Novel Hardware DevelOPMENT
The most demanding engineering problems tend to resist clean disciplinary boundaries. The projects collected here each began in that territory, where the required solution did not yet exist and the path to it had to be assembled from disciplines that had not been combined in that particular way before. In each case the work moved from first principles through iterative lab development and quantitative validation before the system was trusted in deployment. What each required first was figuring out how to build it.
Miniature Smart Material Servohydraulic System
Built from scratch electronics and control systems for a first-of-kind piezoelectric servohydraulic UAV flight control system, from circuit board design through flight certification and successful flight test.
The goal was to replace a conventional control surface actuator on a Scan Eagle UAV with a miniature servohydraulic system driven by a piezoelectric pump scaled down from an existing larger design. There was no established template for this integration. As the sole electronics and control systems engineer on a three-person team, the work included designing and fabricating custom circuit boards for a 200V piezoelectric driver, signal conditioning, position sensing, and embedded control, all sized to stack within a toaster-sized fuselage compartment with no existing mounting protocol. Nearly every stage of development produced an unforeseen problem. The piezoelectric stack required preloading via low-side hydraulic pressure, which made the system sensitive to any fluid leaks. Drive waveform shaping had to be developed to prevent fatigue cracking of the piezo stack. Thermal management required deliberate airflow provisions. The entire system was constrained to 5 amps from the aircraft bus, a hard limit that drove extensive iteration on capacitor selection and NTC current limiting components for the DC/DC converters. After passing g-force certification testing for catapult launch and wire-catch landing, including post-test board reinforcement for massive components, the system flew successfully with no issues.
Adaptive Damping Using Magnetorheological Inchworm Actuation
Designed and built a novel actuator and damper combining magnetorheological fluid clamping with piezoelectric drive, resolving unknowns in magnetic circuit design, fluid mechanics, and material behavior along the way.
Magnetorheological fluid is oil suspended with microscopic iron particles. In the presence of a magnetic field the particles align into chains across the flow path, and flow cannot occur until the applied shear force is sufficient to break those chains. The shear threshold is a function of field strength, which makes the fluid’s resistance to flow controllable in real time. This property is well established for damping applications. The goal here was to push it into a different role entirely, using the fluid’s ability to arrest flow as the clamping mechanism in a piezoelectrically driven inchworm actuator.
The design required resolving a fundamental geometric tradeoff. The cylindrical gap had to be wide enough to produce useful damping variation but narrow enough that the coil current could generate sufficient clamping force. Magnetic circuit analysis, fluid damping models, and coil geometry had to be developed together before any hardware was built. A custom preload fixture incorporating a load cell was developed to assemble the piezoelectric stack under controlled compression, with preload rods selected to maintain force without absorbing too much of the stack’s usable stroke.
The device demonstrated the intended damping behavior. What it did not do was clamp as predicted. The manufacturer’s Bingham plastic model indicated the fluid would arrest flow below a threshold shear stress. In practice the annular fluid volume exhibited compliance that was not captured in the datasheet, stretching under load before flow began rather than locking cleanly. The inchworm action was compromised by this undocumented material behavior. The project ended without funds for a second prototype, but the testing produced a clear understanding of exactly what had occurred and what would be required to address it.
Active Structural Vibration Control of an Aircraft Fin
Designed and instrumented two fin assemblies for wind tunnel testing of a combined active vibration and flow control system addressing a fatigue failure problem with no practical single-discipline solution.
Tactical aircraft carry underwing stores of varying shapes and sizes. The vortices shed by those stores travel aft and produce cyclic loads on the aircraft fin surfaces that were never fully accounted for in the original airframe designs, and the result is premature fatigue failure. Three approaches to the problem were known. Bonding piezoelectric bimorph actuators to the fin surface can stiffen it against cyclic strain, but the power and scale required make it impractical at operational size. Disrupting the flow around the stores can push vortex formation downstream past the fin but introduces drag. Pulsed air jets can provide active flow control but face the same scaling constraints. This project investigated whether combining active vibration control with active flow control could produce a system more effective than either approach alone, at a scale that could realistically be implemented on a full-size aircraft.
The active vibration control system was developed at CSA Engineering in collaboration with Lockheed, who developed the active flow control system and operated the wind tunnel facilities. Two fin assemblies were designed and fabricated, one larger and more flexible for low-speed testing and one shorter and stiffer for transonic and supersonic conditions, each instrumented with piezoelectric bimorph actuators and root strain gages with custom power electronics. Both test campaigns were conducted at Lockheed facilities. The results demonstrated that the combined system outperformed either approach in isolation and suggested that a scaled implementation on a full-size aircraft was feasible.
Piezoelectric Global Shape Control System
Developed analytical methods and custom instrumentation to control the global shape of a flexible structure using 32 individually addressable piezoelectric electrodes, targeting adaptive optics basis shapes from first principles.
The motivation was adaptive optics, where the ability to precisely control the global shape of a flexible structure has applications in precision engineering and optical systems. The approach was to bond a custom piezoelectric material to a thin beam and use an array of 32 individually addressable electrodes etched into the piezo surface to induce controlled bending moments at discrete locations along the structure. The analytical challenge was working in the opposite direction from conventional structural analysis: starting from a desired global shape, calculating the moment distribution required to produce it, and mapping that distribution to the specific voltage inputs for each electrode. Two target shapes were selected for their relevance to adaptive optics, a parabolic profile and a piston shape, and the required electrode voltages were derived analytically for each.
The experimental setup required solving a measurement problem with available resources. A laser displacement probe was used to scan the deformed surface of the beam, and the scan was executed by mounting the entire experiment to a milling machine and using its motion axes as a precision positioning system. The resulting shape measurements were compared against the analytically predicted deformations. The method successfully produced both target shapes, establishing a quantitative relationship between local actuation inputs and device-scale structural deformation.
Multi-modal human-autonomy research platform
Built a custom multi-modal research platform synchronizing EEG, gaze, physiological, motion capture, and behavioral data streams across heterogeneous hardware and software systems to enable three distinct human-autonomy teaming studies.
The three studies that comprised my dissertation research each required simultaneous measurement of human physiological state, attention, and behavior at a level of detail that no existing off-the-shelf platform provided. The solution was a custom multi-modal instrumentation platform built around Lab Streaming Layer, a network-based synchronization framework that integrates data streams from heterogeneous systems into a single timestamped structure regardless of individual sample rates. The modalities combined included EEG, gaze tracking, skin conductivity, heart rate, rigid body motion capture via an OptiTrack system, and behavioral inputs including mouse position, keystrokes, and joystick data, each running on separate computers with separate software.
The platform supported three distinct experimental contexts. A manual grinding study captured tool pose, velocity, and gaze to characterize technique differences across skill levels. A human-swarm interaction study used gaze, EEG, and joystick input while subjects drove configurations of unicycle robots, with the motion capture system doing double duty as both a measurement tool and a real-time control loop input for maintaining robot geometric configuration. A human-robot teaming study combined gaze, skin conductivity, mouse position, and EEG during a simulated collaborative task. The same underlying architecture supported all three without rebuilding.