Travis D. Fields

Flight Testing of a 1-DOF Variable Drag Autonomous Descent Vehicle

This paper details the hardware development and testing of an autonomous descent vehicle. The developed vehicle utilizes a circular parachute and wind data to control the landing location. The benefit of this system over alternative precision aerial delivery techniques is the envisioned ability to use the large inventory of circular parachutes already in use for uncontrolled cargo and personnel deliveries with minimal training or system modifications. Parachute control is obtained via reversibly reefing of the canopy, thereby modifying the descent speed. For this study, the parachute size is assumed to be constant for the remainder of the descent; however, the desired parachute size computation is periodically updated to assist in reducing landing errors due to inaccurate wind data. A small mechanical reeling system has been developed, comprising a microcomputer, RF modem, electronic speed controller, and an electric motor. The hardware is coupled with a quarter-spherical canopy with four suspension lines, similar to those used in automotive drag racing. The combined weight of the parachute-payload system is 53.5N (12.0lb). Flight testing was conducted using a small single engine aircraft (Cessna 172), with preliminary flight testing conducted using an Arcturus T-20 UAV. Release ceilings were approximately 3050m (10,000ft) MSL. Typically, dropsondes were used to collect predicted wind for the descent vehicle. The time needed to collect the wind data, upload it into the descent vehicle software, takeoff, and reach the desired deployment location was approximately two hours. Preliminary testing of the parachute-payload system was performed to determine the appropriate control gains for the motor angle control routine to achieve the desired descent rate. Release altitudes were between 450m (1,500ft) and 610m (2,000ft) AGL. Using Zeigler-Nichols gain tuning rules and an experimental step response, gains were determined for both a Proportional-Integral (PI) controller and a Proportional-Integral-Derivative (PID) controller. Additional testing was conducted to verify the ability of these control gains to achieve a desired descent speed prior to flight testing the full path planning system and control algorithm. Flight testing results demonstrate the ability for the autonomous descent vehicle (ADV) to successfully navigate towards a target line segment when using accurate wind prediction data. As previously published results have noted, when the predicted wind data is inaccurate, the vehicle is not always capable of improving the landing location accuracy compared to an uncontrolled parachute. Additional considerations in developing a descent rate control system for use in circular parachutes are also presented.

Time-Varying Descent Rate Control Strategy for Circular Parachutes

This paper presents a time-varying control methodology for a variable-sized circular parachute to reach a target landing location. A trajectory is calculated for the immediate control horizon using wind forecast data. To create a parachute–payload trajectory, a three-degree-of-freedom kinematic model is developed. Using this, the performance envelope is determined, revealing the potential target range of the system during a descent. Next, this model is further developed into a control methodology to determine the necessary descent rate, to reach the desired landing target, to be controlled via parachute size manipulation. Finally, simulation results are presented to validate the control scheme. Various release locations were simulated with paired uncontrolled/controlled parachute descents from within the performance envelope. Results demonstrate the feasibility of the system, with controlled parachute descents actively navigating toward the target. With accurate wind data, the vehicle can overcome release...

In-flight Landing Location Predictions using Ascent Wind Data for High Altitude Balloons

This paper presents a hardware and software system in which the landing location of a balloon–payload system is continually predicted and improved. Initial prediction parameters are corrected by estimating flight parameters using GPS data. The focus of this study is on small (less than 10 kg payload) balloon systems. For these systems, flight missions typically have a duration of a few hours (minimal loitering at altitude) and descend after the balloon bursts at an altitude in the range of 15-30 km. Prior to launch, weather predictions are typically used to predict the flight path and landing location of the balloon-payload system. The prediction accuracy greatly depends on the stability of wind data, as well as the accuracy in predicting the balloon ascent speed, balloon burst altitude, and parachute descent speed. Differences between the predicted and actual landing locations of 60 km are not atypical. Rather than relying on the pre-flight predictions, the system developed here measures ascent/descent speeds and wind speed during the balloon’s ascent. Unlike the pre-flight predictions, the system provides chase and recovery personnel with an accurate prediction of the landing location that is periodically updated as the mission progresses. During ascent, GPS data is logged to provide both actual ascent speed and up-to-date spatial and temporal wind data. The ascent speed along with the logged wind data and initial estimates of the parachute-payload flight characteristics (mass, parachute size, etc.) are used to correct the flight model up to the time of balloon burst. During descent, the actual descent speed is used to further enhance predictions as the mission progresses. Results calculated from post-processing actual balloon flight GPS data validate the methodology developed. Prediction accuracy is improved from an average predicted landing location error of 36% of the traveled range prior to launch. At the balloon burst location, prediction quality improves to 3.5% of the total range traveled on average. Additionally, real-time, in-flight prediction results verify the ability to perform the in-flight prediction updates on-board a balloon using a microprocessor and off-the-shelf radio communication equipment.

Path Planning of a Circular Parachute Using Descent Rate Control

This paper presents a time-varying control methodology for a variable-sized circular parachute to reach a target landing location. A trajectory is calculated for the immediate control horizon using wind forecast data. In order to create a parachute–payload trajectory, a 3–DOF kinematic model is developed. Using this, the performance envelope is determined, revealing the potential target range of the system throughout a descent. Next, this model is extended to develop a control methodology to determine the descent rate, via parachute size manipulation, needed to reach the desired landing target. Finally, simulation results are presented to validate the control scheme. Various release locations were simulated with paired uncontrolled/controlled parachute descents from within the performance envelope. Results demonstrate the feasibility of the system, with controlled parachute descents navigating towards the target. With accurate wind data the vehicle can overcome release location errors as well as vehicle uncertainties and perform significantly better than an uncontrolled parachute.

Lower Stratospheric Deployment Testing of a Ram-Air Parafoil System

This work continues the flight testing of ram-air parafoils from high altitude weather balloons. Previous work revealed the major challenge of a zero-speed deployment from a balloon combined with the low air density environment. If the parafoil fails to inflate upon initial release at high altitude, tumbling and tangling occur almost instantaneously, thus preventing the parachute from inflating even after it descends into the lower, denser atmosphere. This paper describes further balloon flight testing of two different canopy size systems conducted to collect performance data in the rarely tested high altitude flight regime. It describes problems encountered during testing and considerations for improving the reliability of a ram-air parafoil released in a low-density zero dynamic pressure environment.

Development of a Coupled Dropsonde-Autonomous Descent Vehicle System

This paper describes the benefits of using a dropsonde to automatically provide real time wind data to an autonomous descent vehicle (ADV). Two major sources of wind error are identified: wind errors resulting from large scale temporal and/or spatial differences between forecast and actual wind columns (Type A), and small scale wind fluctuations and/or measurement inaccuracies typically in the form of measurement noise (Type B). An autonomously coupled dropsonde-ADV system was developed in which a dropsonde is released preceding the descent vehicle, with enough lead time to descend to the ground prior to release of the ADV, and automatically telemeters wind data to the ADV, minimizing the more severe Type A wind errors. Simulations of the descent vehicle’s path (computed using actual wind data) are presented to compare the coupled dropsonde-ADV system to other ADV deployment scenarios. The simulations show that the use of a dropsonde to collect wind data can significantly reduce errors in the landing location of the ADV (89% in our simulations) as compared to an ADV that uses only forecasted wind data from prior to the descent. These results quantify the value of using spatiallyand temporallyrelevant wind data, and the potential benefits that can result from incorporating a coupled dropsonde-ADV system.

The use of a steerable single-actuator cruciform parachute for targeted payload return

The concept of using an autonomous HAHO (high altitude, high-opening) parafoil-based system as a solution to the final descent phase of an on-demand International Space Station (ISS) sample return concept was tailored to meet specific constraints defined by the NASA Ames Research Center SPQR (Small Payload Quick-Return) study. Difficulties in consistently and safely deploying ram-air parafoils at high altitudes needs to be suitably addressed prior to utilization in space sample return applications. This paper presents an alternative to the ram-air parachute SPQR concept. Specifically, the feasibility of using a cruciform-type canopy design modified to enable a limited steering capability, similar to what has been accomplished with the traditional round canopy-based Affordable Guided Aerial Delivery System (AGAS). In contrast to the AGAS system the cruciform canopy approach requires only a single actuator, providing a wider range of possible control actions and therefore a more robust guidance technique. Preliminary design and flight test results are provided from initial low-altitude testing of the cruciform canopy-based SPQR-compatible system.

Development and Feasibility of a Non-Invasive, Wireless Parachute Load Distribution Measuring System

In order to improve the accuracy of low-cost circular parachute systems, it is necessary to understand the physical effects present during descent. Historically, parachute loading has been assumed to be relatively symmetric regardless of the canopy mode (inflating or steady state), therefore, typical testing of the parachute loads consists of only a single load cell placed at the suspension line confluence point. Recent investigations have shown significant asymmetric suspension line loads during parachute drop testing; however, only four load cells were used in the investigation. Asymmetric loading impacts both the stresses imposed on suspension lines, as well as the repeatability of parachute performance between tests. This paper discusses the development and preliminary flight testing of a wireless, noninvasive load-cell instrumentation suite capable of capturing loading information on each individual suspension line. The system incorporates low-cost, custom designed strain gauge load cells that clamp onto existing suspension lines. All information is telemetered wirelessly to a single microcontroller unit in the cargo payload. Static and dynamic calibration experiments were conducted to quantify the accuracy of the load cells prior to drop testing. Drop testing was conducted in a controlled, indoor environment with a small-scale 0.91 m (36 inch) circular parachute and a payload weight of 10.2 N (2.29 lbs.). Testing results demonstrate the feasibility of a wireless system; however, additional testing and verification is required prior to conducting a rigorous parachute testing regime.

Flight controller learning based on real-time model estimation of a quadrotor aircraft

Aircraft prototyping and modeling is usually associated with resource expensive techniques and significant post-flight analysis. The NASA Learn-To-Fly concept targets the replacement of the conventional ground-based aircraft development and prototyping approaches with an efficient real-time paradigm. The work presented herein describes a learning paradigm of a quadcopter unmanned aircraft that utilizes real-time flight data. Closed-loop parameter estimation of a highly collinear model terms such as those found on a quadrotor is challenging. Using phase optimized orthogonal multisine input maneuvers, collinearity of flight data decreases leading to fast and accurate convergence of the Fourier transform regression estimator. The generated models are utilized to reconfigure a nonlinear dynamic inversion controller in normal, failure, and learning testing conditions. Results show highly accurate model estimation in different testing scenarios. Additionally, the nonlinear dynamic inversion controller easily integrates the identified model parameters without any need for gain scheduling or computationally expensive methods. Overall, the proposed technique introduces an efficient integration between real-time modeling and control adaptation utilizing the limited computational power of the quadcopter’s microcomputer.

Effective geographically dispersed student teams – A teleoperated systems design case study

In the academic year 2016-17 the University of Missouri-Kansas City and the University of Southampton teamed up for a coordinated capstone design project that integrated expertise from both institutions. The design project was focused on the teleoperated deployment of atmospheric sensing equipment into localised, severe weather events. Seven aerospace engineering undergraduate students at the University of Southampton designed, fabricated, and tested a remotely-piloted aircraft that was capable of delivering atmospheric sensing packages to a target location; they also developed the telemetry-enabled weather observation platforms themselves. Simultaneously, four mechanical engineering undergraduate students from the University of Missouri developed, constructed, and tested the ground-based component of the observation system: a remotely-operated rover that could carry the aircraft and launch it from a pre-selected site in the close proximity of the targeted weather event. In April 2017 the Southampton team traveled from the United Kingdom to Kansas City, MO for a complete operational test demonstration. This paper outlines the motivation for the design activities, as well as the student efforts throughout the project, also looking at the collaborative aspects of the project and the coordination from universities more than 4,000 miles apart.