Peer-to-peer vehicular provision of artificially intelligent traffic analysis
Systems/techniques that facilitate peer-to-peer vehicular provision of artificially intelligent traffic analysis are provided. In various embodiments, a system can be onboard a first vehicle. In various aspects, the system can capture, via one or more cameras, one or more microphones, or one or more other sensors of the first vehicle, roadside data associated with a road on which the first vehicle is traveling. In various instances, the system can execute a deep learning neural network on the roadside data, wherein the deep learning neural network can produce as output a classification label, a bounding box, or a pixel-wise segmentation mask localizing an unsafe driving condition along the road. In various cases, the system can transmit, via a peer-to-peer communication link to a second vehicle traveling on the road, an electronic alert based on the unsafe driving condition.
1 . A system of a first vehicle, the system comprising:
a memory that stores computer executable components; and
a processor that executes at least one of the computer-executable components that:
captures, via at least one sensor of the first vehicle, raw data associated with a road on which the first vehicle is traveling, wherein the at least one sensor comprises at least one of an image sensor or a sound sensor;
generates, using a first deep learning neural network of the first vehicle on the raw data, an output comprising a classification label, a bounding box, and a pixel-wise segmentation mask localizing an unsafe driving condition along the road;
transmits, via a peer-to-peer communication link between the first vehicle and a second vehicle traveling on the road, an electronic alert to the second vehicle based on the unsafe driving condition, wherein the electronic alert comprises the raw data, the classification label, the bounding box, and the pixel-wise segmentation mask localizing the unsafe driving condition along the road; and
receiving, via the peer-to-peer communication link, from the second vehicle, based on an output generated from a second deep learning neural network of the second vehicle using the raw data, the classification label, the bounding box, and the pixel-wise segmentation mask from the first vehicle, an indication of whether the unsafe driving condition was verified by the second deep learning neural network of the second vehicle.
2 . The system of claim 1 , wherein the electronic alert comprises a geolocation stamp recorded by a global positioning sensor or a wi-fi positioning sensor of the first vehicle, and wherein the geolocation stamp indicates a roadside position along the road at which the unsafe driving condition is located.
3 . The system of claim 2 , wherein the electronic alert indicates that the unsafe driving condition is in one or more lanes of the road at the roadside position indicated by the geolocation stamp, and wherein the electronic alert recommends that the one or more lanes be avoided when passing the roadside position.
4 . The system of claim 2 , wherein the electronic alert indicates that the unsafe driving condition is in or outside a shoulder of the road at the roadside position indicated by the geolocation stamp, and wherein the electronic alert recommends that caution be exercised when passing the roadside position.
5 . The system of claim 1 , wherein the electronic alert comprises a timestamp recorded by a clock of the first vehicle, and wherein the timestamp indicates when the unsafe driving condition was localized by the first vehicle.
6 . The system of claim 1 , wherein the at least one sensor comprises the audio sensor, the raw data comprises sounds associated with the road, and wherein the first deep learning neural network, using the sounds:
determines that an object that has fallen onto the road; and
based on a sound associated with the object falling onto the road, identifies a type of the object, wherein the unsafe driving condition comprises the object that has fallen on the road.
7 . The system of claim 1 , wherein the electronic alert comprises an identifier associated with the first vehicle, and wherein the identifier comprises at least one of a license plate of the first vehicle, a first name of an owner of the first vehicle, a first address of the owner of the first vehicle, a second name of a current driver of the first vehicle, or a second address of the current driver of the first vehicle.
8 . The system of claim 1 , wherein the object is a fallen rock, a fallen power line, or a fallen tree branch.
9 . The system of claim 1 , wherein the object is furniture that has fallen off of another vehicle on the road or a damaged tired that has fallen off of another vehicle on the road.
10 . A computer-implemented method, comprising:
capturing, by a system of a first vehicle, via at least one sensor of the first vehicle, data associated with a road on which the first vehicle is traveling, wherein the at least one sensor comprises at least one of an image sensor or a sound sensor;
generating, by the system, using a first deep learning neural network of the first vehicle on the raw data, an output comprising a classification label, a bounding box, and a pixel-wise segmentation mask localizing an unsafe driving condition along the road;
transmitting, by the system, via a peer-to-peer communication link between the first vehicle and a second vehicle traveling on the road, an electronic alert to the second vehicle based on the unsafe driving condition, wherein the electronic alert comprises the raw data, the classification label, the bounding box, and the pixel-wise segmentation mask localizing the unsafe driving condition along the road; and
receiving, by the system, via the peer-to-peer communication link, from the second vehicle, based on an output generated from a second deep learning neural network of the second vehicle using the raw data, the classification label, the bounding box, and the pixel-wise segmentation mask from the first vehicle, an indication of whether the unsafe driving condition was verified by the second deep learning neural network of the second vehicle.
11 . The computer-implemented method of claim 10 , wherein the electronic alert comprises a geolocation stamp recorded by a global positioning sensor or a wi-fi positioning sensor of the first vehicle, and wherein the geolocation stamp indicates a roadside position along the road at which the unsafe driving condition is located.
12 . The computer-implemented method of claim 11 , wherein the electronic alert indicates that the unsafe driving condition is in one or more lanes of the road at the roadside position indicated by the geolocation stamp, and wherein the electronic alert recommends that the one or more lanes be avoided when passing the roadside position.
13 . The computer-implemented method of claim 11 , wherein the electronic alert indicates that the unsafe driving condition is in or outside a shoulder of the road at the roadside position indicated by the geolocation stamp, and wherein the electronic alert recommends that caution be exercised when passing the roadside position.
14 . The computer-implemented method of claim 10 , wherein the electronic alert comprises a timestamp recorded by a clock of the first vehicle, and wherein the timestamp indicates when the unsafe driving condition was localized by the first vehicle.
15 . The computer-implemented method of claim 10 , wherein the electronic alert comprises an identifier associated with the first vehicle, and wherein the identifier comprises at least one of a license plate of the first vehicle, a first name of an owner of the first vehicle, a first address of the owner of the first vehicle, a second name of a current driver of the first vehicle, or a second address of the current driver of the first vehicle.
16 . A computer program product for facilitating peer-to-peer vehicular provision of artificially intelligent traffic analysis, the computer program product comprising a non-transitory computer-readable memory having program instructions embodied therewith, the program instructions executable by a processor of a first vehicle to cause the processor to:
capture, via at least one sensor of the first vehicle, raw data associated with a road on which the first vehicle is traveling, wherein the at least one sensor comprises at least one of an image sensor or a sound sensor;
generate, using first deep learning neural network of the first vehicle on the raw data, an output comprising a classification label, a bounding box, and a pixel-wise segmentation mask localizing an unsafe driving condition along the road;
transmit, via a peer-to-peer communication link between the first vehicle and a second vehicle traveling on the road, an electronic alert to the second vehicle based on the unsafe driving condition, wherein the electronic alert comprises the raw data, the classification label, the bounding box, and the pixel-wise segmentation mask localizing the unsafe driving condition along the road; and
receive, via the peer-to-peer communication link, from the second vehicle, based on an output generated from a second deep learning neural network of the second vehicle using the raw data, the classification label, the bounding box, and the pixel-wise segmentation mask from the first vehicle, an indication of whether the unsafe driving condition was verified by the second deep learning neural network of the second vehicle.
17 . The computer program product of claim 16 , wherein the electronic alert comprises a geolocation stamp recorded by a global positioning sensor or a wi-fi positioning sensor of the first vehicle, and wherein the geolocation stamp indicates a roadside position along the road at which the unsafe driving condition is located.
18 . The computer program product of claim 17 , wherein the electronic alert indicates that the unsafe driving condition is in one or more lanes of the road at the roadside position indicated by the geolocation stamp, and wherein the electronic alert recommends that the one or more lanes be avoided when passing the roadside position.
19 . The computer program product of claim 17 , wherein the electronic alert indicates that the unsafe driving condition is in or outside a shoulder of the road at the roadside position indicated by the geolocation stamp, and wherein the electronic alert recommends that caution be exercised when passing the roadside position.
20 . The computer program product of claim 16 , wherein the electronic alert comprises a timestamp recorded by a clock of the first vehicle, and wherein the timestamp indicates when the unsafe driving condition was localized by the first vehicle.