Detecting physical threats approaching a vehicle
First Claim
1. A method for use at a vehicle, the method comprising:
- determining that the path of a first object is likely to cause the first object and a corresponding second object to travel near the vehicle based on filtered sensor data from one or more sensors externally mounted on the vehicle;
providing the filtered sensor data as input to a neural network;
receiving a threat classification from the neural network classifying the first object and the corresponding second object collectively as a non-vehicular threat to an occupant of the vehicle;
indicating the threat classification in the vehicle cabin; and
automatically maneuvering the vehicle to get away from the first object and the second object.
1 Assignment
0 Petitions
Accused Products
Abstract
The present invention extends to methods, systems, and computer program products for detecting physical threats approaching a vehicle. External sensors on a vehicle capture the environment around the vehicle. Approaching targets detected by the external sensors can be fed into a neural network to recognize and/or classify approaching targets as potential threats. Tracking mechanisms (e.g., Kalman filters, Particle filters, etc.) can leverage temporal information to determine if a threat is approaching a vehicle. When an approaching threat is detected, a vehicle can activate one or more counter measures to deter the threat. When a vehicle includes autonomous driving capabilities, counter measures can include automatically attempting to drive away from an approaching threat.
-
Citations
20 Claims
-
1. A method for use at a vehicle, the method comprising:
-
determining that the path of a first object is likely to cause the first object and a corresponding second object to travel near the vehicle based on filtered sensor data from one or more sensors externally mounted on the vehicle; providing the filtered sensor data as input to a neural network; receiving a threat classification from the neural network classifying the first object and the corresponding second object collectively as a non-vehicular threat to an occupant of the vehicle; indicating the threat classification in the vehicle cabin; and automatically maneuvering the vehicle to get away from the first object and the second object. - View Dependent Claims (2, 3, 4, 5, 6, 16)
-
-
7. A method for use at a vehicle, the method comprising:
-
determining that the path of an object is likely to cause the object to travel near to the vehicle, including; using one or more sensors mounted to the vehicle to monitor an area in proximity to the vehicle for approaching objects; and filtering data from the one or more sensors on a heterogeneous computing platform at the vehicle to determine that the object has a speed and direction indicative of the object approaching space occupied by the vehicle; providing the filtered data for the approaching object as input to a neural network; receiving an indication from the neural network that the approaching object and a second object associated with the approaching object collectively represent a non-vehicular threat to a vehicle occupant based at least in part on identification of the second object; and activating counter measures at the vehicle to address the non-vehicular threat, including automatically attempting to maneuver the vehicle to take the vehicle occupant away from both the approaching object and the second object. - View Dependent Claims (8, 9, 10, 11, 12, 13, 14, 15)
-
-
17. A vehicle, the vehicle comprising:
-
one or more externally mounted sensors for monitoring an area in proximity to the vehicle; one or more processors; system memory coupled to one or more processors, the system memory storing instructions that are executable by the one or more processors; the one or more processors configured to execute the instructions stored in the system memory to; determine that the path of an object is likely to cause the object and a second object to travel near to the vehicle, including; use the one or more externally mounted sensors to monitor the area in proximity to the vehicle for approaching objects; and filter data from the one or more sensors to determine that the object has a speed and direction indicative of the object approaching space occupied by the vehicle; provide the filtered data as input to a neural network; receive a threat probability from the neural network indicating a probability that the object and the second object collectively represent a non-vehicular threat to a vehicle occupant; and control one or more vehicle components, wherein the one or more vehicle components maneuver the vehicle to get away from both the object and the second object. - View Dependent Claims (18, 19, 20)
-
Specification