Morgan State University

Mansoureh Jeihani

Samira Ahangari

Zohreh Rashidi Moghaddam

Arsalan Hassan Pour

Snehanshu Banerjee

Abdollah Dehzangi

Mahmoud Ayman Sayed

Computer-implemented systems and methods are provided for the synchronization of data from asynchronous datasets to form a merged, synchronized data set, and more particularly to systems and methods for merging disparate and asynchronous data sets that include common dimensions, such as space and time dimensions, and more particularly space and time data relating to a driving simulation environment (such as driving simulator data and eye tracking data having disparate time data resolutions), to enable data extraction from and analysis of the merged, synchronized data set. A driving simulator system includes a driver interface having driving controls and a display and creates driving simulator data files for each driver of the driving simulator system having data entries that are captured at a first timing frequency. Likewise, an eye tracking system includes eye tracking cameras for tracking a driver's eye movements when using the driving simulator system and creates eye tracking data files for each driver of the driving simulator system having data entries that are captured at a second timing frequency that is different from the first timing frequency at which driving simulator data is captured. A driving behavior analysis computer is also provided that merges the asynchronous data embodied in each of the driving simulator data files and the eye tracking data files to form a single merged data file based upon an interpolation of data entries in those data files. The merged data files may then be further processed to enable extraction and analysis of the merged data in order to identify various driver behaviors.

System and method for synchronization of asynchronous datasets

Systems and methods are disclosed for detecting both the occurrence and type of driver distraction experienced by a driver, determined through evaluation of sensed vehicle conditions or activities, such as steering, braking, lane changing, etc. as detected by one or more sensors on or associated with the vehicle. That detection of the occurrence and type of driver distraction may then be used to initiate another action, including initiating an audible or visual alert to the driver, taking measures to interfere with or stop operation of the device that is causing the distraction, log the occurrence and type of distraction that occurred, and report the occurrence and type of distraction to an outside monitoring computer (such as one associated with a guardian of the driver, an insurer of the vehicle, a law enforcement authority, or the like). To allow such detection, the system and method set forth herein employ machine learning methods to first train a classifier to classify certain driver behaviors (as evidenced by sensed vehicle movements and conditions) as a distraction or non-distraction event, and if a distraction event is detected then to further classify a type of such distraction event, and then to apply the trained classifier to classify the driver's ongoing driving activity.

System and method for driver distraction detection and classification

Disclosed is a system and method for generating a vehicle speed alert in a driving simulation system that communicates with a remote driving simulator engine and a remote speed optimization engine. The remote speed optimization engine receives data from the vehicle speed alert system after the vehicle speed alert system has received data from the remote driving simulator engine, including the vehicle's simulated distance to a signalized intersection, the time remaining for the simulated traffic light to change, and the current simulated light status (green, yellow, red) of the traffic light. The vehicle speed alert system then receives from the remote speed optimization engine a recommended speed profile for that given instant, and if the driver's current speed does not fall within some maximum difference with the current recommended speed profile, calculates and transmits an alert to an output device to alert the driver of action necessary to achieve the recommended speed profile.

Growing community of inventors

Lutherville, MD, United States of America

Mansoureh Jeihani