Parking Area Video Stream Sharing for Autonomous Vehicle Blind Spots
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face navigation challenges in areas with 'blind spots' for onboard sensors, such as parking lots, due to fixed and dynamic obstacles that obstruct sensor input, increasing the risk of collisions.
Innovation Solution
The implementation of environment mapping using video stream sharing, where a supervisory service advertises Li-Fi based video streams with metadata to autonomous vehicles, allowing them to identify and avoid obstacles through annotated video streams, leveraging machine learning for optimal stream selection and annotation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If autonomous vehicles use onboard sensors for navigation, then they can operate autonomously in areas with high visibility, but their navigation ability is limited in areas with blind spots such as parking lots
Solution Approach 1:
The patent introduces a supervisory service as an intermediary between the environment and the autonomous vehicle. This service collects video streams from multiple cameras positioned throughout the parking area, processes them to identify obstacles and features, and provides annotated environmental data to the vehicle. This intermediary system enables the vehicle to perceive areas beyond its onboard sensor range, resolving the contradiction between reliable autonomous operation and adaptability to environments with blind spots
2Loss of information
If the supervisory service sends multiple video streams to the autonomous vehicle, then the vehicle gains broader environmental awareness, but the data transmission load and processing complexity increase
Solution Approach 1:
The supervisory service extracts only the essential and relevant information from multiple video streams rather than transmitting the complete raw video data. It identifies and extracts key elements such as obstacle positions, parking space availability, traffic flow patterns, and other critical features, then transmits this extracted information to the autonomous vehicle. This extraction approach maintains environmental information completeness while significantly reducing data transmission load and processing complexity
Solution Approach 2:
The patent segments the environmental information into distinct categories and components, such as static obstacles, dynamic obstacles, parking spaces, and traffic flow. Each segment is processed and annotated separately, allowing the system to manage complex environmental data in organized, manageable units that are easier to transmit and process
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances the autonomous vehicle's situational awareness and navigation accuracy by providing a broader environmental view, reducing the risk of collisions in challenging environments by using annotated video streams to detect and avoid obstacles.
Implementation Method 1
A supervisory service of a parking area may send a light fidelity (Li-Fi) based advertisement indicative of an offer by the supervisory service to send video streams of the parking area to an autonomous vehicle
Data Source
AI summary
In one embodiment, a supervisory service of a parking area may send a light fidelity (Li-Fi) based advertisement indicative of an offer to send video streams of the parking area to an autonomous vehicle. The supervisory service may receive an acceptance of the offer by the autonomous vehicle that includes an identifier for the autonomous vehicle. The supervisory service may identify one or more video streams of the parking area as associated with the autonomous vehicle based in part on a location of the autonomous vehicle in the parking area. The supervisory service may annotate the one or more identified video streams with metadata regarding a feature of the parking area. The supervisory service may send the annotated one or more video streams to the autonomous vehicle, wherein the autonomous vehicle uses the metadata of the annotated one or more video streams to avoid the feature of the parking area.