Model adaptation for autonomous trucking in right of way

1. A computer-implemented method comprising:

receiving, at an interface, data from a railroad system that manages a railroad running parallel to a dedicated roadway whose access is limited to autonomous vehicles;

detecting, by each sensor in a plurality of sensors positioned in a fixed location relative to the dedicated roadway whose access is limited to autonomous vehicles, one or more of the autonomous vehicles in a first field of view on the dedicated roadway, and for each detected autonomous vehicle:

generating sensor data for the detected autonomous vehicle based on detecting the detected autonomous vehicle on the dedicated roadway whose access is limited to autonomous vehicles and the data received at the interface from the railroad system;

generating observational data based on the generated sensor data;

determining an action for the detected autonomous vehicle based on the generated observational data, the action indicative of an action the autonomous vehicle to take while traversing the dedicated roadway whose access is limited to autonomous vehicles; and

transmitting data indicative of the action to the detected autonomous vehicle while the autonomous vehicle traverses the dedicated roadway whose access is limited to autonomous vehicles.

2. The computer-implemented method of claim 1 , wherein:

the plurality of sensors:

in response to generating the observational data based on the generated sensor data, determines, from the sensor data and the observational data, data indicating that a train traversing the railroad has derailed off the railroad and fallen across the dedicated roadway;

determines, from the sensor data and the observational data, data indicating that at least some of the autonomous vehicles traversing the dedicated roadway are on a path to collide with the derailed train; and

in response, transmits, to at least some of the autonomous vehicles, data indicating to stop traversing on the dedicated roadway.

3. The computer-implemented method of claim 1 , wherein:

the plurality of sensors:

in response to generating the observational data based on the generated sensor data, determines, from the sensor data and the observational data, data indicating that a train traversing the railroad has derailed off the railroad and fallen across the dedicated roadway;

determines, from the sensor data and the observational data, data indicating that at least some of the autonomous vehicles traversing the dedicated roadway are on a path to collide with the derailed train; and

in response, transmits, to at least some of the autonomous vehicles, data indicating to a second roadway that runs in parallel to the dedicated roadway to avoid the derailed train.

4. A computer-implemented method comprising:

receiving, at an interface, data from a railroad system that manages a railroad running parallel to a dedicated roadway whose access is limited to autonomous vehicles;

detecting, by each sensor in a plurality of sensors positioned in a fixed location relative to the dedicated roadway whose access is limited to autonomous vehicles, one or more autonomous vehicles in a first field of view on the dedicated roadway, and for each detected autonomous vehicle:

generating sensor data for the detected autonomous vehicle based on detecting the detected autonomous vehicle on the dedicated roadway whose access is limited to autonomous vehicles and the data received at the interface from the railroad system;

generating observational data based on the generated sensor data;

determining, from the observational data, the detected autonomous vehicle is within proximity to an end of the dedicated roadway whose access is limited to autonomous vehicles; and

response to determining the detected autonomous vehicle is within proximity to the end of the dedicated roadway whose access is limited to autonomous vehicles, transmitting, to the detected autonomous vehicle, an instruction indicating the detected autonomous vehicle is to switch to a normal processing mode.

5. The computer-implemented method of claim 4 , further comprising:

in response to determining the detected autonomous vehicle has crossed the end of the dedicated roadway and has entered a roadway that runs in parallel to the dedicated roadway, transmitting, to the detected autonomous vehicle, the instruction indicating the detected autonomous vehicle is to switch to the normal processing mode from an enhanced processing mode.

6. The computer-implemented method of claim 5 , wherein the enhanced processing mode causes the detected autonomous vehicle to utilize (i) the sensor data from the plurality of sensors and (ii) additional sensor data from onboard sensors of the detected autonomous vehicle, for producing actions while traversing the dedicated roadway.

7. A computer-implemented method comprising:

receiving, by one or more processors of an autonomous vehicle, data indicative of route guidance for the autonomous vehicle to traverse;

obtaining, by the one or more processors, sensor data from one or more sensors onboard the autonomous vehicle as the autonomous vehicle traverses a first roadway based on the route guidance;

providing, by the one or more processors, the obtained sensor data to a trained machine-learning model, the trained machine-learning model configured to output a likelihood detection of an event that impacts a dedicated roadway whose access is limited to autonomous vehicles;

receiving, by the one or more processors and from the trained machine-learning model, the likelihood detection of the event that impacts the dedicated roadway whose access is limited to autonomous vehicles;

determining, by the one or more processors, whether the likelihood detection of the event indicates whether the autonomous vehicle will avoid the event that impacts the dedicated roadway whose access is limited to autonomous vehicles;

in response to determining that the autonomous vehicle will avoid the event that impacts the detected roadway whose access is limited to autonomous vehicles, instructing, by the one or more processors, the autonomous vehicle to continue along the route guidance by turning left from the first roadway to the dedicated roadway whose access is limited to autonomous vehicles.

8. The computer-implemented method of claim 7 , further comprising:

in response to turning left from the first roadway to the dedicated roadway, receiving, by the one or more processors and from one or more sensors monitoring the dedicated roadway, data that instructs the autonomous vehicle to utilize an enhanced processing mode for producing actions while traversing the dedicated roadway.

9. The computer-implemented method of claim 7 , further comprising:

in response to turning left from the first roadway to the dedicated roadway, receiving, by the one or more processors and from one or more sensors monitoring the dedicated roadway, data that instructs the autonomous vehicle to switch from a normal processing mode to an enhanced processing mode, wherein the normal processing mode is used by the autonomous vehicle for producing actions while traversing the first roadway and the enhanced processing mode is used by the autonomous vehicle for producing actions while traversing the dedicated roadway.

10. A computer-implemented method comprising:

receiving, by one or more processors of an autonomous vehicle, data indicative of route guidance for the autonomous vehicle to traverse;

obtaining, by the one or more processors, sensor data from one or more sensors onboard the autonomous vehicle as the autonomous vehicle traverses a first roadway based on the route guidance;

providing, by the one or more processors, the obtained sensor data to a trained machine-learning model, the trained machine-learning model configured to output a likelihood detection of an event that impacts the first roadway;

receiving, by the one or more processors and from the trained machine-learning model, the likelihood detection of the event that impacts the first roadway;

determining, by the one or more processors, whether the likelihood detection of the event indicates whether the autonomous vehicle will not avoid the event that impacts the first roadway;

in response to determining the autonomous vehicle will not avoid the event that impacts the first way by traversing on the route guidance, instructing, by the one or more processors, the autonomous vehicle to turn left from the first roadway to a dedicated roadway whose access is limited to autonomous vehicles to avoid the event.

11. The computer-implemented method of claim 10 , further comprising:

in response to turning left from the first roadway to the dedicated roadway, receiving, by the one or more processors and from one or more sensors monitoring the dedicated roadway, data that instructs the autonomous vehicle to utilize an enhanced processing mode for producing actions while traversing the dedicated roadway.

12. The computer-implemented method of claim 10 , further comprising:

in response to turning left from the first roadway to the dedicated roadway, receiving, by the one or more processors and from one or more sensors monitoring the dedicated roadway, data that instructs the autonomous vehicle to switch from a normal processing mode to an enhanced processing mode, wherein the normal processing mode is used by the autonomous vehicle for producing actions while traversing the first roadway and the enhanced processing mode is used by the autonomous vehicle for producing actions while traversing the dedicated roadway.

13. A computer-implemented method comprising:

receiving, by one or more processors of an autonomous vehicle, sensor data from one or more sensors onboard the autonomous vehicle as the autonomous vehicle traverses a dedicated roadway whose access is limited to autonomous vehicles;

receiving, by the one or more processors, additional sensor data from one or more other sensors monitoring the dedicated roadway whose access is limited to autonomous vehicles;

providing, by the one or more processors, (i) the received sensor data from the one or more sensors onboard the autonomous vehicle and (ii) the received additional sensor data from the one or more other sensors monitoring the dedicated roadway whose access is limited to autonomous vehicles, to a trained machine-learning model, the trained machine-learning model configured to output a likelihood detection of an event;

receiving, by the one or more processors and from the trained machine-learning model, the likelihood detection of the event;

determining, by the one or more processors, whether the likelihood detection of the event indicates whether the autonomous vehicle will avoid the event; and

in response to determining whether the autonomous vehicle will likely avoid the event, instructing, by the one or more processors, the autonomous vehicle to continue along the route guidance or to leave the dedicated roadway whose access is limited to autonomous vehicles.

14. The computer-implemented method of claim 13 , wherein the sensor data from the one or more sensors onboard the autonomous vehicle comprises data of an environment proximate to the autonomous vehicle, and the sensor data comprises at least one of video data, audio data, LIDAR data, and radar data.

15. The computer-implemented method of claim 13 , wherein the additional sensor data comprises one or more video data, LIDAR data, radar data, or an instruction that indicates an action for the autonomous vehicle to take while traversing the dedicated roadway.

16. A computer-implemented method comprising:

receiving, at an interface, data from a railroad system that manages a railroad running parallel to a dedicated roadway;

detecting, by each sensor in a plurality of sensors positioned in a fixed location relative to the dedicated roadway, one or more autonomous vehicles in a first field of view on the dedicated roadway, and for each detected autonomous vehicle:

generating sensor data for the detected autonomous vehicle based on detecting the detected autonomous vehicle on the dedicated roadway and the data received at the interface from the railroad system;

generating observational data based on the generated sensor data;

instructing the detected autonomous vehicle to switch to an enhanced processing mode;

waiting a predetermined period of time to receive a confirmation from the detected autonomous vehicle that the detected autonomous vehicle has confirmed switching to the enhanced processing mode; and

in response to receiving the confirmation from the detected autonomous vehicle prior to the predetermine time period elapsing, notifying each other sensor in the plurality of sensors that the detected autonomous vehicle has switched to the enhanced processing mode.

17. The computer-implemented method of claim 16 , wherein the notification to each other sensor in the plurality of sensors that the detected autonomous vehicle has switched to the enhanced processing mode indicates to each other sensor that the detected autonomous vehicle is prepared to receive one or more instructions from each of the sensors in the plurality of sensors that monitor the dedicated roadway.

18. A computer-implemented method comprising:

receiving, at an interface, data from a railroad system that manages a railroad running parallel to a dedicated roadway;

detecting, by each sensor in a plurality of sensors positioned in a fixed location relative to the dedicated roadway, one or more autonomous vehicles in a first field of view on the dedicated roadway, and for each detected autonomous vehicle:

generating sensor data for the detected autonomous vehicle based on the detected autonomous vehicle on the dedicated roadway and the data received at the interface from the railroad system;

generating observational data based on the generated sensor data;

instructing the detected autonomous vehicle to switch to an enhanced processing mode;

waiting a predetermined period of time to receive a confirmation from the detected autonomous vehicle that the detected autonomous vehicle has confirmed switching to the enhanced processing mode;

in response to not receiving the confirmation from the detected autonomous vehicle prior to the predetermine time period elapsing, transmitting a first notification to the detected autonomous vehicle to cease driving the dedicated roadway; and

transmitting a second notification to authorities indicating the detected autonomous vehicle being an at risk vehicle.

19. A computer-implemented method comprising:

receiving, at an interface, data from a railroad system that manages a railroad running parallel to a dedicated roadway;

detecting, by each sensor in a plurality of sensors positioned in a fixed location relative to the dedicated roadway, one or more autonomous vehicles in a first field of view on the dedicated roadway;

detecting, by each of the sensors, one or more events on the first field of view on the dedicated roadway for a predetermined period of time;

in response to detecting the one or more events on the first field of view on the dedicated roadway for the predetermined period of time, determining, by each of the sensors, the one or more detected events block traffic on the dedicated roadway; and

transmitting, by each of the sensors and to the interface that communicates with the railroad system, a notification that indicates the railroad system can increase a number of trains running on the railroad while the one or more detected events block traffic on the dedicated roadway to offset a cost utilization of the dedicated roadway.

20. The computer-implemented method of claim 19 , wherein the interface displays data related to the railroad that traverses in parallel to the dedicated roadway and one or more trains traverse the railroad, the data comprising a number of the one or more trains, a direction of the one or more trains traveling on the railroad, and a number of railroads.