Driver quality assessment for driver education

• US 8,915,738 B2
• Filed: 01/24/2012
• Issued: 12/23/2014
• Est. Priority Date: 01/24/2012
• Status: Active Grant

• Current Assignee: Toyota Jidosha Kabushiki Kaisha (Toyota Motor Corporation)
• Original Assignee: Toyota Motor Engineering & Manufacturing North America Incorporated (Toyota Motor Corporation)
• Inventors: Mannino, Matteo T.
• Primary Examiner(s): Egloff, Peter
• Assistant Examiner(s): CRENWELGE, RANDALL NEAL
• Application Number:
  US13/356,917
• Publication Number:
  US 20130189649A1
• Time in Patent Office: 1,064 Days
• Field of Search: 434/65, 434/236, 434/64, 434/69, 340/439, 340/425.5, 340/576, 348/222.1, 382/118, 701/32.7, 701/424, 701/29.2, 701/33, 702/188, 705/4, 705/14.58, 705/11, 455/456.3, 706/20
• US Class Current: 434/65
• CPC Class Codes: G09B 9/052 characterised by provision for recording or measuring trainee's performance devices for psychotechnics, e.g. for vehicle drivers A61B5/16, A61B5/18

Abstract

A driver performance evaluation and apparatus includes means for generating first data stream representing driver actions, a second data stream representing vehicle movements, a filtering function to eliminate short term actions in both data streams, the generation of a raw score which is independent of timing and the adjustment of the raw score based on a comparison of actual driver action timing to an ideal timing sequence to provide a final score which is an indication of driver performance/proficiency.

Claims

1. A training method for evaluating driver performance relative to a standard set by other fully trained drivers comprising:

using a vehicle-mounted camera system to generate a first data stream of driver action sequences representing sequences of driver head position changes and eye gaze direction changes made by a driver during the performance of a series of specific pre-defined vehicle maneuvers;

using a system of vehicle mounted sensors to generate a second data stream representing the series of specific pre-defined maneuvers;

using a processor to produce a score for each driver action sequence correlated to a specific driver-initiated vehicle maneuver from the second data stream as measured against a data set representing ideal driver action sequences, the ideal driver action sequences determined from monitoring different, fully trained drivers performing the same maneuver in an ideal fashion and recording corresponding head position change and eye gaze direction change data from said drivers; and

adjusting each score based on a comparison of the times required by the driver to perform said head position and eye gaze direction actions in each scored sequence relative to a timing standard so as to provide a total numerical score for an evaluator to compare to the driver'"'"'s previous scores to determine the extent to which a driver under examination is progressing.

2. A method as defined in claim 1 including the further step of filtering from the first data stream driver actions having a duration of less than a pre-determined value.

3. A method as defined in claim 1 wherein the first data stream is structured to represent driver action sequences within a moving sampling window such that for each vehicle action in the second data stream, driver actions occurring both early in or prior to as well as late in or after a vehicle action are available for scoring purposes.

4. A method as defined in claim 1 including the further step of determining a running average of the scores.

5. A method as defined in claim 1 wherein the driver actions include looking left, looking partial left for side view mirror check, speed check, straight ahead and center rear view mirror check.

6. A method as defined in claim 1 wherein the vehicle actions include lane changes, backing up and turns.