Characterising Driver Intention via Hierarchical Perception-Action Modelling

We seek a mechanism for the classification of the intentional behaviour of a cognitive agent, specifically a driver, in terms of a psychological Perception-Action (P-A) model, such that the resulting system would be potentially suitable for use in intelligent driver assistance. P-A models of human intentionality assume that a cognitive agent’s perceptual domain is learned in response to the outcome of the agent’s actions rather than vice versa. In this way, the perceptual domain is maintained at an appropriate level of complexity in relation to the agent’s embodied motor capabilities, greatly simplifying visual processing. A hierarchical Perception-Action model further captures the subsumptive task-based hierarchicality implicit within human actions by assuming a parallel subsumption occur within the perceptual domain. Assuming this model enables us to characterise intentions at each level of the P-A hierarchy using a range of perceptual descriptors derived from the UK Highway Code and their correlation with driver gaze behaviour. The problem of classification is thus reconciling high-level protocols (i.e., Highway Code rules) with low-level features. We evaluate generative and discriminative logic-based methods for carrying out this classification based on the control, signal and motor inputs of an instrumented vehicle and find that generative model gives superior intentional classification performance due to the strongly-protocol driven nature of the driving environment.