Generating real-time driver familiarity index for fine-grained dynamic road scenes

1. A computer-implemented method comprising:

generating a spectral signature describing one or more dynamic objects and a scene layout of a current road scene;

identifying, from among one or more scene clusters included in a familiarity graph associated with a user, a road scene cluster corresponding to the current road scene;

determining a position of the spectral signature relative to other spectral signatures comprising the identified road scene cluster;

generating a familiarity index estimating familiarity of the user with the current road scene based on the position of the spectral signature;

determining an assistance level based on the familiarity index of the user; and

providing one or more of an auditory instruction, a visual instruction, and a tactile instruction to the user via one or more output devices of a vehicle at the determined assistance level.

2. The computer-implemented method of claim 1 , wherein generating the spectral signature describing the one or more dynamic objects and the scene layout of the current road scene includes:

capturing a road scene image of a roadway at a particular point in time using one or more sensors of a vehicle directed to the roadway; and

generating the spectral signature including a signature vector using the road scene image.

3. The computer-implemented method of claim 2 , wherein identifying the road scene cluster corresponding to the current road scene includes:

categorizing the current road scene into a macro-scene category using the spectral signature; and

retrieving the road scene cluster including a plurality of historical spectral signatures associated with the macro-scene category, the plurality of historical spectral signatures including a plurality of historical signature vectors, respectively.

4. The computer-implemented method of claim 3 , wherein the position of the spectral signature is a position of the signature vector, and wherein determining the position of the spectral signature relative to the other spectral signatures includes:

determining the position of the signature vector relative to respective positions of the plurality of historical signature vectors of the road scene cluster.

5. The computer-implemented method of claim 3 , wherein the macro-scene category includes one of an urban category, a rural category, a residential category, a construction zone category, a highway category, and an accident category.

6. The computer-implemented method of claim 1 , further comprising:

updating the road scene cluster to include the spectral signature.

7. The computer-implemented method of claim 1 , wherein generating the familiarity index estimating the familiarity of the user with the current road scene includes:

determining a set of nested familiarity zones for the road scene cluster based on a density of spectral signatures within the road scene cluster; and

determining the familiarity index of the user to the current road scene by determining the position of the spectral signature within the set of nested familiarity zones.

8. The computer-implemented method of claim 1 , wherein the assistance level includes one or more of 1) a frequency for providing the one or more of the auditory instruction, the visual instruction, and the tactile instruction to the user, and 2) a level of detail of the one or more of the auditory instruction, the visual instruction, and the tactile instruction.

9. The computer-implemented method of claim 1 , wherein the current road scene is depicted in a road scene image being captured at a particular point in time, the method further comprises:

determining physiological data associated with the user at the particular point in time;

generating a physiological weight value based on the physiological data; and

augmenting the spectral signature using the physiological weight value.

10. The computer-implemented method of claim 9 , wherein the physiological data associated with the user includes one or more of head movement, eye movement, electrocardiography data, respiration data, skin conductance, and muscle tension of the user at the particular point in time.

11. The computer-implemented method of claim 1 , further comprising:

determining historical navigation data associated with the user;

generating a historical navigation weight value based on the historical navigation data; and

augmenting the spectral signature using the historical navigation weight value.

12. The computer-implemented method of claim 11 , wherein the historical navigation data associated with the user includes one or more of a road name, a road type, road speed information, and time information associated with one or more road segments previously travelled by the user.

13. A system comprising:

one or more processors;

one or more memories storing instructions that, when executed by the one or more processors, cause the system to:

generate a spectral signature describing one or more dynamic objects and a scene layout of a current road scene;

identify, from among one or more scene clusters included in a familiarity graph associated with a user, a road scene cluster corresponding to the current road scene;

determine a position of the spectral signature relative to other spectral signatures comprising the identified road scene cluster;

generate a familiarity index estimating familiarity of the user with the current road scene based on the position of the spectral signature;

determine an assistance level based on the familiarity index of the user; and

provide one or more of an auditory instruction, a visual instruction, and a tactile instruction to the user via one or more output devices of a vehicle at the determined assistance level.

14. The system of claim 13 , wherein to generate the spectral signature describing the one or more dynamic objects and the scene layout of the current road scene includes:

capturing a road scene image of a roadway at a particular point in time using one or more sensors of a vehicle directed to the roadway; and

generating the spectral signature including a signature vector using the road scene image.

15. The system of claim 14 , wherein to identify the road scene cluster corresponding to the current road scene includes:

categorizing the current road scene into a macro-scene category using the spectral signature; and

retrieving the road scene cluster including a plurality of historical spectral signatures associated with the macro-scene category, the plurality of historical spectral signatures including a plurality of historical signature vectors, respectively.

16. The system of claim 15 , wherein the position of the spectral signature is a position of the signature vector, and wherein to determine the position of the spectral signature relative to the other spectral signatures includes:

determining the position of the signature vector relative to respective positions of the plurality of historical signature vectors of the road scene cluster.

17. The system of claim 15 , wherein the macro-scene category includes one of an urban category, a rural category, a residential category, a construction zone category, a highway category, and an accident category.

18. The system of claim 13 , wherein the instructions, when executed by the one or more processors, further cause the system to:

update the road scene cluster to include the spectral signature.

19. The system of claim 13 , wherein to generate the familiarity index estimating the familiarity of the user with the current road scene includes:

determining a set of nested familiarity zones for the road scene cluster based on a density of spectral signatures within the road scene cluster; and

determining the familiarity index of the user to the current road scene by determining the position of the spectral signature within the set of nested familiarity zones.

20. The system of claim 13 , wherein the assistance level includes one or more of 1) a frequency for providing the one or more of the auditory instruction, the visual instruction, and the tactile instruction to the user, and 2) a level of detail of the one or more of the auditory instruction, the visual instruction, and the tactile instruction.

21. The system of claim 13 , wherein the current road scene is depicted in a road scene image being captured at a particular point in time, and wherein the instructions, when executed by the one or more processors, further cause the system to:

determine physiological data associated with the user at the particular point in time;

generate a physiological weight value based on the physiological data; and

augment the spectral signature using the physiological weight value.

22. The system of claim 21 , wherein the physiological data associated with the user includes one or more of head movement, eye movement, electrocardiography data, respiration data, skin conductance, and muscle tension of the user at the particular point in time.

23. The system of claim 13 , wherein the instructions, when executed by the one or more processors, further cause the system to:

determine historical navigation data associated with the user;

generate a historical navigation weight value based on the historical navigation data; and

augment the spectral signature using the historical navigation weight value.

24. The system of claim 23 , wherein the historical navigation data associated with the user includes one or more of a road name, a road type, road speed information, and time information associated with one or more road segments previously travelled by the user.