Traffic planning method for a vehicle fleet
A computer-implemented traffic planning method controls a plurality of vehicles which are movable among multiple shared resources. The method includes receiving a transport mission; generating a root node representing an initial resource occupancy of the vehicles; generating a search tree from the root node, in which each edge represents a motion command and each node represents a resource occupancy, wherein each node is associated with a score related to the vehicles' fulfilment of the transport mission; identifying a target node with an acceptable score; and deriving a planned sequence of motion commands corresponding to a path to the target node. The score of a node includes a short-term component, which represents a cost of executing all motion commands from the root node to said node, and a long-term component, which is determined by the resource occupancy that the node represents.
1 . A computer-implemented traffic planning method for controlling a plurality of vehicles which are movable among multiple shared resources in accordance with predefined motion commands, each resource representing space for transport or parking, a communication channel, maintenance machinery or additional equipment for optional temporary use, wherein a vehicle is allowed to move from a first resource to a second resource if the first resource is connected to the second resource and the second resource is not occupied,
the method comprising the following steps to be performed by processing circuitry of a computer system:
receiving a transport mission to be carried out by the vehicles;
generating a root node representing an initial resource occupancy of the vehicles;
generating a search tree from the root node, in which each edge represents a motion command and each node represents a resource occupancy, wherein each node is associated with a score related to the vehicles' fulfilment of the transport mission, and a score is assigned to a newly generated node in the search tree by computing a value of the short-term component and reading a value of the long-term component from a memory;
identifying, among the nodes of the search tree, a target node with an acceptable score; and
deriving a planned sequence of motion commands corresponding to a path from the root node to the target node,
wherein the score of a node includes a short-term component, which represents a cost of executing all motion commands from the root node to said node, and a long-term component, which is determined by the resource occupancy that the node represents; and
performing an initial scoring phase in which the memory is populated with values of the long-term component by evaluating movements of the vehicles with respect to expected battery energy levels of the vehicles.
2 . The method of claim 1 , wherein the evaluated movements of the vehicles are simulated movements.
3 . The method of claim 2 , wherein the simulated movements are random.
4 . The method of claim 2 , wherein the simulated movements are rule-based.
5 . The method of claim 1 , wherein the evaluated movements of the vehicles are observed real-world movements.
6 . The method of claim 1 , wherein the memory is a trainable model.
7 . The method of claim 1 , wherein the movements of the vehicles are evaluated with respect to fulfilment of the transport mission.
8 . The method of claim 1 , wherein the movements of the vehicles are evaluated with respect to an expected total standstill time of the vehicles.
9 . The method of claim 1 , wherein the score of a node further includes a short-term component that represents the vehicles' fulfilment of the transport mission.
10 . The method of claim 1 , wherein the acceptable score of the identified target node is a maximal score in the search tree and/or exceeds a predefined score threshold.
11 . The method of claim 1 , wherein the vehicles are autonomous vehicles.
12 . A computer system comprising processing circuitry configured to control a plurality of vehicles, which are movable among multiple shared resources in accordance with predefined motion commands, each resource representing space for transport or parking, a communication channel, maintenance machinery or additional equipment for optional temporary use, wherein a vehicle is allowed to move from a first resource to a second resource if the first resource is connected to the second resource and the second resource is not occupied,
the computer system being configured to:
receive a transport mission to be carried out by the vehicles;
generate a root node representing an initial resource occupancy of the vehicles;
generate a search tree from the root node, in which each edge represents a motion command and each node represents a resource occupancy, wherein each node is associated with a score related to the vehicles' fulfilment of the transport mission, and the generating of the search tree includes assigning a score to a newly generated node in the search tree by computing a value of the short-term component and reading a value of the long-term component from the memory;
identify, among the nodes of the search tree, a target node with an acceptable score; derive a planned sequence of motion commands corresponding to a path from the root node to the target node,
wherein the score of a node includes a short-term component, which represents a cost of executing all motion commands from the root node to said node, and a long-term component, which is determined by the resource occupancy that the node represents; and
perform an initial scoring phase in which the memory is populated with values of the long-term component by evaluating movements of the vehicles with respect to the vehicles' fulfilment of the transport mission.
13 . The computer system of claim 12 , wherein the memory is a trainable model.
14 . A vehicle comprising the computer system of claim 12 .
15 . A non-transitory computer-readable storage medium comprising instructions which when executed by processing circuitry, cause the processing circuitry to perform the method of claim 1 .