Projects

My research focuses on embedded systems. It is the kind of field where you debug software with an oscilloscope. Of particular interest to me are low-level architectures and system interfaces.

Feel free to contact me if you have any questions: brghena@umich.edu


Tock

Embedded operating systems have traditionally been limited to libraries that abstract hardware and implement common utilities. These systems provide only limited mechanisms, if any, to ensure the safety of drivers and isolate applications. Instead, developers must assume that all code is equally trustworthy and bug free.

Tock is a safe, multitasking operating system for memory constrained devices. It is written in Rust, a type-safe systems language which is used to enforce safety of of kernel components. Tock allows applications, which can be written in any language, to be compiled and loaded independently from the kernel.

[source]

A workshop paper on our experiences building an embedded operating system in Rust was published at PLOS'15.

Abstract: Rust, a new systems programming language, provides compile-time memory safety checks to help eliminate runtime bugs that manifest from improper memory management. This feature is advantageous for operating system development, and especially for embedded OS development, where recovery and debugging are particularly challenging. However, embedded platforms are highly event-based, and Rust’s memory safety mechanisms largely presume threads. In our experience developing an operating system for embedded systems in Rust, we have found that Rust’s ownership model prevents otherwise safe resource sharing common in the embedded domain, conflicts with the reality of hardware resources, and hinders using closures for programming asynchronously. We describe these experiences and how they relate to memory safety as well as illustrate our workarounds that preserve the safety guarantees to the largest extent possible. In addition, we draw from our experience to propose a new language extension to Rust that would enable it to provide better memory safety tools for event-driven platforms.

[workshop paper]


PowerBlade

With an increasing number of miscellaneous electrical loads used in residential and commercial buildings, understanding electricity usage patterns requires metering every load. Existing plug-load power meters are bulky, power-hungry, and require manual user interaction to record energy use. PowerBlade is a new take on power meter designs that has an essentially 2D form factor. Loads literally plug-through it and into the outlet. This means PowerBlade can stay attached to the load plug itself, truly monitoring an individual device, rather than an outlet. Designing a functional power meter with extreme volume constraints requires rethinking the traditional power meter design choices. PowerBlade employs a simplified AC power supply design, a novel method for noninvasively sensing current draw, and a Bluetooth Low-Energy radio to transmit data to nearby smartphones. This work was published and demonstrated at SenSys 2015.

[source] [video]

Abstract:
We present PowerBlade, the smallest, lowest cost, and lowest power AC plug-load meter that measures real, reactive and apparent power, and reports this data, along with cumulative energy consumption, over an industry-standard Bluetooth Low Energy radio. Achieving this design point requires revisiting every aspect of conventional power meters: a new method of acquiring voltage; a non-invasive, planar method of current measurement; an efficient and accurate method of computing power from the voltage and current channels; a radio interface that leverages nearby smart phones to display data and report it to the cloud; and a retro power supply re-imagined with vastly lower current draw, allowing extreme miniaturization. PowerBlade occupies a mere 1" × 1" footprint, offers a 1/16" profile, draws less than 180 mW itself, offers 1.13% error on unity power factor loads in the 2-1200 W range and slightly worse for non-linear and reactive loads, and costs $11 in modest quantities of about 1,000 units. This new design point enables affordable large-scale studies of plug-load energy usage—an area of growing national importance.

[conference paper] [demo paper]

This work was also included in a journal paper covering the topic of perpetual sensing in the built environment.

[journal paper]


Wearabouts

"Who's in lab?"

This is a harder question to answer than it may seem. Determining the location of individuals indoors remains an unsolved problem.

The goal of Wearabouts is to perform room-level localization of individuals. The expected output would include each discovered individual and which room they have been found in. An important piece of the design is to not require individuals to carry any extra devices. We want to use only things they already carry and characteristics of who they are to discover them. Currently, the system uses Bluetooth Low-Energy advertisements from devices like Fitbits to discover people and adds information from RFID door entry and smartphone WiFi connections to increase its accuracy.

[demo] [source]

Terraswarm Annual Meeting 2014:
[poster]

Terraswarm Localization Workshop 2014:
[poster] [presentation]


Thermes

Taking the Monjolo Principle of energy-harvester as a sensor, we created a new platform with a heat harvesting thermoelectric generator. This work was published at ENSsys 2014.

Abstract:
Fine-grained energy metering in homes and buildings provides a promising technique for addressing the unmaintainable energy consumption levels of worldwide buildings. Metering electricity, lighting, natural gas, HVAC, occupancy, and water on a per appliance or room basis can provide invaluable insight when trying to reduce a building's energy footprint. A myriad of sensor designs and systems collect data on particular building aspects, but are often hampered by installation difficulty or ongoing maintenance needs (like battery replacement). We address these common pitfalls for water and heat metering by developing a small, energy-harvesting sensor that meters using the same thermoelectric generator with which it powers itself. In short, the rate at which the harvester captures energy is proportional to the heat production of the monitored appliance or pipe and this relationship allows us to estimate energy use simply based on the sensor's ability to harvest. We prototype our sensor in a bracelet shaped form-factor that can attach to a shower head pipe, faucet, or appliance to provide local hot water or heat metering.

[paper] [presentation]


Green Lights Forever

Starting as a course project in EECS 588, a group of three other students and I investigated the security of an intelligent traffic light system in the United States. This work was published at WOOT 2014.

Abstract:
The safety critical nature of traffic infrastructure requires that it be secure against computer-based attacks, but this is not always the case. We investigate a networked traffic signal system currently deployed in the United States and discover a number of security flaws that exist due to systemic failures by the designers. We leverage these flaws to create attacks which gain control of the system, and we successfully demonstrate them on the deployment in coordination with authorities. Our attacks show that an adversary can control traffic infrastructure to cause disruption, degrade safety, or gain an unfair advantage. We make recommendations on how to improve existing systems and discuss the lessons learned for embedded systems security in general.

[paper] [presentation]


NSF Fellowship

In my 2013 I applied to the NSF Graduate Research Fellowship Program and was lucky enough to be awarded a fellowship. This was the first year with the new essay rules that reduce the application to two essays, Background & Previous Research and Proposed Research, a total of five pages. Below, I have provided my essays as well as the reviews I received as examples for other students applying to the program.

[background and previous research] [proposed research] [reviews]


SensEye

The eyes are the window to the soul. What could a computer do if it knew what you were looking at?

SensEye is an attempt to start answering that question. Low-power real-time gaze detection is finally within our grasp. Using a microcontroller running uCLinux and an FPGA, we can pull real-time images from two small cameras, one facing inwards towards the eye and one facing outwards towards the world. This data can be used to determine which object in view you are observing. Imagine a world where your car would warn you if you were not paying attention, where your phone would have interesting information available on the item you were just observing, and where you could track how you actually spend your time each day.

In order to achieve this goal, our system has to be both energy efficient and low cost. We hope to create such a system and distribute it to software developers in order to spur innovation.

[source]


Oculus-ASR

During my undergraduate studies at Michigan Tech I was part of the Aerospace Enterprise, a group of students building a nanosatellite as part of the University Nanosatellite Program. We won the 6th iteration of the competition and have been finalizing flight hardware. The mission is set to launch using the Falcon Heavy on mission STP-2.

The purpose of the satellite is to help test United States' space situational awareness capabilities. Using spectroscopy, the attitude and configuration of a satellite can be determined if it is first characterized on the ground. Oculus-ASR has been characterized and has the ability to control and observe its orientation and configuration. The Air Force will use our satellite to calibrate their algorithms and enable new advances in space awareness.

JPL Cubesat Symposium 2013:
[presentation]