A flexible scheduling for automobile control using imprecise computation and its fundamental evaluation

Abstract

Imprecise computation is one of the promising schemes in real time systems to adapt the quality of computations to the change of load, keeping the deadlines of tasks in the systems. When overload occurs in the systems, the minimum requirements on the deadline are assured by decreasing the quality of the computation. This paper describes how to apply the concept of imprecise computation to automobile control in the expressway assuming the intelligent transportation system (ITS). The deadline violation of tasks for automobile control in the expressway induces collision of automobiles. Regardless of whether the expressway is congested or not, collision of automobiles must be avoided. To satisfy such requirement, the concept of imprecise computation is effective. This paper proposes a flexible scheduling using imprecise computation to avoid collision of automobiles and increase throughput.