OPTIMIZATION-DRIVEN ADAPTIVE INFOTAINMENT SUPPRESSION IN ADAS-ENABLED VEHICLES: A MULTI-OBJECTIVE RULE SYNTHESIS FRAMEWORK FOR COGNITIVE LOAD-AWARE SECURE HMI DESIGN

Abstract

Advancing integration of Advanced Driver Assistance Systems (ADAS) and in-vehicle infotainment systems has caused a huge increase in cognitive load in contemporary intelligent vehicles. Although adaptive Human-Machine Interfaces (A-HMIs) should control the flow of information according to the driving circumstances, the majority of existing methods are based on empirically constructed rules or heuristics prioritization strategies. These approaches are not formalized and do not model trade-offs between the amount of workload allocated to a driver and the amount of message queues. This paper suggests a multi-objective optimization model structure for adaptive infotainment suppression in ADAS-equipped vehicles. The scheduling problem is modeled as a constrained combinatorial optimization problem, in which the request of the application is either executed or suppressed by referring to the Boolean decision variables. A workload index is based on Driver-Vehicle-Environment (DVE), and a cost function that is a value used to quantify the queue is minimized together in a Non-dominated Sorting Genetic Algorithm II (NSGA-II). Safety-dominant constraints are used to make sure that critical alerts like forward collision warnings take precedence over all the secondary tasks. Findings show that the suggested framework achieves trade-offs between the reduction of cognitive load and queue management and avoids deterministic and interpretable rule synthesis. The technique provides a middle ground between human factors modeling and evolutionary optimization and provides a scalable human factors methodology of intelligent vehicles designed with secure and workload-aware HMI.