Handwriting Evaluation Using Deep Learning with SensoGrip.

Handwriting learning disabilities, such as dysgraphia, have a serious negative impact on children's academic results, daily life and overall well-being. Early detection of dysgraphia facilitates an early start of targeted intervention. Several studies have investigated dysgraphia detection using machine learning algorithms with a digital tablet. However, these studies deployed classical machine learning algorithms with manual feature extraction and selection as well as binary classification: either dysgraphia or no dysgraphia. In this work, we investigated the fine grading of handwriting capabilities by predicting the SEMS score (between 0 and 12) with deep learning. Our approach provided a root-mean-square error of less than 1 with automatic instead of manual feature extraction and selection. Furthermore, the SensoGrip smart pen SensoGrip was used, i.e., a pen equipped with sensors to capture handwriting dynamics, instead of a tablet, enabling writing evaluation in more realistic scenarios.

Comparison of two human-machine-interfaces for cooperative maneuver-based driving.

In the project "Conduct-by-Wire" which is founded by the German Research Foundation (DFG) cooperative maneuver based driving is examined. In this paper two different input devices (gesture recognition and tactile touch display) are compared in a simulator study with 29 participants. It shows that the major advantage of the gesture recognition is that there is no need for the driver to take his gaze off the road. In contrast, the number of gazes at the tactile touch display is significantly higher. The major advantage of the tactile touch display is that no input errors occurred during the test drives. Conversely, the gesture recognition was significantly worse. Nevertheless, further work is needed to decide which input device is the best.

User acceptance of cooperative maneuver-based driving--a summary of three studies.

Modern cars offer drivers support with the help of a number of driver assistance systems. Those systems aim to relieve drivers through assumption of sub parts of the driving task (e.g. in case of an Adaptive Cruise Control by regulation of vehicle speed and time gap to preceding vehicle). Today, systems are controlled and monitored separately which leads to efforts to combine the functionality of all systems in an overlying assistance for drivers. The approach of the University of Technology Darmstadt is called Conduct-by-Wire and can be seen as a cooperative maneuver-based driving paradigm, where the driver gives maneuver command to the systems which are automatically executed. This paper summarizes the results of three studies which investigated the user acceptance of this driving paradigm. Overall, it can be said that the acceptance of the system depends on personal traits of the driver and on the driving situation. Almost all participants are willing to use Conduct-by- Wire for routine tasks such as commuting, which makes the systems interesting for company cars. Still, there remain a number of drivers who are not willing to use such a highly automated system at all.