Autonomous risk assessment for fallen cargo
A method for detecting fallen cargo, the method may include receiving by a computerized system, sensed information related to driving sessions of multiple vehicles; applying a machine learning process on the sensed information to detect fallen cargo and to classify the fallen cargo to fallen cargo classes; estimating, from the sensed information, an impact of at least some of the fallen cargo classes on a behavior of at least some of the multiple vehicles; and determining, based on the impact, at least one suggested vehicle behavior as a response to a detection of at least some of the fallen cargo classes.
1. A method for detecting fallen cargo, the method comprises:
receiving by a computerized system, sensed information related to driving sessions of multiple vehicles;
applying a machine learning process on the sensed information to detect multiple fallen cargo items and to classify each of the multiple fallen cargo items to a fallen cargo class out of fallen cargo classes;
estimating, from the sensed information, an impact of each fallen cargo class of at least two of the fallen cargo classes on a behavior of each vehicle of at least two of the multiple vehicles; and
determining, based on the impact, at least one suggested vehicle behavior as a response to a detection of each fallen cargo class of the at least two fallen cargo classes;
wherein the estimating of the impact comprises estimating a level of danger associated with each fallen cargo class of at least two of the fallen cargo classes; and
wherein for each vehicle of the at least two of the multiple vehicles, the estimating of the level of danger is responsive to a rate of deviation of the vehicle from a previous propagation direction of the vehicle, wherein the previous propagation direction of vehicle is a direction of propagation of the vehicle before reaching a fallen cargo of a class of at least two of the fallen cargo classes.
2. The method according to claim 1 , wherein the suggested vehicle behavior is an optimal vehicle behavior in relation to a safety of the vehicle.
3. The method according to claim 1 , comprising the applying of the machine learning process on visual information of the sensed information.
4. The method according to claim 1 , wherein the estimating of the impact is based at least in part on motion estimation of the vehicle.
5. The method according to claim 1 , wherein the estimating of the impact is based at least in part on sensed information sensed by at least one vehicle sensor.
6. The method according to claim 1 , wherein the suggested vehicle behavior is an optimal vehicle behavior in relation to a combination of a safety of the vehicle and vehicle fuel consumption.
7. The method according to claim 1 , wherein the suggested vehicle behavior is an optimal vehicle behavior in relation to a combination of a safety of the vehicle and brake deterioration.
8. The method according to claim 1 , comprising suggesting a generation of a driver perceivable alert when a fallen cargo is detected.
9. The method according to claim 1 , comprising generating a profile of a driver of a vehicle that is expected to perform the at least one suggested vehicle behavior, wherein the suggesting of the at least one suggested vehicle behavior is also responsive to the profile of the driver.
10. The method according to claim 1 , wherein the fallen cargo classes comprises a first fallen cargo class and a second fallen cargo class that are defined at different resolutions.
11. A non-transitory computer readable medium that stores instructions for:
receiving by a computerized system, sensed information related to driving sessions of multiple vehicles;
applying a machine learning process on the sensed information to detect multiple fallen cargo items and to classify each of the multiple fallen cargo items to a fallen cargo class out of fallen cargo classes;
estimating, from the sensed information, an impact of each fallen cargo class of at least two of the fallen cargo classes on a behavior of each vehicle of at least two of the multiple vehicles; and
suggesting, based on the impact, at least one suggested vehicle behavior as a response to a detection of each fallen cargo class of the at least two fallen cargo classes
wherein the estimating of the impact comprises estimating a level of danger associated with each fallen cargo class of at least two of the fallen cargo classes; and
wherein for each vehicle of the at least two of the multiple vehicles, the estimating of the level of danger is responsive to a rate of deviation of the vehicle from a previous propagation direction of the vehicle, wherein the previous propagation direction of vehicle is a direction of propagation of the vehicle before reaching a fallen cargo of a class of at least two of the fallen cargo classes.
12. The non-transitory computer readable medium according to claim 11 , wherein the suggested vehicle behavior is an optimal vehicle behavior in relation to a safety of the vehicle.
13. The non-transitory computer readable medium according to claim 11 , that stores instructions for the applying of the machine learning process on visual information of the sensed information.
14. The non-transitory computer readable medium according to claim 11 , wherein the estimating of the impact is based at least in part on motion estimation of the vehicle.
15. The non-transitory computer readable medium according to claim 11 , wherein the estimating of the impact is based at least in part on sensed information sensed by at least one vehicle sensor.
16. The non-transitory computer readable medium according to claim 11 , wherein the suggested vehicle behavior is an optimal vehicle behavior in relation to a combination of a safety of the vehicle and vehicle fuel consumption.
17. The non-transitory computer readable medium according to claim 11 , wherein the suggested vehicle behavior is an optimal vehicle behavior in relation to a combination of a safety of the vehicle and brake deterioration.
18. The non-transitory computer readable medium according to claim 11 , that stores instructions for suggesting a generation of a driver perceivable alert when a fallen cargo is detected.
19. A computerized system that comprises a processor and multiple units that are configured to
receive sensed information related to driving sessions of multiple vehicles;
apply a machine learning process on the sensed information to detect multiple fallen cargo items and to classify each of the multiple fallen cargo items to a fallen cargo class out of fallen cargo classes;
estimate, from the sensed information, an impact of each fallen cargo class of at least two of the fallen cargo classes on a behavior of each vehicle of at least two of the multiple vehicles; and
suggest, based on the impact, at least one suggested vehicle behavior as a response to a detection of each fallen cargo class of the at least two fallen cargo classes
wherein the estimating of the impact comprises estimating a level of danger associated with each fallen cargo class of at least two of the fallen cargo classes; and
wherein for each vehicle of the at least two of the multiple vehicles, the estimating of the level of danger is responsive to a rate of deviation of the vehicle from a previous propagation direction of the vehicle, wherein the previous propagation direction of vehicle is a direction of propagation of the vehicle before reaching a fallen cargo of a class of at least two of the fallen cargo classes.