Development of a workload estimator: The influence of surrounding traffic behaviour on driver workload and performance

The consumers’ increasing desire to be connected at all times and the advancement of integrated functionality within the vehicle, increases the risk that drivers could be faced with information overload while driving. Given the importance of human interaction with technology within the vehicle, automobile manufacturers are introducing workload manager systems within the vehicles to help prevent driver overload. However the ability of the system to decide in a timely manner requires anticipation of changes in workload, depending on the capacity of the driver and matching it with the demand expected from the driving task such as the dynamic traffic environment. In relation to the need to understand the influence of traffic demand on driver workload, the work here comprises the systematic manipulation of traffic complexity and exploration of workload measures to highlight which are sensitive to primary task demand manipulated. A within-subjects design was used in the studies explored in this thesis to allow comparison between different manipulated traffic conditions. In the first simulator test, the ability of various objective and subjective workload measures to tap into drivers’ momentary workload was examined. Following the identification of a subjective measure that was sensitive to the influence of lane changes performed by neighbouring vehicle on drivers’ momentary workload, the characteristics of the lane change were explored in the subsequent studies involving single and dual-task conditions. Overall, these studies suggested suppression of non-urgent communications by a workload manager during safety-critical conditions involving critical cut-ins would be advantageous to both younger and older drivers. This thesis offers a novel and valuable contribution to the design of a workload estimator so as to ensure that the driving demand is always within drivers’ capacity to avoid driver overload. Results of these studies have also highlighted the utility of vehicle-based sensor data in improving workload manager functionality.