Cooperative 3D Perception for AAVs Under Wireless Data Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D sensors on autonomous air vehicles (AAVs) face limitations due to occlusions, sensing range, and extreme weather, leading to reduced visibility and data traffic overload, with current data sharing methods failing to optimize information transmission and causing loss of important sensory data.
Innovation Solution
An intelligent cooperative perception framework utilizing information-centric networking (ICN) and deep reinforcement learning to dynamically select and compress sensory data, combined with recurrent neural networks (RNN) for efficient data exchange and fusion, ensuring accurate and timely object detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If AAVs share all detected sensory data to improve cooperative perception accuracy, then object detection precision is improved, but network data traffic exceeds wireless network capacity
Solution Approach 1:
The patent extracts and transmits only the most critical sensory data elements (object type, location, velocity) while discarding redundant detailed sensory information. This selective extraction allows the system to maintain accurate cooperative perception by sharing essential data without overwhelming network capacity with complete sensory streams.
Solution Approach 2:
The patent implements partial action by transmitting a subset of detected object data rather than complete sensory information. The system prioritizes transmitting data about objects that pose potential threats or are relevant to path planning, achieving effective cooperative perception with reduced data volume that fits within network constraints.
2Quantity of substance
If AAVs transmit only essential object information to reduce network load, then network capacity is preserved, but important sensory details are lost
Solution Approach 1:
The patent applies local quality by differentiating the level of data transmission based on object characteristics and contextual importance. Critical objects (those posing threats or requiring attention) receive detailed data transmission, while less important objects receive summarized information, optimizing the balance between information completeness and network efficiency.
Solution Approach 2:
The patent dynamically changes data transmission parameters (data granularity, transmission frequency, compression level) based on network conditions, object priority, and operational context. This adaptive parameter adjustment ensures essential sensory information is preserved while adapting data volume to network capacity constraints.
3Productivity
If AAVs process and compress all sensory data locally before transmission to reduce network traffic, then data transmission efficiency is improved, but on-board computing resources are overwhelmed
Solution Approach 1:
The patent segments the data processing function across multiple levels: basic filtering and prioritization occur on-board the AAV, while more complex compression and optimization are performed at the network edge or cloud infrastructure. This segmentation reduces on-board computational burden while maintaining overall processing efficiency.
Solution Approach 2:
The patent performs preliminary data filtering, prioritization, and basic compression on-board before transmission, preparing data in advance to reduce the burden during actual transmission. This preliminary processing ensures only essential data enters the transmission pipeline, improving efficiency without requiring excessive on-board computational resources during critical operations.
Data Source
AI summary
An intelligent cooperative perception system for autonomous air vehicles (AAVs) comprises at least two AAVs, each of the at least two AAVs being provided with at least one object detection sensor. The at least two AAVs are configured to perform: an information-centric networking (ICN) scheme configured for flexible and efficient communications among the at least two AAVs and infrastructure sensors; a deep reinforcement learning scheme including selecting sensory data to be transmitted, selecting data compression format, mitigating wireless network load, and reducing network latency; an efficient real-time compression scheme of 3D point cloud streams based on recurrent neural network (RNN) algorithms and including significantly reducing data amount exchanged among the at least two AAVs, network load and delay while maintaining accurate cooperative perception; and an effective point cloud fusion scheme including compensating network latency and accurately fusing sensed data from the at least two AAVs and the infrastructure sensors.