Radar-Based UAV Tracking With Bird-Clutter Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Radar systems struggle to effectively track slower-moving objects like unmanned aerial vehicles (UAVs) due to interference from similarly sized and speeded birds, leading to data overload and inefficient operations.
Innovation Solution
A radar-based tracking system coupled with a tracker and classifier that utilizes unsupervised learning to discriminate between UAVs and birds, employing a covariance matrix rotation and hypercube classification to distinguish between different flight patterns and characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If radar sensitivity is increased to detect smaller and slower moving objects like UAVs, then detection capability for UAVs is improved, but the radar system detects excessive extraneous data from birds and other slow-moving objects
Solution Approach 1:
The patent extracts and removes extraneous data (birds and clutter) from the radar detection output through a specialized tracker and classifier system. The tracker is configured to follow slow-moving objects and the classifier discriminates between UAVs and birds, effectively separating the useful signal from the overwhelming clutter background.
Solution Approach 2:
The patent introduces an intermediary classification system between the radar detector and the operator. This intermediary consists of the tracker and classifier components that automatically process and filter the raw radar data, presenting only relevant UAV information to the operator while blocking out bird and clutter returns.
2Measurement precision
If radar sensitivity is increased to detect UAVs, then detection capability is improved, but operational efficiency deteriorates due to data overload
Solution Approach 1:
The system performs self-service by automatically classifying and filtering radar returns without requiring operator intervention. The unsupervised learning model continuously adapts to differentiate UAVs from birds and clutter, allowing the system to maintain high detection capability while preserving operational efficiency through automated processing.
Solution Approach 2:
The patent changes the operational parameters of the radar system by introducing a specialized tracker configured for slow-moving objects and a classifier that dynamically adjusts discrimination thresholds. These parameter changes enable the system to maintain high sensitivity for UAV detection while automatically filtering out clutter based on motion characteristics and pattern recognition.
3Measurement precision
If radar tracks all slow-moving objects, then detection completeness is improved, but discrimination between UAVs and birds becomes difficult
Solution Approach 1:
The patent segments the radar detection process into distinct functional components: a detector that captures all slow-moving objects, a tracker that follows their trajectories, and a classifier that segments the results into UAVs versus birds/clutter. This segmentation allows complete detection while systematically separating useful information from extraneous data.
Solution Approach 2:
The classifier employs unsupervised learning that dynamically adapts to changing environmental conditions and clutter patterns. The system continuously learns from new data to improve discrimination between UAVs and birds, maintaining detection completeness while progressively improving the signal-to-clutter ratio through adaptive parameter adjustment.
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
The system provides accurate and efficient tracking of UAVs by automatically differentiating them from birds and other clutter, enhancing operational efficiency and reducing operator intervention.
Implementation Method 1
Radar has long been employed as an effective tool for identifying and tracking aircraft during flight.
Data Source
AI summary
A system and method for identifying slow-moving and smaller flying objects using one or more radar based sensors is provided. In one or more examples, a radar system can be configured to generate plot data corresponding to flying objects in a given airspace. A tracker can be configured to receive the plot data, and can be configured to generate one or more tracks. The one or more tracks generated by the tracker can then be inputted into a classifier that is configured to distinguish unmanned aerial vehicle (UAV) traffic from birds that are flying in the airspace. In one or more examples, the classifier can generate an N-dimensional hypercube, with each dimension of the hypercube pertaining to a specific attribute of the flying objects. Each track can be converted into a data point within the hypercube and the data points can be clustered to determine whether the track belongs to a bird or a UAV.