterations is less than a second threshold. Still by way of example, the preset condition may be satisfied when a specified number of times or iterations are performed during the training process.

And 540, if the preset condition is met, determining the intermediate model of the current iteration as a target determination model.

Step 550, if it is determined that the predetermined condition is not satisfied, the intermediate model is updated according to the loss function, and then a new iteration is started, for example, step 510 and step 530 are repeated until the predetermined condition is satisfied.

Based on the difference (also referred to as global error) between the test value (e.g., sample fitness) and the expected value (e.g., training label) of the intermediate model in the iteration, the value of the loss function of the intermediate model in that iteration may be determined. Further, whether the loss function result in the iteration satisfies the preset condition can be judged based on the value of the loss function. For example, if the value of the loss function exceeds a preset threshold in the iteration, the intermediate model in the iteration does not satisfy the preset condition. The parameters of the intermediate model may be adjusted and/or updated for use in the next iteration. For example, the values of the parameters may be updated by executing a back-propagation machine learning training algorithm (e.g., a random gradient descent back-propagation training algorithm). If the loss function in the iteration is smaller than the preset threshold value, the condition that the loss function meets the preset condition can be judged, the iteration process can be terminated, and the target determination model is obtained. In some embodiments, after learning is complete, the validation set may be processed to validate the learning results.

FIG. 6 is a schematic diagram of exemplary hardware components and/or software components of an exemplary computing device according to some embodiments of the present application. In some embodiments, one or more elements of system 100 to automate the dispatch of vehicles may be implemented on computing device 600. For example, the processing device 112 may be implemented on the computing device 600 and configured to implement the functions and/or methods disclosed herein.

Computing device 600 may include a communication port 650 connected to network 120 for enabling data communications. Computing device 600 may include a processor (e.g., CPU)620 that may execute program instructions in the form of one or more processors. The exemplary computing device 600 may also include an internal bus 610, as well as various forms of program memory and data memory. For example, hard disk 670, Read Only Memory (ROM)630, or Random Access Memory (RAM)640 may be used to store various types of data files that are processed and/or transmitted by a computer. Exemplary computing device 400 may include program instructions stored in read-only memory 630, random access memory 640, and/or other types of non-transitory storage media that are executed by processor 620. The methods and/or processes of the present specification can be implemented as program instructions. Computing device 600 may also include input/output component 660 for supporting input/output between the computer and other components. The computing device 600 may also receive programs and data in the present application via network communication.

Having thus described the basic concept, it will be apparent to those skilled in the art that the foregoing detailed disclosure is to be regarded as illustrative only and not as limiting the present specification. Various modifications, improvements and adaptations to the present description may occur to those skilled in the art, although not explicitly described herein. Such modifications, improvements and adaptations are proposed in the present specification and thus fall within the spirit and scope of the exemplary embodiments of the present specification.

Claims

1. A method of scheduling an autonomous vehicle, comprising:

acquiring the position information of at least one automatic driving vehicle in a target area through a network;

determining a hot spot area and a non-hot spot area of the target area according to the position information of the at least one automatic driving vehicle;

obtaining vehicle information of at least one candidate autonomous vehicle via a network, the at least one candidate autonomous vehicle being located in the non-hotspot area;

determining at least one target autonomous vehicle among the at least one candidate autonomous vehicle according to vehicle information of the at least one candidate autonomous vehicle;

sending a scheduling instruction to the at least one target autonomous vehicle, the scheduling instruction instructing the at least one target autonomous vehicle to travel from the non-hot spot area to the hot spot area.

2. The method of claim 1,

the vehicle information comprises one or more of usage, residual energy, load receiving state, goodness of comment and hot spot distance;

the hotspot distance is determined based on a calculated distance of the candidate autonomous vehicle from at least one reference point of the hotspot region;

the reference point comprises at least one of a center point or at least one boundary point of the hotspot zone.

3. The method of claim 2, wherein the calculated distance is related to one or more of: a straight-line distance, at least one route distance, at least one travel time; the travel time is related to the road condition.

4. The method of claim 1, wherein determining at least one target autonomous vehicle among the at least one candidate autonomous vehicle based on vehicle information of the at least one candidate autonomous vehicle comprises:

determining the at least one target autonomous vehicle from the at least one candidate autonomous vehicle according to a scheduling preset condition; and the number of the first and second groups,

the scheduling preset condition is related to the vehicle information and the scheduling requirement.

5. The method of claim 1, wherein the method further comprises:

for each of the at least one target autonomous vehicle, determining a dispatchable attribute of the target autonomous vehicle based on the location information and the current time, the dispatchable attribute being used to determine whether to dispatch an order to the target autonomous vehicle when the target autonomous vehicle is within a first preset interval.

6. The method of claim 1, wherein the scheduling instruction includes whether the target autonomous vehicle is allowed to pick up orders in a first preset interval.

7. The method of claim 1, wherein determining hot spot areas and non-hot spot areas of the target area based on the location information of the at least one autonomous vehicle comprises:

aggregating the position information based on an equal-size hexagonal network to determine heat information; and

determining the hotspot region and the non-hotspot region based on the heat information.

8. A dispatch system for an autonomous vehicle, comprising:

the system comprises a position information acquisition module, a position information acquisition module and a control module, wherein the position information acquisition module is used for acquiring the position information of at least one automatic driving vehicle in a target area through a network;

the determining module is used for determining a hot spot area and a non-hot spot area of the target area according to the position information of the at least one automatic driving vehicle;

a vehicle information acquisition module for acquiring vehicle information of at least one candidate autonomous vehicle through a network, the at least one candidate autonomous vehicle being located in the non-hotspot area;

a selection module to determine at least one target autonomous vehicle among the at least one candidate autonomous vehicle based on vehicle information of the at least one candidate autonomous vehicle;

a scheduling module configured to send a scheduling instruction to the at least one target autonomous vehicle, where the scheduling instruction instructs the at least one target autonomous vehicle to travel from the non-hot spot area to the hot spot area.

9. The system of claim 8, wherein:

10. The system of claim 9, wherein the calculated distance is related to one or more of: a straight-line distance, at least one route distance, at least one travel time; the travel time is related to the road condition.

11. The system of claim 8, wherein the selection module is further to:

12. The system of claim 8, wherein the scheduling module is further to:

13. The system of claim 8, wherein the scheduling instruction includes whether the target autonomous vehicle is allowed to pick up orders in a first preset interval.

14. The system of claim 8, wherein the determination module is further to:

15. A scheduling apparatus for an autonomous vehicle, the apparatus comprising at least one processor and at least one memory;

the at least one memory is for storing computer instructions;

the at least one processor is configured to execute at least some of the computer instructions to implement the method of any of claims 1-7.

16. A computer-readable storage medium, wherein the storage medium stores computer instructions, which when executed by a processor, implement the method of any one of claims 1 to 7.