HISTORICAL VEHICLE TRACK VALIDATION IN AN ADAPTIVE ROUTE OPTIMIZATION SOLUTION

This paper details the development of a robust solution for route optimization tailored for commercial vehicle fleets, with a particular emphasis on the specific requirements of small and medium-sized enterprises (SMEs). Our innovative platform integrates several components, including a data ingestion service for real-time GPS data, an integration layer for seamless connectivity with existing customer applications, a sophisticated route optimization engine, and user-friendly interfaces for both web and mobile platforms. A key distinguishing feature of our approach is the incorporation of machine learning (ML) techniques to validate historical route data. This process mitigates the impact of known road hazards, leading to an average reduction of 19.6 % in distance traveled and 14.2 % in route duration when comparing differences between planned and executed routes. Adjusting the optimal route necessitates reliable historical tracks, thus requiring the automatic validation of these tracks with minimal human intervention. In this paper, we describe the implementation of several machine-learning classification models over historical trips and compare the results to select the most suitable model.