Driver monitoring system and method of operating the same
A driver monitoring system for a vehicle and method of operating the driver operating system. The method, in one implementation, involves receiving a plurality of glance aim points for a driver of the vehicle; inputting the plurality of glance aim points into a predictive probability of distraction model to obtain a predictive distraction distribution; determining whether one or more informative glance locations are present in the plurality of glance aim points; comparing the predictive distraction distribution to a predictive distraction distribution threshold when one or more informative glance locations are present in the plurality of glance aim points; and alerting the driver when the predictive distraction distribution satisfies or exceeds the predictive distraction distribution threshold.
1. A method of operating a driver monitoring system for a vehicle, comprising the steps of:
receiving a plurality of glance aim points for a driver of the vehicle;
inputting the plurality of glance aim points into a predictive probability of distraction model to obtain a predictive distraction distribution, wherein the predictive probability of distraction model includes a hidden Markov model (HMM);
determining whether one or more informative glance locations are present in the plurality of glance aim points;
comparing the predictive distraction distribution to a predictive distraction distribution threshold when one or more informative glance locations are present in the plurality of glance aim points; and
alerting the driver when the predictive distraction distribution satisfies or exceeds the predictive distraction distribution threshold.
2. The method of claim 1 , wherein the predictive distraction distribution is a graphical or numerical data scheme representing a crash or near crash probability based on prior data.
3. The method of claim 1 , further comprising the step of fusing the predictive distraction distribution with a predictive saliency distribution when one or more informative glance locations are not present in the plurality of glance aim points.
4. The method of claim 3 , further comprising the step of determining a current driving state and using the current driving state as input when fusing the predictive saliency distribution and the predictive distraction distribution.
5. The method of claim 1 , wherein the predictive probability of distraction model includes using an observation matrix of event probabilities for given glance locations to obtain the predictive distraction distribution.
6. The method of claim 1 , wherein the predictive probability of distraction model includes developing a graphical representation of the plurality of glance aim points.
7. The method of claim 6 , wherein the graphical representation is a histogram plot of the plurality of glance aim points.
8. The method of claim 6 , wherein the graphical representation is used to develop a glance co-occurrence matrix which captures a plurality of glance transitions with the plurality of glance aim points.
9. The method of claim 8 , further comprising the step of feeding the glance co-occurrence matrix into the hidden Markov model (HMM).
10. The method of claim 9 , wherein the feeding is accomplished in a sliding window fashion in one-step increments.
11. A method of operating a driver monitoring system for a vehicle, comprising the steps of:
receiving a plurality of glance aim points for a driver of the vehicle;
inputting the plurality of glance aim points into a predictive probability of distraction model to obtain a predictive distraction distribution, wherein the predictive probability of distraction model includes developing a graphical representation of the plurality of glance aim points, wherein the graphical representation is used to develop a glance co-occurrence matrix which captures a plurality of glance transitions with the plurality of glance aim points, and further comprising the step of factorizing the glance co-occurrence matrix using a non-negative matrix factorization (NMF) technique to develop a plurality of NMF factors;
determining whether one or more informative glance locations are present in the plurality of glance aim points;
comparing the predictive distraction distribution to a predictive distraction distribution threshold when one or more informative glance locations are present in the plurality of glance aim points; and
alerting the driver when the predictive distraction distribution satisfies or exceeds the predictive distraction distribution threshold.
12. The method of claim 11 , further comprising the step of deriving a plurality of features from the plurality of non-negative matrix factorization (NMF) factors.
13. The method of claim 12 , further comprising the step of comparing the plurality of features to an observation matrix of event probabilities for given glance locations and glance transitions to obtain the predictive distraction distribution.
14. A driver monitoring system for a vehicle, comprising:
a driver facing camera; and
an electronic control unit (ECU) operably coupled to the driver facing camera, wherein the ECU is configured to receive a plurality of glance aim points for a driver of the vehicle from the driver facing camera; input the plurality of glance aim points into a predictive probability of distraction model to obtain a predictive distraction distribution, wherein the predictive probability of distraction model includes a hidden Markov model (HMM); determine whether one or more informative glance locations are present in the plurality of glance aim points; compare the predictive distraction distribution to a predictive distraction distribution threshold when one or more informative glance locations are present in the plurality of glance aim points; and alert the driver when the predictive distraction distribution satisfies or exceeds the predictive distraction distribution threshold.