Auditory Obstacle Detection for Occluded Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in detecting obstacles that are obscured from the line of sight of imaging sensors, as existing systems rely heavily on visual data and may fail to identify vehicles or objects hidden by occluding objects.
Innovation Solution
The use of auditory data from microphones, combined with image data, to identify potential obstacles through an audio detection module that preprocesses and classifies audio features using machine learning models, allowing for the detection of vehicles and other objects not visible to cameras, and enabling collision avoidance measures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If imaging sensors (video cameras, RADAR, LIDAR) are used to detect obstacles, then the system can identify visible obstacles and navigate, but it fails to detect obstacles obscured by occluding objects
Solution Approach 1:
The patent segments the detection task into multiple sensory modalities: visual detection for visible obstacles and auditory detection for hidden obstacles. The audio detection module independently processes acoustic information to identify obstacles that are occluded from visual sensors, creating parallel detection pathways that compensate for each other's limitations.
Solution Approach 2:
The patent introduces audio signals as an intermediary medium to detect hidden obstacles. Sound waves can penetrate or绕 around occluding objects that block visual sensors, serving as a mediator that carries information about obstacles invisible to cameras, RADAR, and LIDAR.
2Device complexity
If the system relies heavily on visual data for obstacle identification, then image processing is simplified, but detection capability is reduced for obscured objects
Solution Approach 1:
The patent makes the obstacle detection system multi-functional by equipping it with both visual and auditory detection capabilities. The same navigation system processes both image data and audio data to identify obstacles, allowing a single system to perform multiple detection functions rather than requiring separate specialized systems.
Solution Approach 2:
The patent adds an auditory dimension to the traditionally visual obstacle detection problem. By incorporating audio frequency data alongside visual data, the system transitions from single-dimensional (visual only) detection to multi-dimensional detection, enabling identification of obstacles through multiple physical domains.
Data Source
AI summary
A controller for an autonomous vehicle receives audio signals from one or more microphones. The outputs of the microphones are pre-processed to enhance audio features that originated from vehicles. The outputs may also be processed to remove noise. The audio features are input to a machine learning model that classifies the source of the audio features. For example, features may be classified as originating from a vehicle. A direction to a source of the audio features is determined based on relative delays of the audio features in signals from multiple microphones. Where audio features are classified with an above-threshold confidence as originating from a vehicle, collision avoidance is performed with respect to the direction to the source of the audio features.