Hi there! My name is Peter Jansen, the developer of the Tricorder project, and if you're here you're probably interested in knowing a little more about me.
Background and Story
I have a bit of an eclectic educational background, in part because of my belief in interdisciplinary education.
I am a recent PhD graduate of the Cognitive Science Laboratory at McMaster University in Hamilton, Ontario, Canada. My research interests are broadly in Cognitive Artificial Intelligence, and more specifically in something called Developmental Knowledge Representation. I'm interested in how people, but mostly babies, learn to represent things like concepts and language in their brains, and my research develops computer simulations that run on supercomputers to try and make computers learn language like babies do. My research comes from the perspective of something called Representational Grounding, which means that I believe babies (and eventually, computers) largely learn concepts and language by virtue of their exposure to the world, and to this end I'm also interested in how we can give robots infant-like perception (like vision and hearing and touch).
My undergraduate is a specialized research-oriented degree in Independent Studies done at the University of Waterloo, specializing both in physics (astro and optical), as well as cognitive artificial intelligence, with an official cognitive science option, and combines computer science, physics, computational neuroscience, cognitive psychology, and computational linguistics.
I am currently completing my interdisciplinary Postdoctoral training in adaptive and compressive sensing in the Lab for Engineering Non-Traditional Sensors at the University of Arizona. My research projects involve adaptive spectral imaging as well as rapid prototyping for terahertz optics. I also provide high-performance computing expertise for a massively parallel gigapixel-scale imaging project.
My CV, as well as a graphical portfolio of technical projects, is available online here (pdf).
Because Postdoctoral Fellowships are by nature temporary training positions, I'm actively on the look-out for my next academic or industry position.
I am a broadly interdisciplinary artificial intelligence researcher with interests in cognitive artificial intelligence, robotics, language, and massively
parallel computing. I work best in an openly creative and collaborative environment where my lab or research mates are as generous with their time, energy,
and thoughtful discussions as I will be with mine. As a junior research scientist I have experience teaching and supervising undergraduate and graduate
students, and am an award-winning presenter.
As examples of recent work, my self-organizing neural network model of infant language learning is (to the best of my knowledge) the top performing model in terms of grammar learning and abstraction on infant-scale grammars consisting primarily of nouns and verbs. In other recent work, my computational model of neonate saccades captures much of what is known about newborn visual attention using simple neural mechanisms and representations.
Please feel free to send me a note at email@example.com if you think my expertise and research interests might fit well with your lab or group.
Open Design Projects
I have a creative and non-traditional approach to design, and this is reflected in each of my projects, whether they be designing neural network models, 3D printers, inexpensive handheld sensors, or intuitive graphical presentations. In the context of open source hardware, I especially enjoy working under artificial constraints that allow rethinking traditional design elements in the spirit of improving accessibility. Below are a few of these projects.
Open Selective Laser Sintering 3D Printer (OpenSLS)
Selective Laser Sintering (or SLS) is a rapid prototyping technology that
constructs an object layer-by-layer by selectively fusing (or sintering) a layer of powdered thermoplastic with a laser. Compared to fused-deposition
modelling based 3D printers (such as the RepRap and Makerbot), SLS based printers
have the potential to build objects with more complex geometries and with higher resolution.
The OpenSLS project aims at creating an extremely inexpensive (under $200) SLS 3D printer that can be manufactured almost entirely using a laser cutter and a hand full of nuts and bolts. With these constraints in mind, I was able to design almost entirely laser-cuttable CNC motion elements, including a CNC linear axis, and dual Z-tables to serve as SLS build and feed chambers.
Open Reciprocating Laser Saw
As a spin-off of the OpenSLS project, I had hoped to be able to design a method for the relatively low-power OpenSLS laser to be used both for SLS, as well as
for cutting a small selection of materials (such as 2mm thin absorptive thermoplastic sheets). This would create a hybrid system able to both print complex geometries,
as well as cut relatively large structural pieces, and open the potential for RepRap-style partial self-replication.
The method I designed dynamically adjusts the focus of the laser, allowing it to slowly "bore-cut" through a small selection of thin materials using a reciprocating motion.
Open 3D-Printable Laser Cutter
Commercial laser cutters are incredibly useful tools, but also very expensive. Recently folks in the maker community have designed a number of inexpensive
do-it-yourself alternatives based on aluminum extrusions and a linear bearing system called Makerslide. Alongside this
effort, I designed and fabricated a prototype 3D-printable laser cutter design that could be constructed on an inexpensive
RepRap 3D printer.
While constructing a laser cutter out of thermoplastic — a material that could be nearly instantaneously vaporized by a small beam misalignment — should make anyone a little jittery, the project took the form of designing a set of reusable CNC elements for aluminum extrusion, including stepper motor holders, pillow block bearing mounts, and laser tube holders, and may be of utility for derivative 3D-printable mills and other CNC systems. While I've replaced many of the printable parts with aluminum for safety, I still use many of the elements — such as the stepper motor mounts — from this design experiment.
The project source files are available on Thingiverse, where it was voted by Thingiverse users as the most popular 3D-printable thing for nearly a month.