Nvidia corporation (20240101118). INTERSECTION DETECTION AND CLASSIFICATION IN AUTONOMOUS MACHINE APPLICATIONS simplified abstract

INTERSECTION DETECTION AND CLASSIFICATION IN AUTONOMOUS MACHINE APPLICATIONS

Organization Name

nvidia corporation

Inventor(s)

Sayed Mehdi Sajjadi Mohammadabadi of Santa Clara CA (US)

Berta Rodriguez Hervas of Santa Clara CA (US)

Hang Dou of Fremont CA (US)

Igor Tryndin of Fremont CA (US)

David Nister of Belleview WA (US)

Minwoo Park of Saratoga CA (US)

Neda Cvijetic of East Palo Alto CA (US)

Junghyun Kwon of San Jose CA (US)

Trung Pham of Santa Clara CA (US)

INTERSECTION DETECTION AND CLASSIFICATION IN AUTONOMOUS MACHINE APPLICATIONS - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240101118 titled 'INTERSECTION DETECTION AND CLASSIFICATION IN AUTONOMOUS MACHINE APPLICATIONS

Simplified Explanation

The abstract of the patent application describes a method of using live perception from sensors of a vehicle to detect and classify intersections in real-time or near real-time. This involves training a deep neural network to compute various outputs related to intersections, such as bounding box coordinates, intersection coverage maps, attributes, distances, and distance coverage maps.

• Leveraging live perception from vehicle sensors to detect and classify intersections in real-time or near real-time.
• Training a deep neural network to compute outputs like bounding box coordinates, intersection coverage maps, attributes, distances, and distance coverage maps for intersections.
• Decoding and post-processing the outputs to determine final locations, distances, and attributes of the detected intersections.

---

1. 1. 1. Potential Applications

This technology can be applied in autonomous vehicles, advanced driver-assistance systems (ADAS), and traffic management systems to improve intersection detection and classification capabilities.

1. 1. 1. Problems Solved

1. Enhances the safety and efficiency of autonomous vehicles by accurately detecting intersections in real-time. 2. Improves the performance of ADAS by providing precise information about intersections to assist drivers.

1. 1. 1. Benefits

1. Increased road safety by ensuring timely detection of intersections. 2. Enhanced driving experience through improved intersection classification. 3. Better traffic flow management by accurately identifying intersections.

1. 1. 1. Potential Commercial Applications

1. Automotive industry for autonomous vehicles and ADAS development. 2. Traffic management companies for optimizing traffic flow at intersections.

1. 1. 1. Possible Prior Art

Prior art may include similar patents or research papers on intersection detection using sensor data and deep learning algorithms.

---

1. 1. 1. 1. Unanswered Questions

1. 1. 1. How does this technology handle complex intersection scenarios?

The article does not delve into the specifics of how the deep neural network copes with intricate intersection layouts or challenging traffic conditions.

1. 1. 1. What are the limitations of using live perception from vehicle sensors for intersection detection?

The abstract does not address any potential drawbacks or challenges associated with relying on sensor data for real-time intersection classification.

Original Abstract Submitted

in various examples, live perception from sensors of a vehicle may be leveraged to detect and classify intersections in an environment of a vehicle in real-time or near real-time. for example, a deep neural network (dnn) may be trained to compute various outputs—such as bounding box coordinates for intersections, intersection coverage maps corresponding to the bounding boxes, intersection attributes, distances to intersections, and/or distance coverage maps associated with the intersections. the outputs may be decoded and/or post-processed to determine final locations of, distances to, and/or attributes of the detected intersections.