File Download

  • Find it @ UNIST can give you direct access to the published full text of this article. (UNISTARs only)

Views & Downloads

Detailed Information

Cited time in webofscience Cited time in scopus
Metadata Downloads

A Steering Wheel Mounted Grip Sensor: Design, Development and Evaluation

Suraiya Jahan Liza
Oakley, Ian
Issued Date
Driving is a commonplace but safety critical daily activity for billions of people. It remains one of the leading causes of death worldwide, particularly in younger adults. In the last decades, a wide range of technologies, such as intelligent braking or speed regulating systems, have been integrated into vehicles to improve safety; annually decreasing death rates testify to their success. A recent research focus in this area has been in the development of systems that sense human states or activities during driving. This is valuable because human error remains a key reason underlying many vehicle accidents and incidents. Technologies that can intervene in response to information sensed about a driver may be able to detect, predict and ultimately prevent problems before they progress into accidents, thus avoiding the occurrence of critical situations rather than just mitigating their consequences. Commercial examples of this kind of technology include systems that monitor driver alertness or lane holding and prompt drivers who are sleepy or drifting off-lane. More exploratory research in this area has sought to capture emotional state or stress/workload levels via physiological measurements of Heart Rate Variability (HRV), Electrocardiogram (ECG) and Electroencephalogram (EEG), or behavioral measurements of eye gaze or face pose. Other research has monitored explicitly user actions, such as head pose or foot movements to infer intended actions (such as overtaking or lane change) and provide automatic assessments of the safety of these future behaviors – for example, providing a timely warning to a driver who is planning to overtake about a vehicle in his or her blind spot. Researchers have also explored how sensing hands on the wheel can be used to infer a driver’s presence, identity or emotional state.
This thesis extends this body of work through the design, development and evaluation of a steering wheel sensor platform that can directly detect a driver’s hand pose all around a steering wheel. This thesis argues that full steering hand pose is a potentially rich source of information about a driver’s intended actions. For example, it proposes a link between hand posture on the wheel and subsequent turning or lane change behavior. To explore this idea, this thesis describes the construction of a touch sensor in the form of a steering wheel cover. This cover integrates 32 equidistantly spread touch sensing electrodes (11.250 inter-sensor spacing) in the form of conductive ribbons (0.2" wide and 0.03" thick). Data from each ribbons is captured separately via a set of capacitive touch sensor microcontrollers every 64 ms. We connected this hardware platform to an OpenDS, an open source driving simulator and ran two studies capturing hand pose during a sequential lane change task and a slalom task. We analyzed the data to determine whether hand pose is a useful predictor of future turning behavior. For this we classified a 5-lane road into 4 turn sizes and used machine-learning recognizers to predict the future turn size from the change in hand posture in terms of hand movement properties from the early driving data. Driving task scenario of the first experiment was not appropriately matched with the real life turning task therefore we modified the scenario with more appropriate task in the second experiments. Class-wise prediction of the turn sizes for both experiments didn’t show good accuracy, however prediction accuracy was improved when the classes were reduced into two classes from four classes. In the experiment 2 turn sizes were overlapped between themselves, which made it very difficult to distinguish them. Therefore, we did continuous prediction as well and the prediction accuracy was better than the class-wise prediction system for the both experiments.
In summary, this thesis designed, developed and evaluated a combined hardware and software system that senses the steering behavior of a driver by capturing grip pose. We assessed the value of this information via two studies that explored the relationship between wheel grip and future turning behaviors. The ultimate outcome of this study can inform the development of in car sensing systems to support safer driving.
Ulsan National Institute of Science and Technology (UNIST)
Department of Human Factors Engineering


Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.