Paired Drone AI Inference for Low-Bandwidth Stereo Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multi-camera and multi-drone systems fail to provide a full stereoscopic image by combining separately captured images, limiting their ability to achieve optimized vision and object tracking.
Innovation Solution
A network of drones equipped with AI engines and neural network accelerators processes and selectively transmits image data to create a fused stereoscopic vision, allowing for optimized image capture and object tracking by coordinating camera angles and reducing data transmission through AI-driven decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If multiple drones capture images separately and transmit all image data, then complete visual coverage is achieved, but bandwidth congestion increases and transmission efficiency decreases
Solution Approach 1:
The system extracts only the essential visual information (inference outputs such as object detections, key features, or processed image data) from the complete image data captured by multiple drones. This extraction allows the system to achieve complete visual coverage while transmitting only the necessary information, thereby reducing bandwidth consumption and avoiding congestion.
Solution Approach 2:
The patent segments the image processing workflow into two stages: (1) local inference processing performed by each drone to extract key information, and (2) selective transmission of only this processed information to the server. This segmentation enables the system to maintain complete visual coverage while significantly reducing the data transmission load.
2Reliability
If all captured images are transmitted to the server, then complete object tracking is achieved, but transmission time and network load increase
Solution Approach 1:
Each drone performs preliminary inference processing locally to identify objects of interest and extract relevant tracking information before transmission. This preliminary action ensures that object tracking accuracy is maintained while reducing the amount of data that needs to be transmitted, thereby minimizing transmission time and network load.
Solution Approach 2:
The system implements a feedback mechanism where the server receives inference outputs from multiple drones, processes this information to identify objects of interest, and can send back control signals to coordinate the drones' tracking efforts. This feedback loop ensures reliable object tracking while optimizing data transmission efficiency.
3Ease of operation
If multiple drones operate independently, then individual drone autonomy is maintained, but coordinated stereoscopic vision and object tracking are insufficient
Solution Approach 1:
The patent merges the inference outputs from multiple independently operating drones at the server level. Each drone maintains its autonomy and performs local processing, but the server combines these results to create a coordinated multi-drone system that achieves high-quality stereoscopic vision and accurate object tracking, resolving the contradiction between individual autonomy and coordinated performance.
Data Source
AI summary
Disclosed are methods, devices, and computer-readable media for operating paired or grouped drone devices. In one embodiment, a method is disclosed comprising capturing a first image by a camera installed on a first drone device; processing the first image using an artificial intelligence (AI) engine embedded in the first drone device, the processing comprising generating a first inference output; transmitting the first inference output to a second drone device; receiving a second inference output from the second drone device, the second inference output associated with a second image captured by the second drone device; and transmitting the first image to a processor based on the first inference output and second interference output.