Concept drift modeling for robust autonomous vehicle control systems in time-varying traffic environments. (15th March 2022)

• Record Type: Journal Article
• Authors: Lee, Sangmin; Park, Sung Ho
• Is Part Of: Expert systems with applications. Volume 190(2022)
• Publication Date: 2022-03-15
• Subjects: Concept drift learning -- Time-varying environments -- Autonomous vehicle systems -- Traffic control -- Automated material handling systems; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2021.116206
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: We propose a concept drift modeling framework for a robust vehicle control system. The proposed framework combines a drift-adaptation technique with a drift detector. The proposed drift detection algorithm can detect actual changes in traffic conditions. The proposed framework is useful to achieve adaptive traffic congestion prediction. Abstract: Autonomous vehicle systems (AVSs) are widely used to transfer wafers in semiconductor manufacturing. However, in such systems, robust traffic control is a significant challenge because all vehicles must be monitored and controlled in real time to cope with traffic congestion. Several predictive approaches have been proposed to prevent traffic congestion in stationary traffic environments. However, in real-life traffic situations, concept drifts exist, which are characterized by time-varying traffic conditions that hinder the accurate prediction of congestion. In this study, we propose a concept drift modeling framework for a robust vehicle control system. The proposed method combines a drift-adaptation learning technique with a drift detector to achieve adaptive traffic prediction in time-varying AVSs. We compare the effectiveness of the prediction and efficiency of model updates with representative methods. High-fidelity simulations based on actual data confirm that the proposed method outperforms alternative methods by detecting change patterns and updating prediction models whenever significant concept drifts occur in traffic patterns.