Due to coronavirus, there wasn’t a 2020 HESTEMP conference, but work and progress for the academic year still continued.
Integrated Extremal Control and Explicit Guidance for Quadcopters
The research study aims to create a framework for autonomous control technology for unmanned aerial vehicles with real-time target-relative guidance capabilities, which leverages onboard decision-making to provide targeting and re-targeting solutions. Thus, this paper aims to develop extremal control and guidance functions in the context of the optimal control problem and their integration for applications. Solving the optimal control problem leads to a constant motor thrust case and trivial and nontrivial cases for the variable motor thrust case. As illustrative examples, two quadcopter maneuvers use integrated extremal control and explicit guidance. The first maneuver is the quadcopter taking off to the desired altitude using maximum and then intermediate thrust. The second maneuver has the quadcopter traveling to a waypoint over an agricultural field. The DJI Onboard Software Development Kit provides a method to implement the proposed integration of extremal control and explicit guidance onboard a Raspberry Pi connected to the DJI M100 quadcopter. Simulated and experimental flight tests demonstrate that the integration of extremal control and explicit guidance allows the DJI M100 to reach the desired locations and velocities for both maneuvers.
Utilizing DJI OSDK allows us to implement extremal control and explicit guidance for quadcopters. The video of the flight test is shown below.
JINT article
Correction to JINT article (figure error)
Studying Navigation and Guidance Functions using Unmanned Aerial Systems for Real-World Applications
This work details the methods and approaches ACTUAS members used to study navigation and guidance problems for real-world applications. To explore navigation problems in UAS, ACTUAS members build upon DJI’s platform by adding additional state inputs using a databot sensor. ACTUAS members created an OSDK application to study guidance functions by incorporating the exponential braking law using DJI’s control examples and ACTUAS’s leader’s OSDK application as a reference. ACTUAS members obtained access to a Bixler trainer plane to safely practice real flight and become accustomed to the transmitter and controls. The members are currently preparing for the FAA Part 107 exam to obtain official certification.
[ACTUAS] May 2020 HESTEMP Presentation – Sean and Kevin
Quadcopter Stabilization On Mars
This study analyzes how a quadcopter drone would react in different gravitational fields and atmospheric densities to determine which celestial bodies the drone could stabilize on. After running a series of tests, it was discovered that the drone was unable to stabilize on Mars. Another series of tests were run to answer the question, “what hardware modifications would it take for a drone to stabilize on Mars?” It was found that the motors and batteries need to be upgraded in order for the drone to stabilize. Stable drone flight on other celestial bodies, such as Mars, is desirable because it allows astronomers to learn and explore celestial bodies more efficiently. Rovers are constrained to the ground and have limited exploration capabilities, while drones have a much greater range of movement. For the purpose of this paper, perfectly stable quadcopter drones have zero angular velocity, and stability simulations are conducted in MATLAB. The simulation was created by Gibiansky (2015), and the physical characteristics of a DJI Mavic 2 Pro (M2P) are used in the MATLAB simulation. The simulated quadcopter is dropped into a given atmosphere and gravitational field with some random orientation and it attempts to stabilize. Once the simulation is complete, the max motor spin rate (MMSR) of the simulated quadcopter is compared to the physical limits of the M2P motors to determine feasible flight.
[ACTUAS] May 2020 HESTEMP Presentation – Kian