Temporal CNN rear impact alert system

Patent metadata

Field	Value
Publication number	US-11462020-B2
Application number	US-202016733664-A
Country	US
Kind code	B2
Filing date	Jan 3, 2020
Priority date	Jan 3, 2020
Publication date	Oct 4, 2022
Grant date	Oct 4, 2022

Abstract

The present disclosure discloses a system and a method. In an example implantation, the system and the method can receive an image at a first deep neural network, estimate a distance between an object depicted in the image and a vehicle, wherein the first deep neural network estimates the distance, determine whether the estimated distance is greater than a predetermined distance threshold, and generate an alert when the estimated distance is not greater than the predetermined distance threshold.

First claim

What is claimed is:

1. A system comprising a computer including a processor and a memory, the memory including instructions such that the processor is programmed to: receive an image at a first deep neural network; estimate a distance between an object depicted in the image and a vehicle, wherein the first deep neural network estimates the distance; determine whether the estimated distance is greater than a predetermined distance threshold; generate an alert when the estimated distance is not greater than the predetermined distance thresholds determine whether the vehicle has stopped or is moving in an opposite direction; and actuate the vehicle when the vehicle has not stopped and is not moving in the opposite direction.

2. The system of claim 1 , wherein the processor is further programmed to: cause the vehicle to transition from a non-autonomous mode to an autonomous mode.

3. The system of claim 1 , wherein the first deep neural network comprises at least one of a temporal convolutional neural network or a long short-term memory neural network.

4. The system of claim 3 , wherein the processor is further programmed to continue estimating the distance of the object after the object is no longer depicted within the image.

5. The system of claim 1 , wherein the processor is further programmed to: receive the image at a second deep neural network; classify, via the second deep neural network, at least one object depicted within the image; assign an object type to the at least one classified object; and generate an alert based on the object type.

6. The system of claim 5 , wherein the second deep neural network comprises a convolutional neural network.

7. The system of claim 5 , wherein the object type corresponds to a preassigned risk factor corresponding to the classified object.

8. The system of claim 5 , wherein the processor is further programmed to: determine whether the vehicle has stopped or is moving in an opposite direction; and actuate the vehicle when the vehicle has not stopped and is not moving in the opposite direction.

9. The system of claim 8 , wherein the processor is further programmed to: cause the vehicle to transition from a non-autonomous mode to an autonomous mode.

10. A method comprising: receiving an image at a first deep neural network; estimating a distance between an object depicted in the image and a vehicle, wherein the first deep neural network estimates the distance; determining whether the estimated distance is greater than a predetermined distance threshold; generating an alert when the estimated distance is not greater than the predetermined distance thresholds determining whether the vehicle has stopped or is moving in an opposite direction; and actuating the vehicle when the vehicle has not stopped and is not moving in the opposite direction.

11. The method of claim 10 , further comprising: causing the vehicle to transition from a non-autonomous mode to an autonomous mode.

12. The method of claim 10 , wherein the first deep neural network comprises at least one of a temporal convolutional neural network or a long short-term memory neural network.

13. The method of claim 12 , further comprising continuing to estimate the distance of the object after the object is no longer depicted within the image.

14. The method of claim 10 , further comprising: receiving the image at a second deep neural network; classifying, via the second deep neural network, at least one object depicted within the image; assigning an object type to the at least one classified object; and generating an alert based on the object type.

15. The method of claim 14 , wherein the object type corresponds to a preassigned risk factor corresponding to the classified object.

16. A system comprising a computer including a processor and a memory, the memory including instructions such that the processor is programmed to: train a deep neural network with a set of labeled training images, wherein the set of labeled training images comprises at least one training image depicting an object within a field-of-view of a vehicle camera and at least one training label indicating a distance between the object and the vehicle camera; generate an output based on at least one non-labeled training image at the deep neural network, wherein the output is indicative of a distance between an object depicted in the at least one non-labeled training image and an image source; compare the output with ground truth data; and update at least one weight associated with a neuron of the deep neural network.

17. The system of claim 16 , wherein the distance corresponding to the at least one training label is measured by a vehicle ultrasonic sensor.

18. The system of claim 16 , wherein the deep neural network comprises at least one of a temporal convolutional neural network or a long short-term memory neural network.

19. The system of claim 1 , wherein the vehicle is actuated by controlling one or more of vehicle powertrain controller, vehicle steering controller, and vehicle brake controller.

20. The method of claim 10 , wherein the vehicle is actuated by controlling one or more of vehicle powertrain controller, vehicle steering controller, and vehicle brake controller.

Assignees

• Ford Global Tech Llc

Inventors

• Boston Daniel Lewis
• Morris Phillip
• Dennis Michael

Classifications

• B60W30/08 Primary

  Active safety systems} predicting or avoiding probable or impending collision {or attempting to minimise its consequences · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06V20/58 Primary

  Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads · CPC title

• G06F18/2414

  Smoothing the distance, e.g. radial basis function networks [RBFN] · CPC title