Cloud-Based System Architecture for Driver Assistance in Mobile Machinery

Abstract

Using the example of a wheel loader, this paper presents a cloud-based system architecture enabling intelligent machine behavior. In order to achieve the final goal of a fully automated bucket filling routine, while controlling the loaders engine, travel drive and attachment, different levels of automation are processed gradually. As a first step towards automation, driver assistance can be considered. The paper explains the design choices for a cyber-physicalsystem architecture in the context of construction machinery. This comprises the communication framework and the cloud-application for self-adapting systems (i.e. the MAPE-K loop). As a validation of the architecture and as a demonstrator, a driver assistance functionality has been implemented. Calculations from the cloud-application give the operator feedback about efficiency, loads and task status. A developed visualization app on a tablet serves as user-interface. Concurrent simulation allow an optimization of control algorithms for the machine control and the trajectory planning. Besides changing the parametrization of the underlying models, a solution to change ECU-code at run-time without interrupting the operation is presented. The developed system architecture is the basis for further implementations of adaptive algorithms that improve future machine operation.