Quadrotor Drone System Identification via Model-Based Design and In-Flight Sine Wave Injections

Kaputa, Daniel S.; Owens, K.

doi:10.2514/6.2020-1238

Cited by 4 publications

(1 citation statement)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other parameters namely latency and frame rate must be considered though which point to FPGA-based systems overtaking GPU systems for embedded object tracking [17]. YOLBO was initially run on a desktop PC with an NVIDIA 1070ti GPU and then ported to the TX2 due to the straightforward and similar coding framework, however the ultimate goal is to port YOLBO to the Fusion 2 stereo camera and then then to the Fusion 1 UAV [18] for UAV to UAV tracking.…”

Section: Edge Computing a Edge Devicesmentioning

confidence: 99%

YOLBO: You Only Look Back Once–A Low Latency Object Tracker Based on YOLO and Optical Flow

Kaputa

Landy

2021

IEEE Access

Self Cite

View full text Add to dashboard Cite

One common computer vision task is to track an object as it moves from frame to frame within a video sequence. There are a myriad of applications for such capability and the underlying technologies to achieve this tracking are very well understood. More recently, deep convolutional neural networks have been employed to not only track, but also to classify objects as they are tracked from frame to frame. These models can be used in a tracking paradigm known as tracking by detection and can achieve very high tracking accuracy. The major drawback to these deep neural networks is the large amount of mathematical operations that must be performed for each inference which negatively impacts the number of tracked frames per second. For edge applications residing on size, weight, and power limited platforms, such as unmanned aerial vehicles, high frame rate and low latency real time tracking can be an elusive target. To overcome the limited power and computational resources of an edge compute device, various optimizations have been performed to trade off tracking speed, accuracy, power, and latency. Previous works on motion based interpolation with neural networks either do not take into account the latency accrued from camera image capture to tracking result or they compensate for this latency but are bottlenecked by the motion interpolation operation instead. The algorithm presented in this work gains the performance speedup used in previous motion based neural network inference papers and also performs a novel look back operation that is less cumbersome than other competing motion interpolation methods.INDEX TERMS CNN, classifier, detector, neural network, low latency, tracker, UAV, YOLO, look back, drone, image processing.

show abstract

Section: Edge Computing a Edge Devicesmentioning

confidence: 99%

YOLBO: You Only Look Back Once–A Low Latency Object Tracker Based on YOLO and Optical Flow

Kaputa

Landy

2021

IEEE Access

Self Cite

View full text Add to dashboard Cite

show abstract

Data-Driven System Identification of Quadrotors Subject to Motor Delays

Eschmann,

Albani,

Loianno

2024

2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

Aero-Propulsive Modeling for Propeller Aircraft Using Flight Data

Simmons

Gresham

Woolsey

2023

Journal of Aircraft

View full text Add to dashboard Cite

This paper describes methods to identify an integrated propulsion–airframe aerodynamic model and a decoupled propulsion model for fixed-wing aircraft with propellers using flight data. Propulsion aerodynamics and airframe aerodynamics for propeller aircraft are usually modeled separately, which fails to describe unavoidable interaction effects and propeller performance deviations when integrated on an aircraft. Two novel flight test system identification approaches are presented to develop flight dynamics models with improved characterization of propeller aerodynamics compared to conventional methods. Orthogonal phase-optimized multisine inputs are applied to both the control surfaces and propulsion system to generate data with high-quality information content for model identification. Propulsion explanatory variables derived from propeller aerodynamics theory combined with traditional aircraft modeling variables yield accurate aero-propulsive modeling results and provide propeller performance estimates, which are compared to isolated propeller wind tunnel data. An assessment of model adequacy using flight maneuvers withheld from model identification indicates that the models have good prediction capability. The paper describes application of these methods to a small unmanned aircraft, but the methods are generalizable to many propeller-driven aircraft.

show abstract

Quadrotor Drone System Identification via Model-Based Design and In-Flight Sine Wave Injections

Cited by 4 publications

References 12 publications

YOLBO: You Only Look Back Once–A Low Latency Object Tracker Based on YOLO and Optical Flow

YOLBO: You Only Look Back Once–A Low Latency Object Tracker Based on YOLO and Optical Flow

Data-Driven System Identification of Quadrotors Subject to Motor Delays

Aero-Propulsive Modeling for Propeller Aircraft Using Flight Data

Contact Info

Product

Resources

About