Autonomous UAV Navigation Using Shared Perception and Object Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous aerial vehicle technologies face challenges in efficiently navigating and tracking objects in complex environments, particularly in real-time, due to limitations in perception and object tracking systems, which affect their ability to generate accurate trajectories and maintain safe flight paths.
Innovation Solution
The implementation of a navigation system that utilizes multiple image capture devices and sensors to generate perception inputs, fuse data from various sources, and employ machine learning models for object detection and tracking, allowing for real-time object tracking and trajectory estimation, enabling the vehicle to autonomously maneuver and adjust its flight path accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple image capture devices and sensors are used for perception inputs, then object detection and tracking precision is improved, but device complexity increases
Solution Approach 1:
The system divides the complex perception task into separate modules: multiple image capture devices for visual input, separate sensor devices for additional sensing, machine learning models for object detection, and tracking systems for following objects. Each component handles a specific aspect of perception, improving overall precision while managing complexity through modular architecture.
Solution Approach 2:
The patent combines data from multiple image capture devices and sensor devices into unified perception inputs. The machine learning models process these combined inputs to detect and track objects, merging multiple data sources to achieve higher measurement precision than any single device could provide alone.
2Measurement precision
If machine learning models are employed for real-time object detection and tracking, then tracking precision is improved, but computational energy consumption increases
Solution Approach 1:
The system performs preliminary actions by pre-training machine learning models offline and preparing tracking algorithms before real-time operation. During actual flight, the pre-trained models process perception inputs more efficiently, reducing real-time computational energy consumption while maintaining high tracking precision.
Solution Approach 2:
The machine learning models are designed to operate autonomously on the autonomous aerial vehicle, processing perception inputs and generating tracking outputs without requiring external computational resources. The system serves its own computational needs through onboard processing, optimizing energy usage for real-time operations.
3Reliability
If the vehicle autonomously maneuvers based on real-time perception inputs, then navigation safety is improved, but response time requirements increase system complexity
Solution Approach 1:
The system implements continuous feedback loops where perception inputs from image capture devices and sensors are processed by machine learning models to detect objects, generate tracking information, and feed this back to the motion planning system. The motion planning system uses this feedback to autonomously maneuver the vehicle, adjusting navigation in real-time to maintain safety while managing control complexity through iterative refinement.
Data Source
AI summary
A technique for user interaction with an autonomous unmanned aerial vehicle (UAV) is described. In an example embodiment, perception inputs from one or more sensor devices are processed to build a shared virtual environment that is representative of a physical environment. The sensor devices used to generate perception inputs can include image capture devices onboard an autonomous aerial vehicle that is in flight through the physical environment. The shared virtual environment can provide a continually updated representation of the physical environment which is accessible to multiple network-connected devices, including multiple UAVs and multiple mobile computing devices. The shared virtual environment can be used, for example, to display visual augmentations at network-connected user devices and guide autonomous navigation by the UAV.