Can gaze control steering?

When steering a trajectory, we direct our gaze to locations (1-3 s ahead) that we want to steer through. How and why are these active gaze patterns conducive to successful steering? While various sources of visual information have been identified that could support steering control, the role of stereotypical gaze patterns during steering remains unclear. Here, experimental and computational approaches are combined to investigate a possible direct connection between gaze and steering: Is there enough information in gaze direction that it could be used in isolation to steer through a series of waypoints? For this, we test steering models using waypoints supplied from human gaze data, as well as waypoints specified by optical features of the environment. Steering-by-gaze was modeled using a "pure-pursuit" controller (computing a circular trajectory toward a steering point), or a simple "proportional" controller (yaw-rate set proportional to the visual angle of the steering point). Both controllers produced successful steering when using human gaze data as the input. The models generalized using the same parameters across two scenarios: (a) steering through a slalom of three visible waypoints located within lane boundaries and (b) steering a series of connected S bends comprising visible waypoints without a visible road. While the trajectories on average broadly matched those generated by humans, the differences in individual trajectories were not captured by the models. We suggest that "looking where we are going" provides useful information and that this can often be adequate to guide steering. Capturing variation in human steering responses, however, likely requires more sophisticated models or additional sensory information.

Tracking an occluded visual target with sequences of saccades.

Gaze behavior during visual tracking consists of a combination of pursuit and saccadic movements. When the tracked object is intermittently occluded, the role of smooth pursuit is reduced, with a corresponding increase in the role of saccades. However, studies of visual tracking during occlusion have focused only on the first few saccades, usually with occlusion periods of less than 1 second in duration. We investigated tracking on a circular trajectory with random occlusions and found that an occluded object can be tracked reliably for up to several seconds with mainly anticipatory saccades and very little smooth pursuit. Furthermore, we investigated the accumulation of uncertainty in prediction and found that prediction errors seem to accumulate faster when an absolute reference frame is not available during tracking. We suggest that the observed saccadic tracking reflects the use of a time-based internal estimate of object position that is anchored to the environment via fixations.

Visual anticipation of the future path: Predictive gaze and steering.

Skillful behavior requires the anticipation of future action requirements. This is particularly true during high-speed locomotor steering where solely detecting and correcting current error is insufficient to produce smooth and accurate trajectories. Anticipating future steering requirements could be supported using "model-free" prospective signals from the scene ahead or might rely instead on model-based predictive control solutions. The present study generated conditions whereby the future steering trajectory was specified using a breadcrumb trail of waypoints, placed at regular intervals on the ground to create a predictable course (a repeated series of identical "S-bends"). The steering trajectories and gaze behavior relative to each waypoint were recorded for each participant (N = 16). To investigate the extent to which drivers predicted the location of future waypoints, "gaps" were included (20% of waypoints) whereby the next waypoint in the sequence did not appear. Gap location was varied relative to the S-bend inflection point to manipulate the chances that the next waypoint indicated a change in direction of the bend. Gaze patterns did indeed change according to gap location, suggesting that participants were sensitive to the underlying structure of the course and were predicting the future waypoint locations. The results demonstrate that gaze and steering both rely upon anticipation of the future path consistent with some form of internal model.

Physiological responses to proposals during dyadic decision-making conversations.

A novel conversation-analytically informed paradigm was used to examine how joint decision-making interaction, with its various types of proposal sequences, is reflected in the physiological responses of participants. Two types of dyads-dyads with one depressed and one non-depressed participant (N = 15) and dyads with two non-depressed participants (N = 15)-engaged in a series of conversational joint decision-making tasks, during which we measured their skin conductance (SC) responses. We found that the participants' SC response rates were higher and more synchronized during proposal sequences than elsewhere in the conversation. Furthermore, SC response rates were higher when the participant was in the role of a proposal speaker (vs. a proposal recipient), and making a proposal was associated with higher SC response rates for participants with depression (vs. participants without depression). Moreover, the SC response rates in the proposal speaker were higher when the recipient accepted (vs. not accepted) the proposal. We interpret this finding with reference to accepting responses suggesting a commitment to future action, for which the proposal speaker may feel specifically responsible for. A better understanding of the physiological underpinnings of joint decision-making interaction may help improve democratic practices in contexts where certain individuals experience challenges in this regard.

Humans use Optokinetic Eye Movements to Track Waypoints for Steering.

It is well-established how visual stimuli and self-motion in laboratory conditions reliably elicit retinal-image-stabilizing compensatory eye movements (CEM). Their organization and roles in natural-task gaze strategies is much less understood: are CEM applied in active sampling of visual information in human locomotion in the wild? If so, how? And what are the implications for guidance? Here, we directly compare gaze behavior in the real world (driving a car) and a fixed base simulation steering task. A strong and quantifiable correspondence between self-rotation and CEM counter-rotation is found across a range of speeds. This gaze behavior is "optokinetic", i.e. optic flow is a sufficient stimulus to spontaneously elicit it in naïve subjects and vestibular stimulation or stereopsis are not critical. Theoretically, the observed nystagmus behavior is consistent with tracking waypoints on the future path, and predicted by waypoint models of locomotor control - but inconsistent with travel point models, such as the popular tangent point model.

Humans Use Predictive Gaze Strategies to Target Waypoints for Steering.

A major unresolved question in understanding visually guided locomotion in humans is whether actions are driven solely by the immediately available optical information (model-free online control mechanisms), or whether internal models have a role in anticipating the future path. We designed two experiments to investigate this issue, measuring spontaneous gaze behaviour while steering, and predictive gaze behaviour when future path information was withheld. In Experiment 1 participants (N = 15) steered along a winding path with rich optic flow: gaze patterns were consistent with tracking waypoints on the future path 1-3 s ahead. In Experiment 2, participants (N = 12) followed a path presented only in the form of visual waypoints located on an otherwise featureless ground plane. New waypoints appeared periodically every 0.75 s and predictably 2 s ahead, except in 25% of the cases the waypoint at the expected location was not displayed. In these cases, there were always other visible waypoints for the participant to fixate, yet participants continued to make saccades to the empty, but predictable, waypoint locations (in line with internal models of the future path guiding gaze fixations). This would not be expected based upon existing model-free online steering control models, and strongly points to a need for models of steering control to include mechanisms for predictive gaze control that support anticipatory path following behaviours.

Effects of an Active Visuomotor Steering Task on Covert Attention.

In complex dynamic tasks such as driving it is essential to be aware of potentially important targets in peripheral vision. While eye tracking methods in various driving tasks have provided much information about drivers' gaze strategies, these methods only inform about overt attention and provide limited grounds to assess hypotheses concerning covert attention. We adapted the Posner cue paradigm to a dynamic steering task in a driving simulator. The participants were instructed to report the presence of peripheral targets while their gaze was fixed to the road. We aimed to see whether and how the active steering task and complex visual stimulus might affect directing covert attention to the visual periphery. In a control condition, the detection task was performed without a visual scene and active steering. Detection performance in bends was better in the control task compared to corresponding performance in the steering task, indicating that active steering and the complex visual scene affected the ability to distribute covert attention. Lower targets were discriminated slower than targets at the level of the fixation circle in both conditions. We did not observe higher discriminability for on-road targets. The results may be accounted for by either bottom-up optic flow biasing of attention, or top-down saccade planning.

A computational model for driver's cognitive state, visual perception and intermittent attention in a distracted car following task.

We present a computational model of intermittent visual sampling and locomotor control in a simple yet representative task of a car driver following another vehicle. The model has a number of features that take it beyond the current state of the art in modelling natural tasks, and driving in particular. First, unlike most control theoretical models in vision science and engineering-where control is directly based on observable (optical) variables-actions are based on a temporally enduring internal representation. Second, unlike the more sophisticated engineering driver models based on internal representations, our model explicitly aims to be psychologically plausible, in particular in modelling perceptual processes and their limitations. Third, unlike most psychological models, it is implemented as an actual simulation model capable of full task performance (visual sampling and longitudinal control). The model is developed and validated using a dataset from a simplified car-following experiment (N = 40, in both three-dimensional virtual reality and a real instrumented vehicle). The results replicate our previously reported connection between time headway and visual attention. The model reproduces this connection and predicts that it emerges from control of action uncertainty. Implications for traffic psychological models and future developments for psychologically plausible yet computationally rigorous models of full natural task performance are discussed.