Skip to content
Menu

Research

Nine branches, kept broad on purpose.

ME planning stays wide until a branch genuinely needs deeper design. That keeps the map useful without pretending every subsystem has a five year plan today.

  • Computing & AI

    Primary focus

    The branch with the strongest existing knowledge base at ME, and the first one being explored in depth. Agentic systems, operating systems, distributed compute, infrastructure, robotics intelligence, and scientific computing.

    • Agent native operating system design: permissions, memory, task routing, recovery
    • Command routing that many control surfaces can share safely
    • Model infrastructure and use of strong external models rather than training a frontier model
    • Scientific computing and simulation as shared infrastructure

  • Energy

    Active interest

    Initial focus is power electronics: efficient conversion, control, and distribution. That is the layer compute, robotics, labs, vehicles, satellites, and storage all depend on.

    • Power electronics for high density compute and laboratory loads
    • A smart DC bus for lab benches and factory cells, as the first concrete system to study
    • Protection coordination: which device disconnects first, and how fast, on a fault
    • Storage and grid interaction
    • Longer horizon generation systems

  • Space & Satellites

    Active interest

    Treated as a major branch rather than a subtopic. The emphasis is Earth observation delivering services governments and large organizations actually buy.

    • Imaging, mapping, and weather observation
    • Infrastructure monitoring, disaster response, agriculture, and logistics
    • Communications, orbital infrastructure, launch, and deeper space systems remain in scope

  • Robotics

    In scope

    Industrial, mobile, laboratory, and autonomous systems, including the interfaces that let agents drive them safely.

    • Laboratory automation feeding the research tools platform
    • Mobile and industrial platforms
    • Safe autonomy boundaries between agent intent and actuator command

  • Manufacturing

    In scope

    Automation, advanced factories, materials processing, and scalable production, organized around distributed automated production rather than one large plant.

    • Distributed automated production cells
    • Buy commodity inputs such as steel, wire, fasteners, bearings, and standard electronics where making them adds no strategic value
    • Materials processing tied to the materials branch

  • Materials

    In scope

    Semiconductors, composites, metamaterials, and nanomaterials, chosen by what other ME branches are blocked on.

    • Semiconductors for compute and sensing
    • Metamaterials relevant to displays and optics
    • Composites and structural materials for robotics and space hardware

  • Medicine & Biotech

    In scope

    Emphasis on prosthetics, augmentation, and longevity, with diagnostics and synthetic biology kept in scope.

    • Prosthetics and human augmentation, sharing sensing work with wearables
    • Longevity research
    • Diagnostics and synthetic biology

  • Transportation

    In scope

    Autonomous systems, aerospace, high speed transit, and logistics.

    • Autonomy shared with the robotics branch
    • Aerospace and high speed transit
    • Logistics as a systems problem

  • Frontier Research

    In scope

    Speculative and advanced physics, propulsion, and spacetime work. High risk, high upside, and explicitly labelled as such.

    • Advanced propulsion concepts
    • Spacetime and gravitation
    • Work that fails honestly rather than quietly

Energy, in more detail

Power electronics: efficient conversion, control and distribution. It is the layer compute, robotics, labs, vehicles, satellites and storage all sit on top of, which makes it the useful place to start.

What real distribution has to deal with

  • Voltage limits. Every component has a voltage it is rated for, and insulation that fails above it.
  • Current limits. Conductors, connectors and switches each have a current they can carry continuously.
  • Conductor heating. Current through resistance makes heat. Heat is what sets the real limit long before anything melts.
  • Resistance losses. Energy lost in the wiring is energy paid for and not used, and it grows with the square of the current.
  • Insulation. What separates conductors from each other and from people. It ages, and it is checked, not assumed.
  • Switchgear. Making and breaking a circuit safely, including under fault conditions, is its own engineering problem.
  • Conversion. Changing voltage or converting between AC and DC costs efficiency and produces heat.
  • Cooling. Anything that dissipates power needs somewhere for that heat to go.
  • Fault protection. Detecting a short or an overload and disconnecting fast enough to prevent damage or injury.

Smart DC bus for labs and factory cells

Concept

A shared DC distribution bus for a lab bench or a factory cell, with monitoring and protection built in rather than bolted on. Most equipment on a bench converts incoming AC to DC anyway, often badly and repeatedly, so distributing DC once and well is worth studying.

Open questions

  • What voltage makes sense for a bench and a cell, given the equipment that would actually connect to it
  • How protection coordinates: which device disconnects first, and how quickly, when something faults
  • How a cell reports its own consumption and temperature without adding a second wiring system
  • What happens on a fault, in enough detail that the answer is designed rather than discovered

Electrical work is dangerous and its rules exist because of accidents. Nothing here is a design, a specification or advice, and nothing has been built or energised.