US 12,392,620 B2
Speeding up stochastic routing of vehicles using tiering
Chinya V. Ravishankar, Irvine, CA (US); and Payas Rajan, Riverside, CA (US)
Assigned to The Regents of the University of California, Oakland, CA (US)
Filed by The Regents of the University of California, Oakland, CA (US)
Filed on Aug. 23, 2023, as Appl. No. 18/237,136.
Claims priority of provisional application 63/400,353, filed on Aug. 23, 2022.
Prior Publication US 2024/0068823 A1, Feb. 29, 2024
Int. Cl. G01C 21/34 (2006.01)
CPC G01C 21/3446 (2013.01) [G01C 21/3453 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A vehicle routing method comprising:
storing, at a server that is communicatively coupled with one or more vehicles via a wireless communication network, a graph representing a road network, the graph comprising vertices and edges between vertices, where the vertices represent locations in the road network, and wherein the edges represent road segments between the locations;
augmenting, at the server, the graph representing the road network by adding shortcut edges to the edges, wherein the shortcut edges represent paths comprising a sequence of connected edges in the graph representing the road network;
receiving and storing edge weights representing travel times for travel along edges of the graph, where the edge weights are represented as probability distributions;
partitioning the edges of the augmented graph into multiple tiers including an upper tier and a lower tier, wherein the upper tier has a first set of edges whose weights are specified as stable distributions and/or wherein convolutions of probability distributions of the first set of edges can be feasibly replaced with functional approximations, and wherein the lower tier has a second set of edges wherein convolutions of probability distributions of the second set of edges cannot be feasibly replaced with functional approximations;
receiving, a routing request from a vehicle of the one or more vehicles over the wireless communication network, wherein the routing request includes a vehicle origin and a vehicle destination of the vehicle;
performing a Dijkstra search on the augmented graph from the vehicle origin to the vehicle destination, wherein the Dijkstra search computes convolutions of distributions and sums of functional approximations for paths being computed; and
returning, over the wireless communication network, routing information to the vehicle destination produced by the Dijkstra search, wherein the routing information is configured to guide a human driver or an autonomous driving system.