Route deviation quantification and vehicle route learning based thereon

Abstract

The invention provides a method, device and system for route deviation quantification and vehicle route learning based on the same. In some examples, a method for route deviation quantification of a suggested route is provided. The method includes: obtaining a ground live route based on a filtered trajectory, the filtered trajectory including an inferred start location and an inferred destination location; obtaining a suggested route generated based on the inferred origin location and the inferred destination location; a deviation of the suggested route from the ground live route is quantified by calculating a deviation ratio based on a combined length of a road segment in the suggested route that matches a corresponding road segment in the ground live route and a combined length of a road segment in the ground live route.

Description

technical field [0001] The present disclosure relates broadly, but not exclusively, to methods, devices and systems for route deviation quantification and vehicle route learning based thereon. Background technique [0002] Vehicle routing is critical in the ride-hailing industry for location-based services such as estimated time of arrival (ETA), pricing, navigation, nearest driver search, and more. Accurate route planning is a key factor in providing customers with a superior transport experience by avoiding additional costs (eg, tolls and Electronic Road Pricing (ERP)), heavy traffic and road restrictions. [0003] Traditional vehicle routing is cost-centric and aims to find the lowest-cost route (eg, via the shortest distance or shortest travel time). Shortest path algorithms such as Dijkstra's algorithm and shrinkage hierarchies are used in traditional vehicle routing to suggest the fastest or shortest route. [0004] However, in real life, drivers often choose routes ...

AI Technical Summary

Problems solved by technology

[0006] However, there are multiple challenges in current vehicle routing methods, which include: a. How to quantify the deviation between the route suggested by the vehicle route determination and the driver's actual route, so as to minimize the deviation to improve the route determination results; b. How to efficiently efficiently and effectively learn the driver's voluminous and potentially noisy raw trajectory data to improve routing results; and c. how to improve routing so that it works well across different countries or geographies despite hyperlocalization (For example, in Southeast Asia, traffic rules, road conditions and regulations vary greatly from country to country)

Images