DE102019133337A1 - SYSTEM AND METHOD FOR TRANSMISSION OF FLEET DATA TO A VEHICLE FOR AN ENERGY DEMAND FORECAST FOR THE VEHICLE - Google Patents

Abstract

The invention relates to a system and method for transmitting fleet data (11) to a vehicle (1) for an energy demand forecast (15) of the vehicle (1). The method comprises a grouping (20) of one- or multi-dimensional first characteristic maps (12), each of which comprises a probability distribution of at least one physical dynamic parameter for an individual segment (2) of a route section (3) to be traveled by the vehicle (1) the fleet data (11) in several groups (G1, G2), an assignment (22) in each case a second characteristic map (31, 32) selected from the first characteristic maps (12) of each group (G1, G2) or with the aid of the first Feature maps (12) of each group (G1, G2) is calculated as a representative (R1, R2) for one of the groups (G1, G2), the representatives (R1, R2) having a code book (23) outside the vehicle (1 ) form (24) an assignment (26) of each first characteristic map (12) to one of the representatives (R1, R2) outside the vehicle (1), a transfer of the code book (23) to the vehicle (1) in a first transfer step (27) and of at least one of the assignments (26) to the vehicle (1) in a second transfer step (28), and selecting (29) the second characteristic map (31, 32), which is used as a representative (R1, R2) in the assignment (26) is stored from the code book (23) in the vehicle (1) for creating the energy demand forecast (15) of the vehicle (1). The invention also relates to a software program and a system for transmitting the fleet data (11) to the vehicle (1) for the energy demand forecast (15) of the vehicle (1).

Description

The disclosure relates to a system and method for transmitting fleet data to a vehicle for an energy demand forecast of the vehicle.

Hybrid or electric vehicles are driven by an electric motor, with the necessary electrical energy being stored, for example, in a high-voltage storage device. The high-voltage battery can be charged at a charging station in a house or at a (public) charging station or charging point.

Energy demand forecasts or range forecasts can be created for such vehicles in order, for example, to inform a driver whether a destination can be reached with the available energy supply. For the energy demand forecast or range forecast, an average speed per road section and individual acceleration maneuvers (junction, right-of-way signs, etc.) are generally used in order to derive the energy demand or the range. For the energy demand forecast, a provision of characteristic maps or non-parametric probability distributions from fleet data is required in the vehicle for data protection reasons. In this way, individual driver profile data of the vehicle driver can be combined with the fleet-based characteristic maps to calculate the energy demand forecast without external data storage in the vehicle. The transfer of the fleet data to the vehicle is complex and / or expensive, since a large amount of data has to be transferred to the vehicle in order to transfer a large number of multi-dimensional characteristic maps, such as non-parametric probability distributions. Due to bandwidth limitations, these distributions can not or not always be made available quickly in the vehicle.

It is an object of the present invention to improve the transmission of fleet data to a vehicle for an energy demand forecast of the vehicle while maintaining a high level of precision and reliability of the energy demand forecast. In particular, it is an object of the invention to provide a method for transmitting fleet data to a vehicle for an energy demand forecast of the vehicle and a system for transmitting fleet data to a vehicle for an energy demand forecast of the vehicle, which provides a reliable supply of current fleet data even with bandwidth limitations in the vehicle simple and inexpensive way to enable.

This problem is solved by the subject matter of the independent claims. Advantageous refinements of the invention are specified in the subclaims.

In the computer-implemented method for the transmission of fleet data to a vehicle for an energy demand forecast of the vehicle, one-dimensional or multi-dimensional first characteristic maps, each of which is a probability distribution of at least one physical dynamic parameter, which is preferably specific for a period of time, for example 0:00 to 0:30 o'clock, is, for a single segment of a route section to be traveled by the vehicle, grouped from the fleet data into several groups. A second characteristic map, which is selected from the first characteristic maps of each group or is calculated with the aid of the first characteristic maps of each group, is assigned as a representative to one of the groups, the representatives forming a code book outside the vehicle. An assignment of each first characteristic map to one of the representatives outside the vehicle is formed. The code book is transmitted to the vehicle in a first transmission step and at least one of the assignments is transmitted to the vehicle in a second transmission step. In the vehicle, the second characteristic map, which is stored as a representative in the assignment, is selected from the code book in the vehicle in order to generate the energy demand forecast for the vehicle.

Instead of the first characteristic maps, the second characteristic maps are used in the vehicle to forecast the energy demand. In contrast to the transmission of a total amount of data for the first characteristic maps, a reduced amount of transmitted data for the second characteristic maps is transmitted into the vehicle. By transferring only the representatives of the groups formed to the first characteristic maps in the form of the second characteristic maps, the amount of data of the transferred characteristic maps is reduced into the vehicle by the factor that results from the ratio of all groups to all first characteristic maps. If, for example, there are an average of twenty first characteristic maps in a group, the amount of data is reduced by a factor of twenty compared to the transmission of all the first characteristic maps. In addition, the second characteristic maps required for the respective route are selected with a minimal amount of data to be transmitted into the vehicle, since instead of the first characteristic maps only the assignments of these are made first feature maps to the second maps takes place as representatives of the group in which the respective first feature map was grouped.

The transmission of the code book with the representatives of the groups in the form of a second characteristic map per group can take place asynchronously to the transmission of the assignments into the vehicle. For example, the first step of transferring the code book into the vehicle can take place four times a year with, for example, 81 representatives with five physical dynamic parameters and, for example, 27 float values that describe the probability distributions of the dynamic parameters if the probability distributions are non-parametric, which results in an annual data volume of approximately 0.35 Megabyte leads. A float is a data type. The data type includes real numbers in the language of computers, i.e. numbers with decimal places. In contrast to float values, integer values have no decimal places and are therefore whole numbers. The second transfer step of transferring the assignments into the vehicle can take place, for example, for a distance traveled annually of approximately 9200 km with approximately 195,000 segments, each with an integer value, which is associated with a data volume of approximately 8.7 megabytes per year is. These amounts of data can be transmitted safely and reliably even with bandwidth restrictions depending on the field strength coverage of the radio network used for transmission, weather, trips abroad, etc.

Instead of an exclusively generic, loss-free transmission of all fleet data in the form of the first characteristic maps without data compression, there is a two-part, compressed transmission of the fleet data in the form of the second characteristic maps into the vehicle. In the vehicle, the second characteristic maps approximated with respect to the original first characteristic maps are available at any time and in updated form. With the second characteristic maps selected as required on the basis of the selection made in the second transfer step with the transferred assignments, the creation or calculation of the energy demand analysis is carried out, which includes providing a route section that comprises a plurality of segments or which is divided into a plurality of segments . The at least one first segment is assigned at least one physical dynamic parameter (or a value or a value distribution of the at least one physical dynamic parameter), which was determined based on data from a multiplicity of fleet vehicles. The energy requirement of the vehicle is calculated for the route section by adding up the energy requirements of the plurality of segments, the energy requirement of a respective segment being calculated using the at least one physical dynamic parameter assigned to said segment. The large number of values of the physical dynamic parameters distributed over the length of the route section flow directly into the calculation of the energy requirement. For this purpose, an energy requirement is calculated for each segment. The (total) energy requirement of the route section is determined by adding up the energy requirements of the segments. This enables a physically correct quantification of the route section for an energy demand or range forecast. In addition, through the physically correct quantification and the statistical recording of the dynamic parameter scatter, the driver's influence on the energy demand forecast can be quantified and optimized.

The dynamic parameters or integrals of the dynamic parameters for the at least one segment based on the fleet data can be collected and stored in a backend outside the vehicle. The dynamic parameters or integrals of the dynamic parameters for the at least one segment based on the fleet data are transmitted to the vehicle, for example during manufacture and / or online during operation of the vehicle. In this way, the navigation system can be regularly updated with current fleet data. The fleet dynamics can be requested online for a certain route section at the backend if the route section is to be used. The vehicle can call up and use the received dynamic parameters or integrals of the dynamic parameters for calculating the energy requirement.

Typically, the fleet vehicles are vehicles of the same and / or a similar vehicle type. In particular, the fleet can contain vehicles of the same and / or similar type as the vehicle for which the energy requirement for the route section is calculated (“own vehicle”). The fleet can in particular comprise a large number of third-party vehicles and optionally the own vehicle.

The at least one physical dynamic parameter is preferably selected from the group which comprises a speed and an acceleration or which consists of these. However, the invention is not limited thereto and other suitable physical dynamic parameters can be used. More than two physical dynamic parameters can be used to calculate the energy requirements of the vehicle. For example, a total of five different physical dynamic parameters can be used to calculate the energy requirement. A corresponding route length and a corresponding period of time are preferably assigned to each segment of a multiplicity of segments (“space-time bucket”). The route length can be, for example, a predetermined route interval (for example 50 meters). The time period can be, for example, a predetermined time interval (for example 30 minutes) and / or a time of day. However, the invention is not limited to this, and the route lengths and time periods can be determined in other suitable ways. The segments can have the same route lengths and / or time periods, or can at least partially have different route lengths and / or time periods.

The method preferably further comprises using a respective probability distribution of the at least one physical dynamic parameter, also called dynamic parameter distribution (DPD), for the individual segments of the route section when calculating the energy requirement. For example, at least that value of the segment which has the highest probability can be used for calculating the energy requirement of the individual segment. In other words, there can be a multiplicity of values for a segment, and in particular for all segments. Scattering the values within the individual segments allows a statement to be made about dynamic behavior, for example the acceleration and / or speed, and thus the energy prognosis in the segment and on the route section. This enables a high level of reliability of the range forecast to be achieved.

Alternatively, so-called clusters can be defined or present for the at least one segment if two or more physical dynamic parameters are used. In particular, a corresponding cluster can be assigned to each segment. A cluster consists of sets of the two or more physical dynamic parameters.

The term vehicle includes cars, trucks, buses, mobile homes, motorcycles, etc., which are used to transport people, goods, etc. In particular, the term includes motor vehicles for passenger transport. Additionally or alternatively, the hybrid or electric vehicle according to embodiments can be a purely electric vehicle (BEV) or a plug-in hybrid vehicle (PHEV). However, other forms of drive can also be used, for example in the form of a diesel or gasoline-powered vehicle. The vehicle can also be in the form of a rail vehicle

With the method according to the invention, a reliable transmission of current fleet data therefore takes place in a simple and inexpensive manner for an energy and range prognosis of the vehicle, even with bandwidth limitations in the vehicle, for example a hybrid or electric vehicle.

According to a preferred embodiment of the invention, the first characteristic maps are grouped into the groups using at least one clustering algorithm, the clustering algorithm preferably as a k-means algorithm, self-organizing map (SOM) or bisecting k-means algorithm is performed. An error due to the compression in the form of a deviation of the second characteristic map used in the vehicle for the energy demand forecast from the original first characteristic map, which was replaced by the second characteristic map, can be efficiently kept low when using one of these clustering algorithms.

Particularly preferably, the first feature maps are grouped into the groups using the k-means algorithm, the second feature maps are assigned to the groups as representatives, and the assignment of each first feature map to one of the representatives is formed. After dividing the groups and assigning the second characteristic maps in each case as representatives to the groups, the assignment of each first characteristic map to one of the representatives can thus also take place in an efficient manner with the aid of the k-means algorithm. The use of the k-means algorithm for all method steps outside the vehicle before the transmission steps simplifies the method according to the invention and leads to a slight deviation of the second method in the vehicle Energy demand prognosis feature map used from the original first feature map in such a way that the energy demand prognosis can be created with a high degree of precision.

In a preferred embodiment, a distance measure is used to define a distance between percentiles of two probability distributions of different characteristic maps in order to execute the clustering algorithm, to calculate the representatives for the groups from the first characteristic maps, and to assign the first characteristic map to one of the representatives, with the distance measure is determined as a second water stone metric, first water stone metric, Jensen-Shannon metric or Euclidean distance. In this way, a slight deviation of the second characteristic map used in the vehicle for the energy demand prognosis from the original first characteristic map can be achieved in order to be able to carry out the energy demand prognosis with a high degree of precision.

An advantageous embodiment is when the first and second characteristic maps are present in a recursion in the form of dynamic parameter distributions (DPD), the distance measure is specified, the first characteristic maps are related to one another on the basis of the distance measure in order to group them into the groups, the second feature maps are determined as representatives, the assignment of the first feature maps to the representatives takes place, and the thus determined representatives and first feature maps are related to one another on the basis of the distance measure in order to group the groups, determine the representatives, and the assign the first characteristic maps to the representatives until a change in the thus determined representatives to one another is below a threshold value. In this way, a set of representatives used in the vehicle for the energy demand forecast can be determined efficiently, each of which has a slight deviation from the original first characteristic map.

The second characteristic map is advantageously calculated by averaging all the first characteristic maps of the group which is represented by the second characteristic map. A representative calculated in this way can be generated by a k-means algorithm. Other clustering algorithms can also be used for averaging all the first characteristic maps of the group.

Advantageously, as already stated above, the first transfer step takes place asynchronously to the second transfer step, preferably with the first transfer step once to four times a year and / or the second transfer step when starting every journey or more than 30 times a year, in particular 36 times a year. In this way, an up-to-date and needs-based transmission of the characteristic maps required for the energy demand forecast into the vehicle is possible with only a small amount of data to be transmitted.

The invention also comprises a software program which is set up to carry out the computer-implemented method according to one of the preceding claims when it is carried out on a computer, the computer preferably being a distributed computer system, a part of which is preferably arranged in a cloud computer system . The software program can be set up to be executed on one or more processors, and thereby to execute the method according to the invention. The software program can be stored on one or more storage media.

The invention also comprises a system for transmitting fleet data to a vehicle for an energy demand forecast of the vehicle. It comprises a grouping unit which is set up to group one- or multi-dimensional first characteristic maps, each of which contains a probability distribution of at least one physical dynamic parameter, which is preferably specific for a period of time, for example 0:00 to 0:30, for a single segment comprise a route section to be traveled by the vehicle, from the fleet data into several groups. It further comprises an assignment unit which is set up to assign a second characteristic map, which is selected from the first characteristic maps of each group or is calculated with the aid of the first characteristic maps of each group, as a representative to one of the groups, the representatives having a code book outside the vehicle form, and an allocation unit which is set up to form an allocation of each first characteristic map in each case to one of the representatives outside the vehicle. Also included is a first transmission unit, which is configured to transmit the code book to the vehicle in a first transmission step, and a second transmission unit, which is configured to transmit at least one of the assignments to the vehicle in a second transmission step, and a selection unit, which is configured is for selecting the second characteristic map, which is stored as a representative in the assignment, from the code book in the vehicle for creating the energy demand forecast of the vehicle.

The system according to the invention has advantages and effects corresponding to the method according to the invention. The first and second transmission units can be present as separate units or as an integrated unit, for example a radio unit.

The grouping unit, the assignment unit, and the assignment unit can be combined in a back-end server and / or the selection unit can be comprised by an electronic control unit (ECU) of the vehicle.

• 1 eine schematische Darstellung eines Segments eines Streckenabschnitts, für den eine Energiebedarfsprognose erstellt wird gemäß dem Stand der Technik,
• 2 eine schematische Wahrscheinlichkeitsverteilung eines Dynamikparameters für ein Segment eines Streckenabschnitts, für den eine Energiebedarfsprognose erstellt wird, gemäß dem Stand der Technik,
• 3 ein Schema für eine Übertragung von Flottendaten an ein Fahrzeug für eine Energiebedarfsprognose des Fahrzeugs gemäß dem Stand der Technik,
• 4 ein Schema für eine Übertragung von Flottendaten an ein Fahrzeug für eine Energiebedarfsprognose des Fahrzeugs gemäß einer ersten Ausführungsform der Erfindung,
• 5 ein Schema für eine Übertragung von Flottendaten an ein Fahrzeug für eine Energiebedarfsprognose des Fahrzeugs gemäß einer zweiten Ausführungsform der Erfindung,
• 6 schematisch Merkmalskennfelder, deren Gruppierung in Gruppen, und Repräsentanten der Gruppen, wobei unter Verwendung eines k-means-Algorithmus die Merkmalskennfelder in die Gruppen gruppiert und die Repräsentanten den Gruppen zugewiesen sind, gemäß einer weiteren Ausführungsform der Erfindung.
• 7 ein Ablaufschema für eine Rekursion gemäß einer weiteren Ausführungsform der Erfindung, bei der Repräsentanten und Merkmalskennfelder anhand eines Abstandsmaßes jeweils solange zueinander in Beziehung gesetzt werden, um Gruppen zu gruppieren, die Repräsentanten zu bestimmen, und die Merkmalskennfelder den Repräsentanten zuzuordnen, bis eine Änderung der so bestimmten Repräsentanten unterhalb eines Schwellwertes liegt.

Exemplary embodiments of the disclosure are shown in the figures and are described in more detail below. Show it:

• 1 a schematic representation of a segment of a route section for which an energy demand forecast is made according to the prior art,
• 2 a schematic probability distribution of a dynamic parameter for a segment of a route section for which an energy demand forecast is created, according to the prior art,
• 3rd a scheme for a transmission of fleet data to a vehicle for an energy demand forecast of the vehicle according to the prior art,
• 4th a scheme for a transmission of fleet data to a vehicle for an energy demand forecast of the vehicle according to a first embodiment of the invention,
• 5 a scheme for a transmission of fleet data to a vehicle for an energy demand forecast of the vehicle according to a second embodiment of the invention,
• 6th schematically feature maps, their grouping in groups, and representatives of the groups, the feature maps being grouped into the groups and the representatives assigned to the groups using a k-means algorithm, according to a further embodiment of the invention.
• 7th a flowchart for a recursion according to a further embodiment of the invention, in which representatives and characteristic maps are related to one another based on a distance measure in order to group groups, determine the representatives, and assign the characteristic maps to the representatives until a change of the so certain representatives is below a threshold value.

Unless otherwise noted, the same reference symbols are used below for elements that are the same and have the same effect.

1 shows a schematic representation of a segment 2 of a route section that is to be traversed by the vehicle 1, for which an energy demand forecast is created. In the 1 a segment n is shown as an example.

wobei v die Geschwindigkeit im Segment 2 des Streckenabschnitts, a die Beschleunigung im Segment 2 und ds das Wegintegral über das Segment 2 angibt. Damit können die Werte bzw. Dynamikinformationen über die Beschleunigung a und die Geschwindigkeit v (z.B. quadrierte Signal und/oder ein Mittel) über einen Streckenabschnitt direkt in die physikalische Energiebedarfsrechnung eingesetzt werden.The energy demand for a segment 2 is calculated using at least one, and in particular each of the following expressions: $${\left[{\displaystyle \int v^{2}ds}\right]}_{ZuG},{\left[{\displaystyle \int v^{2}ds}\right]}_{\mathrm{S.}cHub},{\left[{\displaystyle \int ads}\right]}_{ZuG},{\left[{\displaystyle \int ads}\right]}_{\mathrm{S.}cHub},\overline{v},$$

where v is the speed in segment 2 of the route section, a is the acceleration in segment 2 and ds is the path integral over segment 2. The values or dynamic information about the acceleration a and the speed v (eg squared signal and / or a mean) can thus be used directly in the physical energy requirement calculation over a section of the route.

Beispielsweise kann der Energiebedarf für den Streckenschnitt unter Verwendung der folgenden Formel berechnet werden: $$\begin{array}{l}{E}_{Speicher}=\hfill \\ {\displaystyle \sum _{Start}^{Ziel}[\frac{1}{{\eta }_{Zug}}{\displaystyle \int [m\cdot e\cdot a\left(s\right)}+f_R\cdot m\cdot g+\frac{\rho \cdot c_x\cdot A}{2}\cdot v^2\left(s\right)+m\cdot g}\hfill \\ \cdot sin{\left(a\left(s\right)\right)]}_{a\ge a_{Grenz}}ds\hfill \\ +{\eta }_{Schub}{\displaystyle \int [m\cdot e\cdot a\left(s\right)}+f_R\cdot m\cdot g+\frac{\rho \cdot c_x\cdot A}{2}\cdot v^2\left(s\right)+m\cdot g\hfill \\ \cdot sin{\left(a\left(s\right)\right)]}_{a\ge a_{Grenz}}ds+P_{NV,n}\cdot {\Delta t_n]}_n\hfill \end{array}$$

• η_Zug: Wirkungsgrad zwischen einem Rad des Fahrzeugs 1 und einem Hochvoltspeicher (HVS) des Fahrzeugs 1 für Energiefluss von HVS zu dem Rad
• m: Fahrzeuggesamtmasse
• e: Rotationsmassenzuschlagsfaktor
• s: Weg
• a(s): Beschleunigung über den Weg
• f_R: Rollwiderstandskoeffizient
• g: Erdbeschleunigung
• ρ: Dichte der Luft
• c_x: Luftwiderstandskoeffizient
• A: Stirnfläche des Fahrzeugs 1
• v(s): Geschwindigkeit über den Weg
• α(s): Steigung über den Weg
• η_Schub: Wirkungsgrad zwischen Rad und Hochvoltspeicher (HVS) für Energiefluss von Rad zum HVS
• P_NV,n: Nebenverbraucherleistung im Segment n
• 
  Benötige Zeit zum Durchfahren des Segmentes n

The expressions are separated into recuperation / overrun route and pull route. The separation can depend on a limit acceleration: $$a_{Grenz}\left(s\right)\forall \left({\mathrm{F.}}_{\mathrm{R.}ad}=0\right)=-\frac{f_{\mathrm{R.}}\cdot G}{e}-\frac{\rho \cdot c_x\cdot \mathrm{A.}}{2\cdot m\cdot e}\cdot v^2\left(s\right)-\frac{G\cdot \text{sin}\left(a\left(s\right)\right)}{e}$$

For example, the energy requirement for the route cut can be calculated using the following formula: $$\begin{array}{l}{\mathrm{E.}}_{\mathrm{S.}peicHer}=\hfill \\ {\displaystyle \sum _{\mathrm{S.}tart}^{Ziel}[\frac{1}{{\eta }_{ZuG}}{\displaystyle \int [m\cdot e\cdot a\left(s\right)}+f_{\mathrm{R.}}\cdot m\cdot G+\frac{\rho \cdot c_x\cdot \mathrm{A.}}{2}\cdot v^2\left(s\right)+m\cdot G}\hfill \\ \cdot sin{\left(a\left(s\right)\right)]}_{a\ge a_{Grenz}}ds\hfill \\ +{\eta }_{\mathrm{S.}cHub}{\displaystyle \int [m\cdot e\cdot a\left(s\right)}+f_{\mathrm{R.}}\cdot m\cdot G+\frac{\rho \cdot c_x\cdot \mathrm{A.}}{2}\cdot v^2\left(s\right)+m\cdot G\hfill \\ \cdot sin{\left(a\left(s\right)\right)]}_{a\ge a_{Grenz}}ds+{\mathrm{P.}}_{NV,n}\cdot {\Delta t_n]}_n\hfill \end{array}$$

• η _train : efficiency between a wheel of vehicle 1 and a high-voltage storage device (HVS) of vehicle 1 for energy flow from HVS to the wheel
• m: total vehicle mass
• e: Rotational mass addition factor
• s: way
• a (s): Acceleration over the path
• f _R : coefficient of rolling resistance
• g: acceleration due to gravity
• ρ: density of the air
• c _x : drag coefficient
• A: Frontal area of the vehicle 1
• v (s): speed on the way
• α (s): slope over the path
• η _thrust : efficiency between wheel and high-voltage storage (HVS) for energy flow from wheel to HVS
• P _{NV, n} : Ancillary consumer power in segment n
• 
  Time required to drive through segment n

Typically they are fleet vehicles 10 Vehicles of the same or a similar type of vehicle. In particular, the fleet can accommodate a wide variety of vehicles 10 of the same and / or similar type as the vehicle 1 , for which the energy requirement for the route section is calculated ("own vehicle"). The fleet can in particular include a large number of third-party vehicles 10 and optionally your own vehicle 1 include.

2 shows an exemplary probability distribution 12th the values of a dynamic parameter for the segment n, and in particular Δt, [∫v ² ds; ∫a ds] Pull _{/ push} .

A respective probability distribution, that is to say a scatter, of the values of the at least one physical dynamic parameter can be calculated for each segment n of the route section. For example, only that value of the segment that has the highest probability can be used to calculate the energy requirement of the individual segment. In another example, a threshold value can be established for the probability, with those values of the at least one physical dynamic parameter being used for calculating the energy requirement which have a probability that is equal to or greater than the threshold value. In this way, the reliability of the energy forecast for the segment and thus the entire route section can be increased.

A location of a driver model can be used in the probability distribution 12th of the segments by matching the probability distribution 12th can be determined with at least one value of the at least one driver-specific dynamic selection parameter. For example, a driver-specific acceleration with the probability distribution 12th of accelerations of the fleet vehicles 10 be compared to the location of the driver-specific acceleration in the probability distribution 12th of accelerations of the fleet vehicles 10 to determine. Similarly, a driver-specific speed with the Probability distribution 12th of the speeds of the fleet vehicles 10 be compared to the location of the driver-specific speed in the probability distribution 12th of the speeds of the fleet vehicles 10 to determine.

If the fleet dynamics parameters are the probability distribution 12th several different values are available for calculating the energy requirement. By including the driver-specific dynamic selection parameters, the correct value for calculating the energy requirement for the route section can be selected. In particular, the calculation of the individual driver's energy requirement of the vehicle can be based on the location of the driver model in the probability distributions 12th of the segments n. An energy requirement can thus be calculated with high precision. The learned driver behavior can thus be correlated with the fleet dynamics, whereby a position of the driver behavior in the probability distribution 12th determined and from this the energy prognosis, which is individual for the driver, can be derived. A savings potential or optimized energy consumption can then be determined, for example. Based on the determined optimized energy consumption, a recommendation for action can be output which is suitable to achieve the optimized energy consumption, such as a reduced travel speed.

3rd shows a scheme for a transfer of fleet data 11 to the vehicle 1 for an energy demand forecast of the vehicle 1 . From the fleet data 11 , the driving profiles of the fleet vehicles 10 for a large number of route sections and their segments are first characteristic maps 12th in the form of probability distributions of at least one physical dynamic parameter, for example the speed v and / or the acceleration a, outside the vehicle 1 , for example in a backend server 5 , calculated and in a memory 4th saved by the backend server 5 can be included. The ones in memory 4th stored first characteristic maps 12th are, for example, transferred to the vehicle for a distance traveled annually of approximately 9200 km with approximately 195,000 segments, each with five physical dynamic parameters and 27 float values, which is associated with a data volume of approximately 8.6 gigabytes per year.

In the vehicle 1 becomes a driver model / driver profile 13th or a classification of the driver dynamics in the fleet dynamics for the learning of the driver model. For example, for a route section n traveled through and all segments of this route section, the speed traveled and the acceleration that occurred are used to create the driver profile 13th in the vehicle 1 recorded. The driver profile-related data is subject to data protection and should not be stored outside the vehicle, even if storage outside the vehicle is possible and / or permitted. For each segment that has not yet been passed through 2 of the route section 3rd the driver dynamics are predicted based on the dynamic parameters of the fleet dynamics and the learned driver dynamics. The dynamic parameters, namely the speed v and the acceleration a, of both dynamics go into the energy requirement calculation 14th a, to a driver-specific energy consumption E for the route section 3rd to determine. A driver-specific determination of dynamic information takes place for a driver-specific energy consumption forecast. The driver's influence on the energy consumption for a range forecast can thus be precisely quantified, whereby a high level of precision in the range forecast can be achieved.

4th shows a scheme for a transfer of fleet data 11 to the vehicle 1 for an energy demand forecast of the vehicle 1 according to the invention. As in 3rd are shown from the fleet data 11 , the driving profiles of the fleet vehicles 10 for a large number of route sections and their segments comprise first characteristic maps 12th in the form of probability distributions of at least one physical dynamic parameter, for example the speed v and / or the acceleration a, outside the vehicle 1 , for example in the backend server 5 , calculated and in the memory 4th saved. The memory 4th can be part of the backend server 5 be or through a distributed computer system (not shown) with or without the backend server 5 outside the vehicle 1 be educated.

Grouping takes place in the backend server 20th of the first characteristic maps 12th which each have a probability distribution of at least one physical dynamic parameter, for example the speed v and / or the acceleration a, for an individual segment 2 of the vehicle 1 section to be traveled through 3rd include, from the fleet data 11 into several groups in one grouping unit 21 of the backend server 5 instead of. This is followed by the assignment 22 in each case of a second characteristic map that is derived from the first characteristic maps 12th selected for each group or with the aid of the first characteristic maps 12th each group in the backend server 5 is calculated as a representative R1 , R2 to one of the groups in an assignment unit assigned by the grouping unit 21 may be included, the representatives R1 , R2 a code book 23 outside the vehicle 1 form. Formation then takes place 24 an assignment 26th of each first characteristic map 12th each to one of the representatives R1 , R2 in an allocation unit 25th of the backend server 5 , i.e. outside the vehicle 1 , and transferring the codebook 23 to the vehicle 1 in a first transfer step 27 and from those assignments 26th , which are the first characteristic maps 12th of the segments 2 concern those for calculating the energy requirement 14th required to the vehicle 1 in a second transfer step 28 . The transfer steps 27 , 28 can be done, for example, by radio according to the GSM / 3G, 4G and / or 5G standard, or via a WLAN network in one or separate transmission units. There is a selection in the vehicle 29 of the second characteristic map, which is used as a representative R1 , R2 in the assignment 26th is stored from the code book 23 in a selection unit 30th instead of the energy demand forecast of the vehicle 1 based on the energy requirement calculation 14th to create.

The first transfer step 27 of the code book 23 into the vehicle 1 takes place according to the in 4th The exemplary embodiment shown four times a year with 81 representatives with five physical dynamic parameters and 27 float values, which leads to an annual data volume of approximately 0.35 megabytes. The second transfer step 28 the transfer of the assignments 26th into the vehicle 1 takes place in the exemplary embodiment shown for a distance traveled annually of approximately 9200 km with approximately 195,000 segments, each with an integer value, that is, an integer value, which is associated with a data volume of approximately 8.7 megabytes per year. Therefore, a total of only approximately 9.05 megabytes are required to transmit the second characteristic maps to the vehicle each year 1 transfer. Compared to the amount of data for transferring the first characteristic maps 12th to the vehicle 1 of about 8.6 gigabytes per year, as shown in 3rd is shown, there is a reduction in the annual amount of data transmitted for the preparation of the energy demand forecast of approximately 1000.

In the vehicle 1 becomes the driver model / driver profile 13th or a classification of the driver dynamics in the fleet dynamics for the learning of the driver model. The driver profile-related data are subject to data protection and are not transferred outside of the vehicle 1 saved. For each segment that has not yet been passed through 2 of the route section 3rd finds a prediction of the driver dynamics based on the fleet dynamics on the basis of the assignments 26th selected second characteristic maps and the learned driver dynamics of the driver profile 13th instead of. The dynamic parameters of both dynamics go into the energy requirement calculation 14th a, to a driver-specific energy consumption E for the route section 3rd to determine.

5 shows a scheme for a transmission of fleet data to a vehicle for an energy demand forecast of the vehicle according to a further embodiment of the invention. The scheme includes a system for transferring the fleet data 11 to the vehicle 1 for the energy demand forecast 15th of the vehicle 1 . The system includes the grouping unit 21 that is set up for grouping 20th the one or more dimensional, in the memory 4th of the backend server 5 stored first characteristic maps 12th that each have a probability distribution of at least one physical dynamic parameter for the individual segment 2 of the vehicle 1 section to be traveled through 3rd include, from the fleet data 11 in several groups. The system further comprises an assignment unit which is set up for assignment 22nd each of a second characteristic map 31 , that from the first characteristic maps 12th selected for each group or with the aid of the first characteristic maps 12th each group is charged as a representative R1 , R2 to one of the groups, with the representatives R1 , R2 a code book 23 outside the vehicle 1 in the backend server 5 form. The system also includes the allocation unit 25th that is set up for education 24 the assignment 26th of each first characteristic map 12th in each case to one of the representatives, the first transmission unit which is set up to transmit the code book 23 to the vehicle 1 in the first transfer step 27 and a second transmission unit which is set up to transmit those assignments 26th , which are the first characteristic maps 12th of the segments 2 concern that for the energy demand forecast 15th required to the vehicle 1 in the second transfer step 28 .

Finally, the system includes the selection unit 30th that is set up to select 29 of the second characteristic map 31 , that as a representative in the respective assignment 26th is stored from the code book 23 in the vehicle 1 for energy demand calculation 14th taking into account the driver profile 13th to the energy demand forecast 15th of the vehicle 1 to obtain. The selection of the second characteristic map 31 and the energy demand calculation 14th taking into account the driver profile 13th takes place in an electronic control unit 6th , Electronic Control Unit ECU, inside the vehicle 1 instead of. So that is the selection unit 30th to choose 29 of the second characteristic map 31 in the ECU 6th arranged.

6th shows schematically first feature maps 12th , grouping them into groups G1 , G2 , and representatives R1 , R2 of the groups G1 , G2 , using the illustrated k-means algorithm as the objective function 34 the characteristic maps 12th in the groups G1 , G2 grouped and the representatives R1 , R2 the groups G1 , G2 are assigned, according to a further embodiment of the invention. A second characteristic map in each case 31 , 32 is made by means of the k-means algorithm 34 with the aid of the first characteristic maps 12th each group G1 , G2 calculated and as a representative R1 , R2 one of the groups G1 , G2 assigned.

The aim of k-Means is in the in 6th embodiment of the invention shown, the data record on first characteristic maps 12th so divide into k groups that the sum of the squared deviations as a distance function 35 of the second characteristic maps 31 , 32 as group priorities R1 , R2 is minimal. Mathematically, this corresponds to the optimization of the objective function 34 with the data points x _i in the form of the first characteristic maps 12th and the group focal points c _j in the form of the first characteristic maps 31 , 32 of groups j in the form of groups G1 , G2 . The assignment 26th of each first characteristic map 12th each to one of the representatives R1 , R2 also takes place by means of the k-means algorithm 34.

After drawing the boundary between the first characteristic maps 12th to form the groups G1 , G2 (see limits between the first characteristic maps 12th in the form of lines in 6th ) the assignment takes place 26th of each first characteristic map 12th each to one of the representatives R1 , R2 by assigning the respective first characteristic map 12th to his group G1 , G2 . As an alternative or in addition to the k-means algorithm, other algorithms, such as the self-organizing map (SOM) or the bisecting k-means algorithm, can be used to determine the groups G1 , G2 , the representative R1 , R2 and the assignments 26th be used.

7th shows a flow chart for a recursion 43 according to a further embodiment of the invention, in which the representatives R1 , R2 and the first feature maps 12th based on a distance measure 45 in each case as long as they are related to the groups G1 , G2 to group the representatives R1 , R2 to be determined, and the second feature maps 31 , 32 the representative R1 , R2 to be assigned until a change in the representative determined in this way R1 , R2 is below a threshold value. In the recursion 43 the first 12 and second 31, 32 characteristic maps are available in the form of dynamic parameter distributions (DPD), 42. For this purpose, a fleet of fleet vehicles 10 Driving profiles on, 39, which are sent to a backend, for example in the form of the backend server 5 , are sent, 40. In the backend, the driving profile parameters, driver profile parameters DP, are integrated and stored for each map segment, 41 in order to provide the distributions of the dynamic parameters, dynamic parameter distributions DPD, 42.

The distance measure 45 is used to define a distance, for example, of percentiles of two probability distributions of different characteristic maps 12th , 31 , 32 to use the distance function, for example 35 of the k-means algorithm (see 6th ), and is specified, 46. After an initialization 44 are based on the distance 45 the first characteristic maps 12th related to each other, 47 to divide them into groups G1 , G2 to group the second characteristic maps 31 , 32 recalculate as a representative, 48, and the assignment of the first characteristic maps 12th to the representatives R1 , R2 to undertake, 49. The representatives determined in this way R1 , R2 and first feature maps 12th are based on the distance 45 each as long as related to each other, 50.47 to the groups G1 , G2 to group the representatives R1 , R2 to be determined, and the first characteristic maps 12th the representative R1 , R2 to be assigned until a change in the representatives determined in this way from one another between two repetitions is below a predetermined threshold value. After that, the final representatives R1 , R2 in the code book 23 entered, 51, and the final assignment of the first characteristic maps is found 12th to the final representatives R1 , R2 instead, 52.

Although the invention has been illustrated and explained in more detail by preferred exemplary embodiments, the invention is not restricted by the disclosed examples and other variations can be derived from them by the person skilled in the art without departing from the scope of protection of the invention. It is therefore clear that there is a multitude of possible variations. Embodiments cited by way of example only represent examples that are not to be interpreted in any way as a limitation, for example, of the scope of protection, the possible applications or the configuration of the invention. Rather, the preceding description and the description of the figures enable the person skilled in the art to specifically implement the exemplary embodiments, whereby the person skilled in the art, with knowledge of the disclosed inventive concept, can make various changes, for example with regard to the function or the arrangement of individual elements mentioned in an exemplary embodiment, without the To leave the scope of protection that is defined by the claims and their legal equivalents, such as further explanations in the description.

The features of the invention described with reference to the illustrated embodiments, for example the use of the k-means algorithm 34 for grouping the groups G1 , G2 and making representatives R1 , R2 , can also be used in other embodiments of the invention, e.g. B. the recursion 43 , whose flow chart is in the 7th shown and described with reference to this figure, unless otherwise stated or is prohibited for technical reasons.

Claims (10)

Computer-implemented method for the transmission of fleet data (11) to a vehicle (1) for an energy demand forecast (15) of the vehicle (1), comprising the steps: - Grouping (20) of one-dimensional or multi-dimensional first characteristic maps (12), each of which contains a probability distribution of at least one physical dynamic parameter, which is preferably specific for a period of time, for example 0:00 to 0:30, for a single segment (2nd ) comprise a route section (3) to be traveled by the vehicle (1), from the fleet data (11) into several groups (G1, G2), - Assigning (22) a second characteristic map (31, 32), which is selected from the first characteristic maps (12) of each group (G1, G2) or is calculated with the aid of the first characteristic maps (12) of each group (G1, G2), as a representative (R1, R2) to one of the groups (G1, G2), the representatives (R1, R2) forming a code book (23) outside the vehicle (1), - Forming (24) an assignment (26) of each first characteristic map (12) to one of the representatives (R1, R2) outside the vehicle (1), - Transferring the code book (23) to the vehicle (1) in a first transfer step (27) and of at least one of the assignments (26) to the vehicle (1) in a second transfer step (28), and - Selecting (29) the second characteristic map (31, 32), which is stored as a representative (R1, R2) in the assignment (26), from the code book (23) in the vehicle (1) to create the energy demand forecast (15) of the Vehicle (1). The computer-implemented method according to Claim 1 , the first feature maps (12) being grouped (20) into the groups (G1, G2) using at least one clustering algorithm, the clustering algorithm preferably as a k-means algorithm (34), self-organizing map (SOM) or bisecting k-means algorithm is executed. The computer-implemented method according to Claim 2 , the first feature maps (12) being grouped into the groups (20) using the k-means algorithm (34), the second feature maps (31, 32) each as representatives (R1, R2) of the groups (G1, G2 ) are assigned (22), and the assignment (26) of each first characteristic map (12) to one of the representatives (R1, R2) is formed (24). The computer-implemented method according to Claim 2 or Claim 3 , wherein a distance measure (45) for establishing a distance between percentiles of two probability distributions of different characteristic maps (12, 31, 32) is defined (46) in order to execute the clustering algorithm, from the first characteristic maps (12) the respective representatives (R1, R2) to calculate (22) for the groups (G1, G2), and to assign the first characteristic map (12) to one of the representatives (R1, R2) in each case (24), the distance measure (45) being the second water stone metric, first Wasserstein metric, Jensen-Shannon metric or Euclidean distance is determined. The computer-implemented method according to Claim 4 , the first (12) and second (31, 32) characteristic maps being present in a recursion (43) in the form of dynamic parameter distributions (DPD), the distance measure (45) being specified (46), the first based on the distance measure (45) Feature maps (12) are related to one another (47) in order to group them into groups (G1, G2) (20), the second feature maps (31, 32) are determined as representatives (R1, R2) (48), the assignment of the first characteristic maps (12) to the representatives (R1, R2) takes place (49), and the thus determined representatives (R1, R2) and first characteristic maps (12) are related to one another on the basis of the distance measure (45) to group (20) the groups (G1, G2), determine the representatives (R1, R2) (48), and assign (49) the first feature maps (12) to the representatives (R1, R2) until a change the representatives (R1, R2) determined in this way are below a threshold value in relation to one another. The computer-implemented method according to one of the preceding claims, wherein the second characteristic map (31, 32) is calculated by averaging all first characteristic maps (12) of the group (G1, G2) represented by the second characteristic map (31, 32) ( 22, 48). The computer-implemented method according to one of the preceding claims, wherein the first transmission step (27) takes place asynchronously to the second transmission step (28), preferably with the first transmission step (27) once to four times a year and / or the second transmission step (28) at each occurrence Trip or more than 30 times a year, especially 36 times a year. Software program which is set up to carry out the computer-implemented method according to one of the preceding claims when it is carried out on a computer, the computer preferably being a distributed computer system, a part of which is preferably arranged in a cloud computer system. System for the transmission of fleet data (11) to a vehicle (1) for an energy demand forecast (15) of the vehicle (1), comprising: - A grouping unit (21) which is set up to group (20) one- or multi-dimensional first characteristic maps (12), each of which contains a probability distribution of at least one physical dynamic parameter, which is preferably specific for a period of time, for example 0:00 to 0:30 Clock, is, for a single segment (2) of a route section (3) to be traveled by the vehicle (1), from the fleet data (11) in several groups (G1, G2), - An assignment unit which is set up to assign (22) a second characteristic map (31, 32), which is selected from the first characteristic maps (12) of each group (G1, G2) or with the aid of the first characteristic maps (12) of each group ( G1, G2) is calculated as a representative (R1, R2) for one of the groups (G1, G2), the representatives (R1, R2) forming a code book (23) outside the vehicle (1), - an assignment unit (25) which is set up to form (24) an assignment (26) of each first characteristic map (12) to one of the representatives (R1, R2) outside the vehicle (1), - a first transmission unit which is set up to transmit the code book (23) to the vehicle (1) in a first transmission step (27) and a second transmission unit which is set up to transmit at least one of the assignments (26) to the vehicle ( 1) in a second transfer step (28), and - A selection unit (30) which is set up to select (29) the second characteristic map (31, 32), which is stored as a representative (R1, R2) in the assignment (26), from the code book (23) in the vehicle for Creation of the energy demand forecast (15) of the vehicle (1). System according to Claim 9 , in which the grouping unit (21), the assignment unit, and the assignment unit (25) are combined in a backend server (5) and / or the selection unit (30) from an electronic control unit (6) (ECU) of the vehicle (1 ) is included.

Images