A METHOD FOR CONTROLLING A VEHICLE

Abstract

The invention relates to a method for controlling a vehicle (1), the method comprising—establishing (S1) a plurality of example situations, wherein each example situation is characterized by a plurality of example situation features, including at least a velocity of the vehicle (VS), a velocity change of the vehicle (VCS), and a road inclination (αS), —determining (S2), for each of plurality of the example situations, a example situation cost (CS) dependent on a cost for operating the vehicle in the respective example situation, —subsequently obtaining (S3) topology data, indicative of a topology of a route (RT) to be travelled by the vehicle, —determining (S4-S8, S41), based at least partly on the route topology, and at least a plurality of the example situation costs (CS), a velocity profile (VP) for the vehicle along the route (RT).

Claims

1. A method for controlling a vehicle, the method comprising, establishing a plurality of example situations, wherein each example situation is characterized by a plurality of example situation features, including at least a velocity of the vehicle in the example situation, a velocity change of the vehicle in the example situation, and a road inclination in the example situation, determining, for each of a plurality of the example situations, an example situation cost dependent on a cost for operating the vehicle in the respective example situation, subsequently obtaining topology data, indicative of a topology of a route to be travelled by the vehicle, determining, based at least partly on the route topology, and at least a plurality of the example situation costs, a velocity profile for the vehicle along the route.

2. A method according to claim 1, characterized in that the example situation costs are determined by means of a vehicle model.

3. A method according to claim 1, characterized by establishing an approximation function, based partly on at least a plurality of the example situations, and based partly on the example situation costs for the example situations on which the approximation function is partly based.

4. A method according to claim 3, characterized by establishing, in dependence on the topology data, a plurality of predicted situations, and determining, for each of at least a plurality of the established predicted situations, by means of the approximation function, a predicted situation cost indicative of a cost for operating the vehicle in the respective predicted situation, wherein the velocity profile is determined partly based on the determined predicted situation costs.

5. A method according to claim 1, characterized in that determining the velocity profile comprises establishing a sequence of positions along the route.

6. A method according to claim 5, characterized by associating with each of a plurality of, or all of, the positions, a respective altitude of the position, and/or a respective road inclination at the position.

7. A method according to claim 5, characterized by determining for each of a plurality of pairs of adjacent positions, a predicted situation cost for operating the vehicle from one to the other of the adjacent positions, based at last partly on a predicted vehicle velocity at the pair of adjacent positions, and a predicted velocity change from one to the other of the adjacent positions.

8. A method according to claim 7, characterized in that the predicted situation costs are determined by means of an approximation function established, based partly on at least a plurality of the example situations, and based partly on the example situation costs for the example situations on which the approximation function is partly based.

9. A method according to claim 7, characterized in that the velocity profile is determined partly based on the determined predicted situation costs.

10. A method according to claim 5, characterized in that determining the velocity profile comprises setting up a matrix for the route, with the sequence of route positions, and one or more vehicle velocity values for each position.

11. A method according to claim 5, characterized by determining, for a pair of adjacent positions in the sequence of positions, a vehicle velocity at one of the positions, and determining a plurality of candidate predicted velocity changes, each providing a respecting vehicle velocity at the other of the positions.

12. A method according to claim 11, characterized by determining, for each a plurality of, or all of, the candidate predicted velocity changes, a candidate predicted situation cost, based at least partly on at least one of the example situation costs.

13. A method according to claim 12, characterized in that at least a plurality of the candidate predicted situation costs are determined, by means of an approximation function established based partly on at least a plurality of the example situations, and based partly on the example situation costs for the example situations on which the approximation function is partly based.

14. A method according to claim 13, characterized in that the approximation function is established by an interpolation of two or more of the example situation costs.

15. A method according to claim 12, characterized by adjusting the candidate predicted situation costs, based at least partly on the time of driving between the adjacent positions.

16. A method according to claim 15, characterized by determining, subsequently to the adjustment of the example situation costs, and/or the candidate predicted situation costs, the velocity profile for the route, by selecting, for each pair of adjacent positions, one of the candidate predicted situation costs, such that the sum of the candidate predicted situation costs, selected throughout the route, are minimized.

17. A method according to claim 1, characterized in that the example situation costs are determined separately from the vehicle.

18. A method according to claim 1, characterized by establishing an approximation function, based partly on at least a plurality of the example situations, and based partly on the example situation costs for the example situations on which the approximation function is partly based, for determining predicted situation costs for respective predicted situations of the vehicle.

19. A method according to claim 18, characterized by storing the approximation function in the vehicle.

20. A method according to claim 18, characterized by determining an actual cost for operating the vehicle, and adjusting the approximation function in dependence on the determined actual cost.

21. A method according to claim 1, characterized by determining an actual cost for operating the vehicle, and adjusting an example situation cost in dependence on the determined actual cost.

22. A method according to claim 1, characterized in that the plurality of example situation features characterizing the example situations include a load of the vehicle.

23. A computer program comprising program code means for performing the steps of claim 1 when said program is run on a computer, or a group of computers.

24. A computer readable medium carrying a computer program comprising program code means for performing the steps of claim 1 when said program product is run on a computer, or a group of computers.

25. A control unit, or a group of control units, configured to perform the steps of the method according to claim 1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0062] With reference to the appended drawings, below follows a more detailed description of embodiments of the invention cited as examples.

[0063] In the drawings:

[0064] FIG. 1 is a schematic view of a vehicle and a stationary control unit.

[0065] FIG. 2 is a diagram depicting steps in a method, according to an embodiment of the invention, for determining a velocity profile of the vehicle in FIG. 1.

[0066] FIG. 3 shows a table with parameters used for the velocity profile determination.

[0067] FIG. 4 shows a contour of a route to be travelled by the vehicle.

[0068] FIG. 5 and FIG. 6 depict a matrix for determining the velocity profile of the vehicle.

[0069] FIG. 7 is a flow diagram depicting steps in a method according to a more general embodiment of the invention.

DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS OF THE INVENTION

[0070] FIG. 1 shows a vehicle 1. In this example, the vehicle 1 is a truck with a semitrailer. However, the invention is equally applicable to other types of vehicles, such as cars, buses, dump trucks, and mining vehicles.

[0071] The vehicle includes a powertrain. The powertrain includes a propulsion arrangement. Embodiments of the invention are applicable to a variety of propulsion arrangements. The propulsion arrangement may include an internal combustion engine. The vehicle may be arranged to be driven by an engine only. The propulsion arrangement may include an electric motor. The propulsion arrangement may be a hybrid arrangement with an engine and a motor. In this example, the vehicle is a fully electric vehicle.

[0072] The vehicle comprises a vehicle control unit 101, arrange to control functions of the vehicle, such as its propulsion, and braking. The control unit is arranged to control the propulsion arrangement. The control unit is arranged to control a braking system of the vehicle. The vehicle control unit 101 may be provided as a single physical unit, or as a plurality of physical units, arranged to communicate with each other.

[0073] The vehicle also comprises vehicle equipment for wireless communication 102. The vehicle control unit 101 is arranged to receive data via the vehicle communication equipment 102.

[0074] FIG. 1 also shows a stationary control unit 201. The stationary control unit 201 may be provided as a computer. The stationary control unit 201 is connected to stationary equipment for wireless communication 202. The stationary control unit 201 is arranged to send data via the stationary communication equipment 202.

[0075] With reference to FIG. 2, a method according to an embodiment of the invention will be described.

[0076] The method comprises establishing S1 a plurality of example situations for one or more vehicles. In this example, the example situations are determined by the stationary control unit 201. In this example, the example situations are described as referring to a single vehicle, or a single type of vehicle, such as a specific vehicle model. However, in some embodiments, some of the example situations may refer to one type of vehicle, while other example situations refer to one or more other types of vehicles.

[0077] Each example situation is characterized by a plurality of example situation features. The example situation features include a velocity of the vehicle, a velocity change of the vehicle, a road inclination, and a load of the vehicle. An example situation feature may represent a value of the velocity, the velocity change, the road inclination, or the load. The values of these parameters may be stepped from one example feature to another. Further, different parameter values may be distributed to different example situations. Each example situation may be characterized by a unique combination of example situation features.

[0078] The example situation features may also include a length of a segment of the route, e.g. expressed as a distance between two positions along the route.

[0079] The method further comprises determining S2, for each of a plurality of, or all of, the example situations, an example situation cost. The example situation cost is dependent on a cost for operating the vehicle in the respective example situation. As exemplified below, each example situation cost may also be dependent on the time of travelling, or the velocity, at the respective example situation.

[0080] The example situation costs are determined by means of a vehicle model. The vehicle model may include a model of the powertrain. The model may include energy losses of the powertrain. The vehicle model may include a model of the propulsion arrangement. Where the propulsion arrangement is an electric hybrid propulsion arrangement, or a fully electric propulsion arrangement, the vehicle model may include a model of the electric motor, and a model of an electric storage device, such as a battery, or a battery pack. The electric storage device model may include a model of the state of health of the storage device. The electric storage device model may include a battery degradation model. The storage device capability may be dependent on a state of charge of the storage device.

[0081] The vehicle model may further include a model of a braking system of the vehicle. The braking system may include service brakes. The braking system may include a function of regenerative braking by means of the motor and the storage device. Further, the vehicle model may include a model of a road friction. The method may aim to minimize the use of the service brakes. For this, example situations entailing, as indicated by the vehicle model, the use of service brakes, may be assigned a relatively high example situation cost.

[0082] From the velocity, the velocity change, and the road inclination of an example situation, a required force, torque, and/or power, of the powertrain, may be determined. The vehicle model may include constraints of the vehicle, or the powertrain. Such constraints may indicate whether or not a candidate for an example situation is feasible. For example, there might be a limit of a vehicle acceleration in dependence on the propulsion arrangement capacity, and the road inclination. Also, there might be a limit of a negative vehicle acceleration in dependence on a capacity of the braking system, or the road friction. In addition, there might be a limit of a negative vehicle acceleration in dependence on a negative road inclination. The example situation costs may be determined S2 only for feasible example situations.

[0083] As exemplified below, a route to be travelled by the vehicle may be divided into segments. The segments may be of different length, or of equal length. Where the segments are of equal length, the segments are herein also referred to as unit segments. Also, each example situation may involve a travel of the vehicle through a unit segment. Thereby, an example situation may be characterized partly by a segment length, a vehicle velocity at the entry of the unit segment, and a vehicle velocity at an exit of the segment, as well as an inclination of the segment.

[0084] The example situation costs are determined S2 by the stationary control unit 201. Thus, the example situation costs are determined separately from the vehicle. Each example situation cost may be a cost for operating the vehicle in the respective example situation. The cost for operating the vehicle may be dependent on the energy consumption of the vehicle at the respective example situation. The cost for operating the vehicle in the respective example situation is determined by means of the vehicle model.

[0085] Thus, each example situation, with its example situation features, is related with a respective of the example situation costs.

[0086] The method further comprises establishing, for determining predicted situation costs based on predicted situation features of respective predicted situations of the vehicle, an approximation function based on at least a plurality of the example situation costs CS, and values of the example situation features of the example situations for which the example situation costs were determined.

[0087] The approximation function is sent to the vehicle control unit 101. Alternatively, data representing the determined example situation features, and the example situation costs are sent to the vehicle control unit 101. The sending may be done by means of the communication equipment 102, 202. The approximation function is stored in the vehicle, in the vehicle control unit 101, or accessible thereto.

[0088] The approximation function is stored in the vehicle control unit 101. Alternatively, the example situation features, and the example situation costs, are stored in the vehicle, in the vehicle control unit 101, or accessible thereto. In the vehicle control unit 101, the example situation features, and the example situation costs, may be provided in the form of a table. FIG. 3 shows an example of such a table. The table may present, for each combination of a vehicle load LS, a road inclination αS, a vehicle velocity VS, and vehicle velocity change VCS, an example situation cost CS. As stated, the example situation features may also include a length of a segment of the route, not shown in FIG. 3.

[0089] Thus, for determining the velocity profile, the vehicle control unit 101 is not burdened with the computationally demanding work of determining the costs for operational situations. Instead, the vehicle control unit uses the approximation function, and/or the example situation features, and the example situation costs, as exemplified below.

[0090] The velocity profile may be determined while the vehicle is stationary, e.g. for a planned route. Alternatively, the velocity profile may be determined repetitively while the vehicle is travelling, for a route forming a part of a trip of the vehicle. The route may be determined, e.g. by means of a journey planning device, such as a GPS (Global Positioning System) unit. For this presentation, such a device may be understood as forming a part of the vehicle control unit 101. The determination of a velocity profile may be done at regular time intervals, or driving distance intervals, e.g. every 100 metres.

[0091] Reference is made also to FIG. 4. The method comprises obtaining S3 topology data, indicative of a topology of a route RT to be travelled by the vehicle. The topology data may be obtained from map data. A velocity profile for the vehicle along the route is determined, based at least partly on the route topology, and at least a plurality of the stored example situation costs CS. This determination comprises establishing S4 a sequence of positions P0-P9 along the route. Thereby, the route is divided into segments ST1-ST9. In this example, the segments are of equal length, herein referred to as a unit segment. The length of the route segments is the same as the length of the segments of the stored example situations. It should be noted that in some embodiments, lengths of the segments of the route may differ from each other. It should be noted that the number of positions in FIG. 4 is, for simplicity of this presentation, considerably lower, than in a practical embodiment of the invention. For example, the length of the segments may be metres. This would mean that for a route that is 1 km, there would be 101 positions, one every 10 metres.

[0092] In dependence on the topology data, an altitude ALT is associated with each position P0-P9. In this example, the road inclination of each segment ST1-ST9 is assumed to be constant. As exemplified in FIG. 4, the road inclination α of a segment ST4 can be determined by the altitudes of the positions P3, P4 at the ends of the segment ST4, and the length of the segment.

[0093] Reference is made also to FIG. 5. Determining the velocity profile comprises setting up a matrix MX for the route RT. The matrix includes the sequence of route positions P0-P9. For each position, a plurality of vehicle velocity values V1-V9 are provided. It should be noted that the number of positions and velocities in the matrix in FIG. 5 is, for simplicity of this presentation, considerably lower, than in a practical embodiment of the invention.

[0094] The velocity values are indicated in FIG. 5 with circles. A pair of a position and a velocity, in the matrix, is herein also referred to as a node. The velocity values are, for each position, stepped, discrete values. The size of the steps can be chosen with an aim to obtain a suitable balance between a relatively low size of the task of computing the velocity profile, and the accuracy of the approximation of the computation. For example, the velocity steps may be 0.1 m/s.

[0095] Consideration is made for velocity limits, for example imposed by road curves, or legal speed limits. For this, some velocity values may be removed in the matrix MX. This is exemplified in FIG. 5 by removing the nodes of some velocities V6-V9 at some of the positions P5, P6. Also, lower velocity limits may be taken into consideration for the matrix.

[0096] The method further comprises determining, based at least partly on the route topology, and the plurality of the example situation costs CS, a route cost. The route cost is dependent on a cost of operating the vehicle when travelling through the route. As exemplified in FIG. 5, determining the route cost comprises determining S5, for a pair of adjacent positions P2, P3 in the sequence of positions P0-P9, a vehicle velocity V5 at one of the positions P2, and determining a plurality of candidate predicted velocity changes, each providing a respecting vehicle velocity V1-V9 at the other of the positions P3. The candidate predicted velocity changes are in the example in FIG. 5 indicated with lines between the nodes in question. The candidate predicted velocity changes in this example are designated P2V5-P3V1, P2V5-P3V2, P2V5-P3V3, etc.

[0097] The method further comprises determining S6, for each of a plurality of the candidate predicted velocity changes, a candidate predicted situation cost, based at least partly on at least one of the stored example situation costs CS, by using the approximation function, as exemplified below. One or more of the candidate predicted situation costs may be determined by using the approximation function. Thereby, a candidate predicted situation cost may be determined by an interpolation of the two or more of the example situation costs CS.

[0098] Alternatively, one or more of the candidate predicted situation costs may be determined by identifying a respective stored example situation cost CS of an example situation. The example situation may be an example situation characterized by example situation features LS, αS, VS, VCS matching features of the candidate predicted velocity change, e.g. the vehicle velocity V5 at one of the positions P2, the vehicle velocity V1-V9 at the other of the positions P3, and the road inclination a. The example situations of the example situation costs CS may be identified by providing the features of the respective candidate predicted velocity changes, i.e. the velocity V5 at one of the positions P2, the velocity V1-V9 at the other of the positions P3, and the road inclination α, e.g. in the form of a vector, for a search in the storage of example situations.

[0099] The method further comprises identifying non-feasible candidate velocity changes. For example, certain velocity changes might not be possible within the length of the segment between the positions P2, P3, for example, due to a limitation of the road friction, a capacity limitation of the powertrain, and/or a capacity limitation of the brake system. For example, in FIG. 5, the candidate velocity changes P2V5-P3V1 and P2V5-P3V9 might not be feasible. Whether or not a candidate velocity change is feasible may be determined when the candidate predicted situation cost thereof is determined. The approximation function may indicate that the candidate velocity change is non-feasible. Alternatively, a storage of example situations may indicate that the candidate velocity change is non-feasible. Thus, the method may comprise determining, for each of the feasible candidate predicted velocity changes, a candidate predicted situation cost, based at least partly on at least one of the example situation costs CS.

[0100] The velocity profile is determined such that a weighting function, dependent on the cost of operating the vehicle when travelling through the route, and the duration of the vehicle travelling through the route, is optimized, e.g. minimized. The weighting function may be provided in any suitable way. The weighting function provides a balance between operating costs and the cost of time, which is an indication of the productivity of the vehicle.

[0101] For the operation cost and productivity balance, the example situation costs CS, and/or the candidate predicted situation costs, may be adjusted, based at least partly on the time of driving between the adjacent positions P2, P3. In this example the candidate predicted situation costs are adjusted S7. For example, the weighting function can be

CCadj=(1−w)*CC/CCopernom+w*Δt/Δtnom  (1)

where CCadj is the adjusted candidate predicted situation cost, CC is the non-adjusted candidate predicted situation cost, i.e. the operating cost for the candidate velocity change, CCopernom is a reference operating cost to normalize the operating cost CC, Δt is the time for travelling between the positions P2, P3, Δtnom is a reference time to normalize the time Δt, and w is a weight factor for balancing the operating cost and time. The weighting factor w may be in the interval 0 to 1. The referencing operating cost CC opernom may be determined, for example, by making it equal to the median of the predicted situation costs of all segments along the route. Alternatively, referencing operating cost CC opernom may be determined, for example, by making it equal to the mean value of the predicted situation costs of all segments along the route.

[0102] An alternative for the weighting function may be:

CCadj=(1−w)*(CC1/CC1nom+CC2/CC2nom)+w*Δt/Δtnom  (2)

where CC1 is a first type of operating cost, e.g. the cost of energy usage, and CC2 is a second type of operating cost, for example a cost of wear, e.g. a state of health cost.

[0103] Equations (1) and (2) make the cost and time dimensionless. A more general weighting function could be:

CCadj=(1−w)*f1+w*f2  (3)

where f1 is at least based on the operating cost, and f2 is at least based on time.

[0104] In some embodiments, the example situation costs CS are adjusted before being used for the approximation function determination. In some embodiments, the example situation costs CS are adjusted before being stored. Thereby, the computational task of the vehicle control unit 101 is reduced.

[0105] In addition to the adjustments S7 of the candidate predicted situation costs, the velocity profile for the route RT is determined by selecting, for each pair of adjacent positions P2, P3, one of the candidate predicted situation costs, such that the sum of the candidate predicted situation costs, selected throughout the route, are minimized S8. This optimization may be done by a shortest path algorithm, for example a Bellman-Ford algorithm. For this, one of the velocities at the last position P9 of the route, has to be selected. Since the velocity at the beginning of the path, may be known as the present velocity of the vehicle, end values for the shortest path algorithm are therefore available. The velocity at the last position P9 may be selected in any suitable manner, e.g. as being equal to a legal speed limit at the end position P9.

[0106] It should be noted that the adjustments S7 of the candidate predicted situation costs may be done simultaneously with the selection, throughout the route, of the candidate predicted situation costs, for the minimization S8 of the sum of the candidate predicted situation costs. However, in some embodiments, when the time balanced candidate predicted situation costs have been obtained for all feasible candidate predicted velocity changes in the matrix MX, the velocity profile for the route RT is determined by selecting, for each pair of adjacent positions P2, P3, one of the candidate predicted situation costs, such that the sum of the candidate predicted situation costs, selected throughout the route, are minimized S8.

[0107] The optimization S8 may be done in a stepwise manner. As an alternative to the shortest path algorithm, the velocity at a present position P0 (FIG. 4) of the vehicle 1 may be used as a start node in the matrix MX. Thereby, the lowest adjusted candidate predicted situation cost, for travelling between the first two adjacent positions P0, P1, may be determined as exemplified above. The candidate predicted velocity change of the lowest adjusted candidate predicted situation cost will give a velocity at the next position P1. The process for determining the lowest adjusted candidate predicted situation cost is then repeated for the next pair of adjacent positions P1, P2.

[0108] The result of the optimization S8 will be a series of selected candidate predicted velocity changes, which together form the determined velocity profile VP, as exemplified in FIG. 6.

[0109] The method may further comprise determining an actual cost for operating the vehicle, and adjusting one or more of the example situation costs CS in dependence on the determined actual cost. For this, features of a driving situation, for which the actual cost was determined, may be matched to the example situation features of one or more example situations. Thereby, the accuracy of the stored example situation costs may be improved based on real costs.

[0110] Reference is made to FIG. 7, depicting steps in a method according to a more general embodiment of the invention. The method may be carried out by a vehicle control unit 101 and a stationary control unit 201, as exemplified above. The method comprises establishing S1 a plurality of example situations, wherein each example situation is characterized by a plurality of example situation features, including at least a velocity of the vehicle VS, a velocity change of the vehicle VCS, and a road inclination αS. The method further comprises determining S2, for each of a plurality of the example situations, an example situation cost dependent on a cost for operating the vehicle in the respective example situation. The method also comprises subsequently obtaining S3 topology data, indicative of a topology of a route to be travelled by the vehicle. In addition the method comprises determining S41, based at least partly on the route topology, and at least a plurality of the example situation costs, a velocity profile for the vehicle along the route.

[0111] It is to be understood that the present invention is not limited to the embodiments described above and illustrated in the drawings; rather, the skilled person will recognize that many changes and modifications may be made within the scope of the appended claims.