Patent Application
20060259206
The disclosure relates to vehicle operator monitoring systems. More specifically, the disclosure relates to vehicle operator attentiveness imaging for adjusting vehicle systems.
Each year numerous automobile accidents are caused by vehicle operator distractions. The National Highway Traffic Safety Administration (NHTSA) estimates that vehicle operator distraction is directly involved in twenty to thirty percent of all automobile accidents or roughly 1.6 million automobile accidents in the U.S. annually. Visual distraction of the vehicle operator is attributed to many accidents. A need therefore exists to provide a real-time monitoring of the visual distraction of the vehicle operator.
The present disclosure is described, by way of example, with reference to the accompanying drawings, in which:
Referring to
The video imaging camera 16 may include CCD/CMOS active-pixel digital image sensors mounted as individual chips onto a circuit board (not shown). One example of a CMOS active-pixel digital image sensor is Model No. PB-0330, commercially available from Photobit, which has a resolution of 640H×480V. The use of digital image sensors for the video imaging camera 16 also allows for the detection of stereo information. The video imaging camera 16 may also be coupled to an eye tracking processor (not shown). The eye tracking processor may include a frame grabber for receiving the video frames generated by the video imaging camera 16. The video imaging camera may also be coupled to a video processor for processing the video frames. The video processor includes memory, such as random access memory (RAM), read-only memory (ROM), and other memory as should be readily apparent to those skilled in the art. Other features of the vehicle operator attentiveness systems of
Referring to
The present disclosure utilizes the identification of vehicle operator attentiveness determined from distracted, ‘non-forward’ ocular profiles 50b, 50c so that countermeasures can be taken to enhance the operation of vehicle systems. The real time duration and frequency of the vehicle operator's distracted, ‘non-forward’ view may be captured to identify if the vehicle operator 20 is distracted, if, for example, a particular task is being conducted by the vehicle operator 20, such as, for example, radio-tuning, that distracts the vehicle operator 20 from maintaining attentiveness of the road ahead in the forward view window 75.
Referring to
Referring now to steps 120-124, a general examination of the ocular profiles 50a, 50b, 50c is conducted by locating, determining, and storing an imaged frame of the vehicle operator's ocular features at 52a-54c to determine the attentiveness type of the captured ocular profile 50a, 50b, 50c. Steps 126-128 cycle the general examination of steps 120-124 within a time interval, Y, to capture sequentially imaged frames of the ocular profile 50a, 50b, 50c to determine a proportional amount of time that a vehicle operator 20 may be classified as having an attentive ocular profile 50a or distracted ocular profile 50b, 50c. Because the examined ocular profiles 50a, 50b, 50c are captured sequentially at step 124 on a frame-rate basis, real time data can be calculated at step 130 to determine the level of vehicle operator visual distraction.
The real time data calculated at step 130 is a percentage of a series of saved data from step 124. According to an embodiment, the calculated percentage may relate to distracted ocular profiles 50b, 50c captured over a given time interval, Y. The calculation at step 130 is determined by summing frames of distracted ocular profiles 50b, 50c over the time interval, Y, and dividing the summed frames of distracted ocular profiles 50b, 50c over the time interval, Y, by a total series of frames captured over the time interval, Y, that includes attentive and distracted ocular profiles 50a, 50b, 50c. The expression for determining the calculation at step 130 is:
PORT=n/N
where “PORT” stands for Proportion of Off-Road glance Time, “n” is the number of frames that the vehicle operator's ocular profile is classified as a distracted ocular profile 50b, 50c, and ‘N’ is a total number of predetermined series of frames (i.e. both attentive ocular profiles 50a and distracted ocular profiles 50b, 50c) to be captured over the time interval, Y.
To calculate PORT, a counter value, X, is initialized at zero and the time interval, Y, is set to any desirable value in step 102. The values of X and Y are compared at step 126. Once the value, X, of the counter is greater than or equal to the value of the time interval, Y, which is associated with the total number of frames, N, the PORT is calculated at step 130. If the criteria at step 126 is not met, the algorithm is advanced to step 128 where the counter value, X, is incremented by any desirable value, Z. This loop at steps 120-128 is continuously cycled until the criteria at step 126 is met. The value of the counter, X, time interval, Y, and incrementation value, Z, may be any desirable value, such as, for example, 0.10 seconds, 0.25 seconds, 0.50 seconds, 1 second, 2 seconds, 3 seconds, 5 seconds, 10 seconds, 15 seconds, 30 seconds, or 60 seconds.
Utilizing the calculated PORT value at step 130, a visually-distracted vehicle operator 20 may timely respond to potentially dangerous driving conditions upon application of the PORT value at step 132. PORT is found to be directly correlated with vehicle operator distraction and safety variables such as brake reaction times, lane departures, and standard deviation of lane position; as PORT increases, reaction times, lane departures, and standard deviations of lane position all increase. Accordingly, the relationship between PORT and safety criteria is reliable across a wide range of conditions. As such, PORT can be implemented in several manners to mitigate visual distraction and enhance vehicle operator safety.
According to an embodiment, a PORT threshold value can be established at step 132 as a reference value to determine when there is an excessive level of vehicle operator distraction. For example, a predetermined PORT threshold value may be set at 0.50 (i.e. 50% of the captured frames include distracted ocular profiles 50b, 50c), which is used as the basis for comparing a calculated PORT value from step 130. In application, according to an embodiment, if the calculated PORT at step 130 is 0.75 (i.e. 75% of the captured frames include distracted ocular profiles 50b, 50c), the predetermined threshold value of 0.50 has been exceeded, and thus, gentle braking pattern warnings can be delivered to the vehicle's braking system at step 132 to remind the vehicle operator 20 to pay more attention to the forward view window 75.
Using the above-described braking system example, the first application of the calculated PORT information at step 132 may take place approximately at the value of the predetermined time interval value, Y, that is set at step 102 if the vehicle operator 20 is determined to be in a distracted state. Once the application takes place at step 132, the algorithm is advanced to step 134 where the counter value, X, and the time interval value, Y, are incremented by the incrementation value, Z. After the incrementation is registered at step 134, the algorithm is returned to step 104 so that the ocular profile of the vehicle operator 20 may be examined at subsequent timeframes. Accordingly, the ocular profile data for a subsequent timeframe is stored step 124, PORT is calculated for the subsequent timeframe at step 130, and the application of the PORT data is conducted once again at step 132. Upon feeling the gentle braking patterns, the vehicle operator 20 may become more attentive to the road (i.e. the vehicle operator transitions to an attentive ocular profile 50a), thereby drawing down the calculated PORT value from 0.75. As the vehicle operator 20 maintains attentiveness of the forward view window 75, the calculated PORT value eventually falls below the predetermined PORT threshold value of 0.50 and the gentle braking pattern warnings are ceased, thereby restoring normal operation of the vehicle 10.
The calculation and application of the PORT data at steps 130, 132 for subsequent timeframes may take place for any desirable period. For example, ocular profiles 50a, 50b, 50c may be examined over a running (i.e. limitless) time interval, Y. Accordingly, when the vehicle operator 20 keys the vehicle, the time interval Y may be continuously incremented until the vehicle 10 is keyed-off so that attentiveness of the vehicle operator 20 may be determined over the entire operating period of the vehicle 10. Alternatively, the calculated and applied PORT data may be utilized for recent (i.e. limited) time intervals, Y. Accordingly, the most recent, newly-captured ocular profile data may ‘bump’ the oldest ocular profile data such that the calculated PORT value from step 130 over a limited time interval, Y, can be constrained to the most recent ocular profile data. For example, if Y is set to be equal to 5 seconds and Z is equal to 1 second, the first PORT calculation at step 130 takes place from the time interval of 0-5 seconds; accordingly, the subsequent calculation of PORT values at step 130 may be over the limited time interval range of 1-6 seconds, 2-7 seconds, 3-8 seconds, etc. Thus, the applied PORT data at step 132 may be refined such that the PORT data is related to the most recent vehicle operator attentiveness information.
According to an embodiment, safety warning systems, such as, for example, forward collision warning (FCW) systems, can be enhanced using the calculated PORT information from step 130. As is known, FCW systems include high levels of “nuisance alerts.” Because FCW system alerts are associated with the timing of a potential collision, the real-time calculated PORT information may be utilized for comparison to the threshold value in step 132 as described above to more appropriately time and reduce the number of alerts. Thus, according to an embodiment, if a calculated PORT value of 0.75 exceeds a predetermined PORT value of 0.50, step 132 may modify the alert sent to the FCW system so that the distracted vehicle operator 20 may receive the alert earlier than normal to provide additional time for the vehicle operator 20 to avoid a crash. Conversely, if the calculated PORT value is 0.10 (i.e. 10% of the captured frames include distracted ocular profiles 50b, 50c), which is below the predetermined PORT value of 0.50, step 132 may apply a signal to the FCW system that maintains, disables, or modifies the alert so that a generally non-distracted vehicle operator 20 will be less likely to be annoyed by alerts since the vehicle operator's ocular profile 50a is focused on the road in the forward view window 75 approximately 90% of the time.
According to another embodiment, a vehicle operator attentiveness alert/calculated PORT value from step 130 may be utilized in a similar manner as described above in an automated cruise control (ACC) system, which is shown generally at 250 in
In application, the vehicle operator attentiveness alert/PORT information from step 216 is used at step 218 as feedback for adjusting the ACC system 250 based on the alertness level of the vehicle operator 20. According to an embodiment, adjustments can be made to the ACC system at step 218 to increase the following distance (i.e., by reducing throttle) to allow for longer reaction time when a vehicle operator 20 is distracted. In an another embodiment, the ACC system 250 may apply the brakes at step 218 in response to slower-moving or stationary vehicle/objects 11a-11f during periods of vehicle operator distraction. The braking at step 218 may be applied to the vehicle 10 in a manner to alert the vehicle operator 20 so as to acquire vehicle operator attention quickly. For example, a low automatic braking may provide a very powerful stimulus for returning the vehicle operator's attention to a forward view window 75. The adjusted ACC state at step 218 may occur for any desired period. If desired, steps 202, 204, 216, and 218 may be continuously looped until the vehicle operator maintains an attentive ocular profile 50a for an expected time interval or period.
While the invention has been specifically described in connection with certain specific embodiments thereof, it is to be understood that this is by way of illustration and not of limitation, and the scope of the appended claims should be construed as broadly as the prior art will permit.
