Search

Monitoring drivers' mental workload facilitates initiating and maintaining safe interactions with in-vehicle information systems, and thus crucial for delivering adaptive human machine interaction solutions with reduced impact on the primary task of driving. In this article, we tackle the problem of workload estimation from driving performance data. First, we present a novel on-road study for collecting subjective workload data via a modified peripheral detection task in naturalistic settings. Key environmental factors that induce a high mental workload are identified via video analysis, e.g. junctions and behaviour of vehicle in front. Second, a supervised learning framework using state-of-the-art time series classifiers (e.g. convolutional neural network and transform techniques) is introduced to profile drivers based on the average workload they experience during a journey. A Bayesian filtering approach is then proposed for sequentially estimating, in (near) real-time, the driver's instantaneous workload. This computationally efficient and flexible method can be easily personalised to a driver (e.g. incorporate their inferred average workload profile), adapted to driving/environmental contexts (e.g. road type) and extended with data streams from new sources. The efficacy of the presented profiling and instantaneous workload estimation approaches are demonstrated using the on-road study data, showing $F_{1}$ scores of up to 92% and 81%, respectively.

Most research into haptic feedback for in-car applications has used vibrotactile feedback. In this article, two simulator studies investigate novel thermal feedback during driving for a lane change task. The distraction and time differences of audio and thermal feedback were investigated in the first, with results showing that thermal feedback does not increase lane deviation, but the time to completed lane change is 1.82s longer in the thermal than the audio condition. The second experiment explored the difference in variable changes of the thermal stimuli on the recognition rate and false positive recognition at the return to the neutral temperature. Variable alterations can have different effects on these tasks and are not mirrored for the directions of temperature change. This suggests that the design of thermal stimuli is highly dependent on what result should be maximized: recognition rate or minimal additional changes at the return to the neutral temperature.

Two studies investigated the concept of following a lead vehicle as a navigational aid. The first video-based study (n=34) considered how drivers might use a real-world lead vehicle as a navigational aid, whist the second simulator-based study (n=22) explored how an Augmented Reality (AR) virtual car, presented on a head-up display (HUD), may aid navigation around a complex junction. Study 1 indicated that a lead vehicle is most valued as a navigation aid just before/during a required maneuver. During the second study the dynamic virtual car (which behaved like a real vehicle) resulted in greater confidence and lower workload than a static virtual car that “waits” at the correct junction exit, but resulted in more gaze concentration. It is concluded that a virtual car may be a valuable element of a navigation system, in combination with other forms of information, to completely fulfil all a driver's navigational task requirements.

As the characteristics of in-vehicle human-machine interfaces (HMIs), the driving task, and the expectations and behavior of drivers have evolved, so too should our thinking and approach to HMI design and evaluation. This panel will present background and perspectives on the current status, emerging needs, challenges, and opportunities in this area. A key focus for the panel is an emphasis on attention support. Detailed contextual description is provided below for reference to allow panelists to keep their opening remarks relatively brief to allow for substantive question and discussion time with the audience.

Predictive touch technology aims to improve the usability and performance of in-vehicle displays under the influence of perturbations due to the road and driving conditions. It fundamentally relies on predicting and early in the freehand pointing movement, the interface item the user intends to select, using a novel Bayesian inference framework. This article focusses on evaluating facilitation schemes for selecting the predicted interface component whilst driving, and without physically touching the display, thus touchless. Initially, several viable schemes were identified in a brainstorming session followed by an expert workshop with 12 participants. A simulator study with 24 participants using a prototype predictive touch system was then conducted. A number of collected quantitative and qualitative measures show that immediate mid-air selection, where the system autonomously auto-selects the predicted interface component, may be the most promising strategy for predictive touch.

In present-day highly-automated vehicles, there are occasions when the driving system disengages and the human driver is required to take-over. This is of great importance to a vehicle &#x02BC s safety and ride comfort. In the U.S state of California, the Autonomous Vehicle Testing Regulations require every manufacturer testing autonomous vehicles on public roads to submit an annual report summarizing the disengagements of the technology experienced during testing. On 1 January 2016, seven manufacturers submitted their first disengagement reports: Bosch, Delphi, Google, Nissan, Mercedes-Benz, Volkswagen, and Tesla Motors. This work analyses the data from these disengagement reports with the aim of gaining abetter understanding of the situations in which a driver is required to takeover, as this is potentially useful in improving the Society of Automotive Engineers &#x0028 SAE &#x0029 Level 2 and Level 3 automation technologies. Disengagement events from testing are classified into different groups based on attributes and the causes of disengagement are investigated and compared in detail. The mechanisms and time taken for take-over transition occurred in disengagements are studied. Finally, recommendations for OEMs, manufacturers, and government organizations are also discussed.

The purpose of this study was to evaluate an alleged EEG amplifier’s, the Freer Logic BodyWave, ability to record EEG by comparing its signals to those recorded at the same time by a proven EEG amplifier, the Emotiv EPOC+. The BodyWave’s circuitry is similar to that of an Electro-dermal Activity meter and alleges to record surface-EEG from a limb of the user. The EPOC+ has been proven in previous studies to record signals of high-enough quality to be used in research, whilst the BodyWave is unproven. In order to compare their recordings, both EEGs were worn simultaneously whilst different stimuli induced various brain states. 15 participants were recruited and each subjected to two different experimental paradigms meant to induce increased attention and increased relaxation. Both headsets achieved a classification accuracy well above chance when differentiating between the attention and relaxation tasks. Further testing provided evidence that data from the alpha and beta frequency bands was contributing the most to the classifier’s accuracy. However the total lack of correlation between the signals of the two devices strongly implies that the BodyWave is not measuring EEG. Despite this these results show that in some cases it can be used in place of an EEG to perform a similar function. It also proved to be much more robust against muscle artefacts compared to the EPOC+.